Image of the Week
Edited and moderated by Stacey Mills, MD, Pathology Network's Image of the Week Blog is a forum for the discussion of interesting and often diagnostically challenging pathology images.
Friday, May 10, 2013
A 48-year-old man had gradually progressing decreased hearing in his right ear over several months. Routine physical examination demonstrated a conductive hearing loss. An otoscopic exam showed an apparent mass behind the tympanic membrane and a clinical diagnosis of "glomus tumor" was made. An excision was performed. The small mass was adherent to the ossicles but was easily removed with minimal blood loss. Representative sections are shown below.
The lesion consists of loosely cohesive aggregates of polygonal epithelioid cells with prominent glassy cytoplasm imparting a plasmacytoid or rhabdoid-like appearance. The cells are relatively uniform with no evidence of mitotic activity.
What is your diagnosis?
This is an example of a middle ear adenoma (MEA), a benign lesion of middle ear lining mucosa that can have a quite variable microscopic appearance. In addition to the discohesive plasmacytoid cell appearance seen in this example, these tumors may form sheets, nests, trabecula, glandular structures and other adenomatous patterns.
Immunohistochemical studies will frequently demonstrate neuroendocrine differentiation in the neoplastic cells of MEA, but this feature doesn't not correlate with more aggressive clinical behavior and such tumors should not be labelled as "carcinoid tumors" or "well-differentiated neuroendocrine carcinomas." However, in acknowledgement of this feature, the term "middle ear neuroendocrine tumor" has also been applied to these lesions. The latter term, however, imparts no information about the malignant/benign status of the lesion, and for that reason, I prefer to label these as middle ear adenomas.
MEAs are slowly growing tumors that may have been present for many months or even years. They are usually cured by simple excision, but they may recur. They do not invade bone, but may tightly adhere to the middle ear ossicles and exhibit limited extension along branches of the facial nerve. Although MEAs are often confused clinically with paraganglioma ("glomus tumor") of the middle ear, they are easily distinguished histologically and by their strong cytokeratin positivity. At the time of surgery MEAs lack the hypervascularity and often excessive bleeding associated with the resection of a middle ear paraganglioma. MEAs should also be distinguished from the much more aggressive papillary endolymphatic sac / temporal bone tumor that is frequently associated with von Hippel-Lindau syndrome. MEAs lack papillary structures.
Wednesday, April 10, 2013
A 65-year-old woman presented with a flank mass. A biopsy was performed and the resultant specimen is illustrated below.
The mass consists of a cellular fibroinflammatory proliferation with prominent histiocytes, plasma cells and lymphocytes. The histiocytic component predominates and consists of markedly enlarged cells with eosinophilic to foamy cytoplasm and easily identified intracytoplasmic lymphocytes and plasma cells. Immunohistochemical studies showed that the histiocytic cells were positive for S100 protein and negative for CD1a.
What is your diagnosis?
This is an extremely nice example of sinus histiocytosis with massive lymphadenopathy (SHML), also known as Rosai-Dorfman disease. Hint: To get a disease named after you, give it is helpful to give it a very long name originally! The histiocytic cells demonstrate florid emperipolesis, a classic finding in this lesion. As the term, sinus histiocytosis with massive lymphadenopathy implies, this process was first described in and most frequently involves lymph nodes. However, virtually any anatomic site may be involved with skin, sinonasal region, soft tissues and bone being common extranodal locations. SHML occurs over a wide age range from newborns to the elderly.
Although the histiocytic cells of SHML share S100 protein positivity with Langerhans cell proliferations, they are not considered to belong to the family of Langerhans or dendritic cells and their lack of reaction for CD1a and Langerin are of value in making this distinction diagnostically. SHML is typically indolent and self limited, although prognosis does correlate with the number of involved sites. Patients with persistent or progressive disease have been variably treated with surgery, radiation, and chemotherapy. Patients with widespread disease often have an associated immune dysfunction. Although SHML is occasionally a cause of death, I am not aware of any cases undergoing malignant transformation.
Monday, March 25, 2013
A 56-year-old man complained of a mass involving his right foot. MRI showed a 4 cm mass adjacent to and perhaps involving the anterior tibialis tendon. The mass was also closely approximated to the first metatarsal-phalangeal joint. A needle biopsy was performed and the resultant images are seen below.
The mass consists of a bland but hypercellular proliferation of spindled cells, some with more prominent vacuolated cytoplasm. Mitotic figures are scarce and there is no evidence of necrosis. The interface with the adjacent tissue is not easily visualized in this biopsy.
What is your diagnosis?
Additional studies were performed and the tumor cells were noted to be negative for cytokeratin, EMA, CD56, CD34, and desmin. They were positive for S100 protein, HMB-45, and Melan-A. The HMB-45 staining is seen below.
FISH studies showed that the SYT gene was intact, but the EWS gene was rearranged. Since breakapart probes were utilized, the fusion partner was not determined but is presumed to be the ATF1 or. less likely, the CREB1 gene.
Accordingly, this is an example of a clear cell sarcoma. Interestingtly, the heterogeneous group of neoplasms harboring EWS-ATF1/CREB1 mutations are somewhat unified by the strong, but not universal, tendancy for tumors in this group to have at least focally clear cytoplasm due to glycogen deposition.
Although once interpreted as "melanomas of soft parts" because of their often obvious melanogenesis, these tumors are fundamentally different from cutaneous or mucosal melanomas as the latter lack EWS-ATF1/CREB1 mutations and, instead, often contain BRAF mutations which are lacking in clear cell sarcomas. The tumors typically involve the deep tissues of the extremities and, as in this case, the foot is the most common location.
The microscopic appearance is somewhat variable with cells ranging from highly spindled to somewhat more epithelioid in appearance. A distinct nested or fascicular growth pattern is common. Cytoplasm varies from clear (due to cytoplasmic glycogen) to eosinophilic. Melanin may be visible microscopically and in some instances, grossly.
As noted in this case, melanoma-related markers such as S100 protein, HMB-45, Melan-A, MiTF and others are typically positive and of considerable diagnostic value.
Although rather bland in histologic appearance, and thus not amenable to conventional grading systems for soft tissue sarcomas, clear cell sarcoma is a high-grade malignancy with local recurrence rate of 15-40% and a metastatic rate, typically to lungs and lymph nodes, approaching 50%. Late metastases have been described several decades after original diagnosis, so long-term follow up is clearly required. Therapy is surgical and adjuvant methods have thus far had little effect.
Wednesday, March 13, 2013
A 77-year-old woman presented with a subcutaneous abdominal mass. She had a history of a resected parotid gland tumor an unknown number of years earlier. The abdominal mass was excised and histologic sections from it are shown below.
Microscopically, the tumor has a remarkably heterogeneous appearance and consists of nests, cords, sheets, and tubules of epithelial cells embedded in a stroma that varies from fibrous to myxo-chondroid. Well-formed hyaline cartilage is not seen, although several areas approach this appearance. The lesion is at least partially surrounded by a fibrotic capsule. The epithelial cells lack any significant nuclear pleomorphism and mitotic figures are extremely rare. The neoplastic cells are positive for smooth muscle actin and S100 protein (not shown).
What is your diagnosis?
This lesion has all of the microscopic features of a salivary gland-type mixed tumor. Histologically identical tumors may occur in the skin where they have been termed "chondroid syringomas" as well as "cutaneous mixed tumors." The appearance of this case and the history of a prior parotid gland tumor was intriguing. The parotid tumor was ultimately obtained and proved to be a histologically identical mixed tumor. This leads to the obvious question of whether the subcutaneous tumor is an independent cutaneous mixed tumor or whether the lesion represents a so-called benign metastasizing mixed tumor. The latter is an extremely rare by well-recognized phenomenon. We have seen only a small handful of such cases over the last three decades and often the secondary lesion involves the skin, raising the issue of whether these are indeed metastases or independent tumors.
Several studies have addressed the issue of benign metastasizing mixed tumor and have invariably concluded that there are no clinical or histologic features that allow one to predict which of these tumors might rarely pursue this clinical course. As the name implies, the presence of isolated metastases in this setting does not predict a malignant clinical course and the patients do extremely well. A number of other benign tumors have been noted to be associated with "bengn metastases" including, for example, uterine leiomyomas and giant cell tumors of bone.
Molecular techniques could potentially be brought to bear to compare these two lesions and possibly answer this question. My own prejudice in this instance is to fall back on "Occam's razor" and assume that the simplest answer, ie. one tumor with a "benign" metastasis is the correct one, but I have no proof as yet to support that interpretation. Regardless, this patient should have an excellent prognosis.
Wednesday, February 13, 2013
A 70-year-old woman was noted on physical exam to have a uterine cervical mass. A cervical conization was performed and several images from the resultant specimen are shown below.
There is a prominent lymphoplasmacytic background highly evocative of a lymphoproliferative process. Embedded within the inflammatory background are ill-formed nests and cords of larger cells with variable eosinophilic cytoplasm. The nuclei in the latter cells are considerably enlarged, but nuclear chromatin is uniform from cell to cell. Cell borders are indistinct and the larger nuclei appear to float in a syncytium of cytoplasm. Mitotic figures, including atypical forms are easily found in this cell population.
What is your diagnosis?
Several immunohistochemical stains were performed and are shown below.
CD45 (LCA) (above) strongly stains the prominent background lymphoid infiltrate, but the larger cells are negative.
The larger cells were strongly positive for cytokeratin as seen above.
The cytokeratin positive cells also strongly and diffusely expressed p16 seen above.
In-situ hybridization for human papillomavirus was not performed but would be expected to be strongly positive.
This is an example of a lymphocyte-rich variant of cervical squamous cell carcinoma that has also been referred to as lymphoepithelioma-like carcinoma because of its strong resemblance to nasopharygeal lymphoepithelioma. The latter tumor, of course, is causally related to Epstein-Barr virus whereas the current tumor is undoubtedly due to human papillomavirus. Tumor site overides histologic appearance with regard to causation. Some earlier studies utilizing PCR instead of in-situ hybridization purported to show that EBV was present in a substantial number of these tumors. However, this almost certainly represents false positivity due to EBV-containing lymphocytes common in the general population. In contrast high-risk HPV is easily demonstrated by in-situ hybridization in these tumors.
The main reason for recognizing this histologically distinct variant is to avoid confusing this process with a lymphoma. The advent of type-specific immunohistochemical markers makes short work of this distinction provided the proper stains are ordered. If an epithelial marker such as cytokeratin, p63 or EMA is not employed, the large numbers of adjacent lymphocytes can give a false sense of positivity in the epithelial cells on CD45 staining. Tumors with this histologic appearance occur at a wide variety of anatomic sites where the above differential also occurs.
Whether cervical lymphoepithelioma-like carcinoma has clinically distinctive features is not entirely clear. Some studies have suggested that stage for stage these tumors have less frequent lymph node metastases and better overall prognosis when compared to conventional squamous carcinoma of the cervix. Other studies have also shown that all types of cervical carcinoma with a prominent inflammatory infiltrate tend to have a better overall prognosis.
Thursday, January 31, 2013
A 56-year-old man without symptomatology presented for routine annual physical examination and was found to have a 1.5 cm. left testicular mass. The mass was confirmed by ultrasonography and an orchiectomy was performed. Images from the resultant specimen are shown below.
Within the testis parenchyma is a sharply demarcated mass composed of sheets of cells with abundant eosinophilic cytoplasm. There is no evidence of glandular differentiation, although some entrapped seminiferous tubules are noted within the lesion. The neoplastic cells have relatively unifrom nuclei with medium-sized nucleolil. Mitotic figures are rare and there is no evidence of necrosis.
What is your diagnosis?
This is a straightforward example of the most common sex-cord tumor in the testis, a Leydig cell tumor. When Leydig cell tumors present in pre-pubertal males they are typically associated with clinically obvious isosexual pseudoprecocity. It adults, however, signs of androgen excess are not typically obvious, with the exception that about 1/3rd of patients will develop gynecomastia.
The sheet-like growth pattern of polygonal cells seen in the current case is the most typical pattern, although the cells also may form ribbon, cords, and small nests in a more prominent fibrous stroma, or exhibit a clearly spindled morphology. Although Reinke crystals are often obvious in hyperplastic Leydig cells, in my experience they can be quite difficult or impossible to find in Leydig cell tumors. I could not find any convincing examples in the current case.
Although seldom required, immunohistochemical markers may be of value in recognizing Leydig cells. The tumors (and their normal conterparts) are strongly positive for inhibin-alpha, melan-A, calretinin, and vimentin. There may be patchy positivity for low molecular weight cytokeratins.
Leydig cell tumors may be confused with the eosinophilic histiocytes of malakoplakia, but the latter process is invariably associated with other forms of inflammation and has characteristic Michaelis-Gutmann bodies. If necessary, as for example in small biopsy specimens, immunohistochemistry will allow ready distinction. The neoplasm most likely to be confused with Leydig cell tumor is the large cell calcifying variant of the Sertoli cell tumor. The latter are more commonly mulifocal, may exhibit intratubular growth, have more prominent stroma and, as the name indicates, frequent calcifications. Immunohistochemical studies show considerable overlap with Leydig cell tumors. About a third of patients with large cell calcifying Sertoli cell tumors have Carney syndrome or Peutz-Jeghers syndrome.
Less than 10% of Leydig cell tumors have been said to behave in a malignant and in all likelihood that number is inflated due to over reporting of aggressive cases. As with all endocrine tumors predicting malignant behavior is notoriously difficult. Factors associated with malignancy have included increasing age, larger tumor size, infiltrating margins, necrosis, significant nuclear atypia, increased mitotic rate, atypical mitotic figures, aneuploidy, vascular invasion, and increased expression of p53. Gynecomastia has been associated with a decreased risk of malignant behavior and no malignant cases have been documented in prepubertal boys.
Wednesday, January 16, 2013
A 67-year-old man underwent a right colectomy for a cecal mass. Microscopic sections from the resultant specimen are shown below.
The tumor consists of sheets of neoplastic cells with greatly enlarged but relatively uniform nuclei with prominent nucleoli. Mitotic rate is brisk with numerous mitotic figures easily identified, including a few overtly atypical forms. No glandular or squamous differentiation is present. There is a prominent lymphoid infiltrate associated with and infiltrating the tumor cells.
What is your diagnosis?
This is an example of medullary (undifferentiated, non-neuroendocrine) carcinoma of the colon. The tumor cells were negative for both CK20 and CDX2 (not shown) indicating absence of the conventional colonic adenocarcinoma phenotype. Immunohistochemical staining for mismatch repair protein MLH1 is shown below.
The tumor cell nuclei are completely negative with strong positivity in the associated lymphoid cells.
Immunhistochemical reactivity for mismatch repair protein PMS2 was similarly absent, while MSH2 and MSH6 were intact. Thus, this tumor has features of a carcinoma with microsatellite instability (MSI). About 2/3rds of MSI carcinomas are sporadic with the remainder associated with germline mutations (HNPCC). Sporadic MSI cancers appear to be associated with hypermethylation of MLH1 as opposed to somatic gene mutation. Sporadic cases are also frequently associated with BRAF (V600E) mutations, and appear to arise via the serrated polyp pathway, whereas germline mutation cases seem to arise via adenomatous changes. MSI cases typically have wild-type p53, which is typically mutated in microsatellite stable carcinomas.
In addition to the medullary pattern seen in this case, MSI colonic carcinomas also tend to show mucinous, signet-ring, and other poorly differentiated phenotypes. The most helpful features to suspect an MSI tumor are right-sided location, generally poor (or absent) differentiation, and the presence of tumor infiltating lymphocytes. The latter was a prominent feature of the current case. There may also be a Crohn's-like reaction in the deeper subserosal tissues.
MSI colonic carcinomas have been noted to typically present at low stage and to have a better survival, stage for stage than microsatellite stable (MSS) carcinomas when treated surgically. MSI carcinomas do not respond well to 5-FU chemotherapy. Because of differences in prognosis and therapeutic response knowledge of mismatch repair protein status is important in colorectal carcinomas. Although, as noted above, there are features which should lead to a strong suspicion of an MSI carcinoma, there are enough outlier cases to support reflexive testing for mismatch repair proteins on all resected colonic carcinomas.
Wednesday, January 02, 2013
An 8-year-old boy presented with a polypoid translucent mass that could be easily seen in his upper oropharynx. On endoscopy the mass was seen to be originating from the maxillary antrum and extending from it on a long, attenuated stalk. A snare excision was performed and the specimen was submitted to pathology.
Microscopically the surface consisted of variably hyperplastic and attenuated squamous, intermediate and respiratory mucosa. Features of a Scheiderian papilloma were not present. The stroma contained scattered variably atypical, often greatly enlarged cells. Because of concern for a sarcoma, the case was referred to us in consultation. Several H&E images of the lesion are shown below.
What is your diagnosis?
The stroma is loose and somewhat edematous and constains mild acute and chronic inflammation. Scattered throughout the stroma but never forming more hypercellular aggregates, are greatly enlarged stellate stromal cells. These cells exhibit both nuclear and cytoplasmic enlargement with accordingly nearly unchanged nuclear to cytoplasmic ratios. Mitotic figures are not noted. Immunohistochemical stains for cytokeratins, muscle specific actin, smooth muscle actin, and desmin are negative. This is an example of an antrochoanal polyp with marked, but completely benign, stromal atypia.
Longstanding stromal-epithelial polyps at a wide variety of anatomic sites, including vaginal polyps, inflammatory nasal polyps, fibrous papules of the nose, and antrochoanal polyps are well known to often show considerable stromal atypia, perhaps relating to persistent trauma. Antrochoanal polyps, unlike more typical inflammatory nasal polyps, are prone to occur in younger children and in this population confusion of this benign change with a botryoid sarcoma can be a devastating pitfall. Fortunately the distinction is easy to make if one is aware of this lesion and, if necessary, confirms the absence of rhabdomyogenous differentation with appropriate immunohistochemical stains.
Unlike the botryoid variant of embryonal rhabdomyosarcoma, antrochoanal polyps with stromal atypia lack the hypercellular "cambium layer" typical of the former tumors. The atypical stromal cells are also considerably larger than the cells of embryonal rhabdomyosarcoma and lack more than rare mitotic figures, which should be typical in appearance. Some caution with immunohistochemistry is warranted, as these cells may show some minimal myofibroblastic differentiation (lacking in this case), but they will lack staining for skeletal muscle markers. Most commonly they stain only for vimentin.
Tuesday, December 04, 2012
A 55-year-old woman was noted to have nodular breast masses on routine mammographic evaluation. A needle biopsy of one of the masses was performed. The resultant specimen is seen at several magnifications below.
The nodule consists of fibroinflammatory tissue with the inflammatory component being predominantly lymphocytic, with scattered plasma cells. There is a clear-cut lobular accentuation to the inflammation, but it also extends into the surrounding dense fibrous tissue.
What is your diagnosis?
Based on the histologic findings an inquiry was made with the referring pathologist regarding whether this patient was a diabetic. It was determined that the patient was indeed an insulin-dependent diabetic, and this is an excellent example of diabetic mastopathy.
This condition is thought to be an autoimmune response, possibly due to cross reacting antibodies developed to injected non-human insulin. Although often an incidental finding, it can occasionally produce painful nodules or multiple nodules detected on mammography but often not palpable. Patients are typically young to middle-aged women.
The fibrotic stroma is thought to be derived from activated myofibroblasts and these cells will often react, at least focally, for smooth muscle actin. Although not well seen in this more well-developed example, in "early" diabetic mastopathy, the proliferating myofibroblasts may have a distinctly epithelioid appearance leading to confusion with infiltrating carcinoma. If necessary a negative cytokeratin stain should alleviate that concern.
The lymphocytic infiltrate in this condition consists predominantly of B-cells, with only a scattered T-cell component. Histologically identical lesions may occasionally occur in women who are not diabetic. The terms "lymphocytic mastopathy" or "lymphocytic lobulitis" have been applied to these cases, although it seems likely that this is essentially the same process as in the diabetic patients.
Tuesday, November 06, 2012
A 52-year-old Asian man with a history of bilateral submandibular gland and cervical lymph node enlargement underwent a biopsy of his right submandibular gland. The resultant specimen is shown in the images below.
There is a dense inflammatory infiltrate consisting predominantly of eosinophils, lymphocytes and plasma cells. Lymphoid germinal centers are prominent. An area of eosinophilic abscess formation is present in the upper center of the first image. In the background, there is a pronounced proliferation of post-capillary venules, many of which show marked peri-venular fibrosis. Peripheral blood eosinophilia was also noted.
What is your diagnosis?
This is an excellent example of Kimura's disease, a rather rare chronic inflammatory disorder of unclear pathogenesis. The clinical history in this case is quite typical; an Asian man with recurrent involvement of the deep subcutaneous tissues of the neck. Over the years there has been considerable confusion, particularly in the Western literature, of Kimura's disease with angiolymphoid hyperplasia with eosinophilia (histiocytoid / epithelioid hemangioma). The latter entity has different histologic features, particularly with regard to the nature of the endothelial cells, and does not show the marked ethnic and male prevalence of Kimura's disease.
Histologically, Kimura's disease is characterized by three components: 1. A cellular inflammatory infiltrate rich in eosinophils and associated with follicular lymphoid hyperplasia, 2. A proliferation of post-capillary venules, 3. A fibrocollagenous background.
Kimura's disease is a chronic disorder with an indolent clinical course, often waxing and waning over time. Patients typically have symptoms ranging from months to multiple years. The current patient had a 7-year history with three documented recurrences. As in the current case, peripheral eosinophilia is almost always present. Surgery is currently the mainstay of therapy, though steroids and radiation therapy have also been used. Recurrence after treatment is common.
The differential diagnosis of Kimura's disease is extensive and includes, at a minimum, Hodgkin's lymphoma, Langerhan's cell histiocytosis, Castleman's disease, florid lymphoid hyperplasia, angiolymphoid hyperplasia with eosinophilia (epithelioid hemangioma), parasitic infection, drug reactin, and Churg Strauss disease.
For an excellent review on this topic the reader is referred to the AFIP series of 21 cases published in AJSP 28:505-512,2004.
Wednesday, October 17, 2012
A 46-year-old woman underwent a laparotomy for an adnexal mass. At the time of surgery a nodular abnormality was noted in her spleen and a splenectomy was also performed. The adnexal mass was a benign ovarian cyst. The spleen contained a nodular grossly fibrotic mass measuring several cm in diameter. Sections from the splenic mass were sent for consultation. Representative microscopic images are seen below.
There are distinctly nodular proliferations of capillaries surrounded by variable concentric fibrosis. Entrapped blood vessels show perivascular hyalinized fibrosis. There is a scattered mixed inflammatory infiltrate consisting mainly of lymphocytes, plasma cells, and histiocytes within the fibrotic tissue and surrounding the vascular nodules.
What is your diagnosis?
This is an example of so-called sclerosing angiomatoid nodular transformation
(SANT) of the spleen. The angiomatous nodules are considered to represent altered, non-neoplastic red pulp entrapped by the surrounding fibrotic reaction. The cause of SANT is unclear, but similar nodular transformations have been noted in response to metastatic carcinomas.
At least some SANT have been considered to represent IgG4-related sclerosing lesions, though in the current case the total number of plasma cells was not overly high, and only a small number were of IgG4 type.
Most SANT are discovered as incidental findings, often in middle-aged women. No examples of this apparently reactive lesion have recurred or otherwise behaved in an aggressive fashion. SANT may be confused with other benign vascular lesions of the spleen including littoral cell hemangiomas and conventional hemangiomas. More importantly, it should not be mistaken for a borderline or overtly malignant vascular lesion such as "hemangioendothelioma," angiosarcoma, or Kaposi's sarcoma. The appearance of SANT is remarkably similar from case to case and awareness of this entity should avoid confusion.
Thursday, October 04, 2012
A 49-year-old woman with a history of fungal sinusitus underwent a nasal biopsy because of concern for recurrent disease. She was being treated for a relapse of her recurrent acute myelocytic leukemia (AML). The result of the biopsy is shown below in a series of three images.
What is your diagnosis?
The ciliated respiratory mucosa of the nasal cavity contains scattered cells with bizarre, irregularly shaped, enlarged nuclei. Some nuclei contain prominent nucleoli. The bizarre cells in many instances can be seen to contain surface cilia. Mitotic figures are not identified.
This is an example of chemotherapy induced severe epithelial atypia. Chemotherapy induced epithelial atypia is well described in the urinary bladder, GI tract, breast, lungs, and other organs, but has not been well described in the upper aerodigestive tract, undoubtedly because this region is much less frequently biopsied. The exception is a single article by Westra and colleague in the American Journal of Surgical Pathology in 2001.
This article reviewed 11 examples of sinonasal biopsies with cytologic features identical to the current case. Most of their patients, as in this case, had AML and were receiving alkylating agents, typically cyclophosphamide. In their series, this change was present in 14% of patients currently or recently undergoing chemotherapy with these agents. Alkylating agents are most commonly associated with epithelial atypia at other anatomic sites, as well.
The changes seen in the current case may be easily confused with dysplasia or viral cytopathic effect, both misdiagnoses that will lead to unnecessary therapy. Frozen section artifacts may add to this confusion. Viral cytopathic effect can be excluded by the absence of true inclusions and, if necessary, with immunohistochemical stains. Importantly, viral (or fungal) infection may of course co-exist with atypia!
Excluding true dysplasia may be much more difficult, particularly if the atypia occurs in squamous rather than ciliated mucosa. A history of treatment with alkylating agents, particularly for leukemia is obviously critical. If needed, immunostaing for Ki-67 will show extremely low reactivity ( <1% ) in the atypical cells in contrast to true dypslasia.
It is well-known that stromal atypia, usually related to radiation, may well continue to exist for the remainder of the patient's life. However, because of mucosal turnover, these changes appear to regress relatively soon after cessation of therapy.
Wednesday, September 12, 2012
A 16-year-old boy with no obvious genetic anomalies underwent an orchiectomy for a cryptorchid testis. The findings are shown in the multiple images below.
How many different anomalies do you see in the images above?
The seminiferous tubules show moderate peritubal fibrosis with no evidence of germ cells (germ cell aplasia, "Sertoli-only syndrome"). The interstitial Leydig cells are hyperplastic with easily detected Reinke crystals. In addition, there are multiple approximately 1 mm hyperplastic Sertoli cell nodules composed of nests and tubules of immature Sertoli cells.
All of these changes are commonly found in post-pubertal resections of cryptorchid testes. Cryptorchidism occurs in about 2% to 8% in newborns and decreases to 1% to 2% within a few months due to delayed descent. Persistent exposure of the cryptorchid testis to the normal body temperature of the abdomen or inguinal canal, as opposed to the cooler temperature in the scrotum, leads to progressive changes. These changes become particularly advanced after the onset of puberty.
The risk of subsequent germ cell neoplasia is dramatically increased in comparison to normally descended testes, and prevention of this is the primary reason for resection.
Surgically lowering the testis in early infancy, and anchoring it in the scrotum with an orchiopexy has shown some promise for maintaining fertility in adult life, although the results have been mixed, as many cryptorchid testes have other anomalies present even at birth.
Friday, August 24, 2012
A 57-year-old woman had a 2 cm mass involving the superficial lobe of her right parotid gland. A superficial parotidectomy was performed. The resultant specimen, seen in two images below consisted of a distinctly glandular neoplasm with abundant eosinophilic luminal material somewhat resembling the colloid of a follicular thyroid lesion. The luminal secretions were strongly PAS-positive (not shown). The individual cells also contained focally prominent cytoplasmic vacuoles. Nuclei were mildly variable, with some cells having prominent intranuclear cytoplasmic pseudoinclusions.
What is your diagnosis?
When I first encountered the seminal article describing this entity and published by Skalova and colleagues in the May 2010 issue of AJSP, my initial reaction was probably the same as most readers, "Why isn’t this just acinic cell carcinoma?" The authors clearly anticipated this response and go to great and convincing lengths to document why this tumor is identical at the morphologic, immunohistochemical, and molecular level, to secretory carcinoma of the breast, and hence the term "mammary analogue secretory carcinoma" (MASC) has been applied to these neoplasms.
I’m sure that these tumors have been interpreted as acinic cell carcinomas in the past, and it is highly doubtful that this misdiagnosis caused any patient disservice. Morphologically these tumors have a stereotypical appearance of rather uniform glandular cells with moderate eosinophilic cytoplasm and prominent eosinophilic extracellular material. They lack the other cell types (indeterminate, clear, etc) often encountered in acinic cell carcinoma.
The tumors are positive for a panel of markers also found in mammary secretory carcinoma, including MUC1, MUC4, GCDFP-15, STAT5a, and mammaglobin (seen below). They lack staining for estrogen, androgen, and progesterone receptors, as well as Her2, markers also typically lacking in secretory carcinoma of the breast.
Most convincingly, the authors showed that 13 of 14 MASC tested had a characteristic t(12;15)(p12;q25) translocation producing the ETV6-NTRK3 fusion. This translocation was not found in any of 12 acinic cell carcinomas tested and the ETV6-NTRK3 fusion was also found in all three mammary secretory carcinomas that the authors tested. It seems highly likely that an antibody to the resultant fusion protein will eventually be available and should prove to be relatively specific for these neoplasms in both breast and salivary gland.
Based on a limited number of cases, the prognosis for this tumor appears good with only one of 16 patients dying of disease, although several patients required surgical excision of local recurrences. This is similar to the behavior of acinic cell carcinomas.
Wednesday, August 15, 2012
A 35-year-old woman underwent a Whipple resection for a partially cystic pancreatic mass. The resultant specimen seen at several magnifications below was composed of relatively uniform polygonal cells with ovoid to elongated nuclei, some with nuclear folds. Eosinophilic intracytoplasmic and extracellular globules were quite numerous. Mitotic figures were rare. The cells grew in a variety of architectural patterns including solid sheets, glandular structures, rosettes, nests, and papillary formations. There were areas of cyst formation with associated hemorrhage and necrosis.
What is your diagnosis?
This is an example of an uncommon pancreatic neoplasm known by a large number of terms. The one currently preferred by the WHO is "solid-pseudopapillary neoplasm" of the pancreas. The eponym "Gruber-Franz tumor" has also been applied to this tumor. The histologic appearance of these tumors immediately brings to mind a differential diagnosis that includes islet cell tumors (pancreatic endocrine neoplasms) and acinar cell carcinoma of the pancreas. This distinction is of critical importance because solid pseudopapillary tumors have an excellent prognosis with a less than 10% recurrence rate.
The immunohistochemical staining pattern is of considerable value in making this distinction.
The tumors are typically negative or only weakly positive for cytokeratin as seen above.
Tumor cells will often express some neuroendocrine differentiation, exemplified in this image by positivity for CD56 seen above.
One of the most helpful stains in the differential diagnosis is a beta catenin, seen above. Solid-pseudopapillary neoplasms of the pancreas show intranuclear accumulation of beta-catenin protein and have been shown to have an associated mutation of the beta-catenin gene. Although not illustrated, the abundant eosinophilic globules are PAS positive and also react with antibodies to alpha-1-antitrypsin and alpha-1-chymotrypsin.
Because of its apparent multidirectional differentiation with both acinar and neuroendocrine immunohistochemical features, much ink has been expended discussing the "cell of origin" of these tumors. In the end, as I've discussed in a blog, we simply don't know and probably shouldn't care. A non-committed stem cell seems like a good guess!
Friday, August 03, 2012
A 52-year-old man underwent upper endoscopy for ill-defined upper GI symptoms. A 3 cm. polypoid mass with focal surface ulceration was noted in his gastric antrum. The mass was subsequently resected and images from the resultant specimen can be seen below.
Surface ulceration with granulation tissue can clearly been seen in the low-power image. Beneath this is a polypoid submucosal mass that consists of bland spindled cells embedded in a loose connective tissue matrix. There are rare mitotic figures but no atypical forms are identified. The stroma contains numerous small, predominantly thin-walled blood vessels. There is a background mixed inflammatory infiltrate with prominent eosinophils.
What is your diagnosis?
This is an example of an inflammatory fibroid polyp, a lesion that can occur anywhere in the gastrointestinal tract but most commonly involves the antral region of the stomach. The pathogenesis of these lesions is unclear and it has been suggested that it is a reactive process in response to localized trauma. Neurogenic differentiation was once suggested but has subsequently been disproven. The spindled cells typically show fibroblastic or myofibroblastic features The lesion is typically centered in the submucosa but larger examples (lesions up to 9 cm have been described) may extend into the muscularis propria.
Immunohistochemically the spindled stromal cells are reactive for vimentin, and most will stain for CD34. Variable positivity for smooth muscle actin, desmin, and factor XIIIA has also been noted.
Tuesday, July 24, 2012
A 60-year-old woman presented with a bulky mass involving her lower nasal cavity and one maxillary sinus. A biopsy specimen was obtained yielding the tissue seen below.
Microscopically, the tumor consisted of sheets of small blue cells with nuclei showing variably prominent eosinophilic nucleoli. Most nuclei were round, although some showed lobation. The cells had scant often somewhat eccentrically located cytoplasm. Apoptotic cells were common and mitotic figures were easily found. There was no light microscopic evidence of squamous or glandular differentiation. Immunohistochemical stains for epithelial and hematolymphoid markers were completely negative.
What is your diagnosis?
The neoplastic cells were strongly positive for malignant melanoma markers including S100 protein, HMB-45, and Melan-A, seen above. This is an example of a sinonasal malignant melanoma displaying small cell morphologic features. In this, and many other anatomic locations, malignant melanoma must always be considered in the differential diagnosis of virtually any light microscopically undifferentiated neoplasm, as mucosal melanomas may have a wide variety of histologic appearances including small cell, sarcomatoid, spindle cell, rhabdoid, epithelioid, and pleomorphic patterns. The differential diagnosis is accordingly large and includes olfactory neuroblastoma, a variety of undifferentiated carcinomas, lymphomas, PNET, synovial sarcoma, embryonal rhabdomyosarcoma, undifferentiated pleomorphic sarcoma, and others!
Melanin pigment, if present, can be helpful in the diagnosis but is often absent, particularly from the small cell forms of the tumor. Other helpful, but often absent features include junctional change and a nested or theque-like growth pattern. The presence of large, eosinophilic nucleoli, as in this case, is also suggestive but, of course, may be seen in many other forms of neoplasia.
Interestingly, most sinonasal melanomas, and most mucosal melanomas in general lack the BRAF gene mutation commonly seen in cutaneous melanomas. Conversely, they often show gains in chromosomes 1q and 6p, changes not typical of cutaneous disease. A subset of mucosal melanomas will express CD117 (KIT), and this correlates with mutations in the kit gene. This suggests a role for tyrosine kinase inhibitor therapy in these patients.
Sinonasal malignant melanoma generally occurs in older individuals, and has an aggressive biologic behavior with a 5-year survival of less than 30%. Conventional (non-gene-targeted) chemotherapy and radiation therapy have shown little effect on these tumors.
Thursday, July 05, 2012
A 7-year-old girl presented with a bulky mass involving her lower nasal cavity with extension into the maxillary sinuses. An enlarged cervical lymph node was also noted. Both the nasal mass and the lymph node were biopsied.
The resultant specimens, shown microscopically above, consisted predominantly of sheets of undifferentiated malignant cells with little visible cytoplasm, and uniform enlarged nuclei with one or more medium-sized nucleoli. Mitotic figures were numerous. There were many apoptotic tumor cells, as well as foci of overt necrosis.
As seen above, in some areas, the undifferentiated cells appeared to undergo abrupt squamous differentiation with the production of sharply demarcated nests of rather bland squamous cells, devoid of mitotic activity.
Occasionally (above image), there were small cystic spaces (top of image) lined by bland squamous cells and surrounded by the undifferentiated cell population.
The undifferentiated tumor cells were negative for hematolymphoid and neuroendocrine markers. They were also negative for EBV and HPV viral RNA/DNA. A p63 stain showed strong positivity in the squamous nests and focal positivity in the undifferentiated cells. Cytokeratin and EMA stains were diffusely positive as well.
What is your diagnosis?
This is an example of a NUT Midline Carcinoma (NMC). The most common translocation involving the NUT gene is the t(15;19) (q13;p13.1). This translocation fuses the NUT gene on chromosome 15 to the BRD4 gene. In about one-third of cases, however, the NUT gene is fused to a different partner gene (NUT variant). Rare cases have found the gene to be fused to the BRD3 gene at 9q34.2.The BRD proteins are known to bind transcriptionally active chromatin and stimulate transcriptional activity. The function of the NUT protein is not known, although in is only constitutively expressed in early germ cells and within the brain (ciliary ganglion). It is believed that the fusion protein can be localized to the nucleus secondary to its fusion to the BRD4 and BRD3 proteins. Some recent data have suggested that the fusion proteins block epithelial differentiation as si-RNA-induced withdrawal of the fusion proteins leads to squamous differentiation and cell cycle arrest.
Histologically, these tumors typically are composed of sheets of undifferentiated cells. Large areas of coagulative necrosis may be present. The cells have little amphophilic or eosinophilic cytoplasm. Nuclei have irregular contours, although they tend to be somewhat uniform in size, with fine to vesicular chromatin and prominent nucleoli. Mitotic figures and apoptotic bodies are numerous. Squamous differentiation may be seen and does not necessarily predict the partner gene (see below). The differentiation may be subtle with slight streaming of the malignant cells or it can be very distinct, with sharply demarcated mature squamous cells and extracellular keratin formation. Some authors, including ourselves, have noted that the keratinization may appear abrupt, with sheets of immature cells juxtaposed to well-differentiated, mature squamous nests. The well-differentiated areas may even become cystic.
The differential diagnosis for NMCs includes other undifferentiated carcinomas and poorly differentiated (basaloid) squamous cell carcinomas. Within the upper aerodigestive tract, the differential diagnosis includes pediatric small blue cell tumors, melanoma, olfactory neuroblastoma, high-grade hematologic malignancies, endocrine carcinomas and other undifferentiated or poorly differentiated carcinomas including SNUC and EBV-driven lymphoepitheliomas that occur most frequently in the nasopharynx.
Identification of clear-cut epithelial differentiation rules out many of the tumors in the differential (although one should remember that some mesenchymal tumors can rarely show epithelial differentiation, e.g., primitive neuroectodermal tumor). For tumors without obvious differentiation seen histologically, immunohistcochemistry can be used. Epithelial differentiation can almost always be demonstrated with pankeratin immunococktail. Negative results for other antigens such as S100 protein, CD45, muscle antigens, neuroendocrine antigens, germ cell antigens, are also helpful as cytokeratin or EMA immunoreactivity can be seen, at least focally, with many of these tumors. Immunoreactivity for p63 is also helpful as it points to squamous differentiation and was seen in 4 of 5 NMCs that we tested. Reactivity makes some diagnoses, such as SNUC much less likely.
NMCs have not been shown to be associated with carcinoma-related viruses such as EBV and HPV. Thus, positive in-situ hybridization for these viruses excludes the diagnosis of NMC.
Friday, June 22, 2012
A 57-year-old, postmenopausal woman presented with spotty vaginal bleeding. An endometrial curettage was performed yielding the tissue seen at increasing magnifications below.
The lining of relatively normal appearing, weakly proliferative endometrial glands is abruptly transformed to a papillary and micropapillary proliferation of cuboidal cells with moderate pleomorphism, increased mitotic activity and scattered apoptotic bodies.
What is your diagnosis?
An immunohistochemical stain for p53 was performed and the results are shown below.
An immunohistochemical stain for Ki67 (not illustrated) showed increased labelling in the p53-positive regions.
This is an example of endometrial intraepithelial carcinoma (EIC), a lesion considered to be the in-situ precursor of uterine serous papillary carcinoma. EIC is most commonly encountered as a secondary finding in association with (invasive) uterine serous papillary adenocarcinoma. It is rare as a purely "in-situ" lesion, as in this case. EIC may also be a marker for serous papillary carcinoma outside of the endometrium, though this association is, again, not full defined. Because of this association, however, the presence of EIC should lead to meticulous staging and, as most women are postmenopausal, hysterectomy and salpingectomy seem fully appropriate. Not surprisingly, the presence of extrauterine disease is THE critical prognostic factor in patient with pure EIC. Some authors have noted that the distinction between EIC and uterine serous papillary caricnoma confined to the endometrium may occasionally be difficult and have suggested combining these two variants under the rubric of "minimal uterine serous carcinoma." In our experience, however, this distinction has almost always been straightforward.
The histologic features of EIC are reasonably well defined and the lesion is not overly difficult to recognize once encountered. Strong p53 staining (and increased Ki67) positivity may be helpful in problematic cases. Unfortunately, the designation of EIC leads to some potential confusion with so-called EIN or endometrial intraepithelial neoplasia, a distinctly different system of categorizing pre-malignant non-serous endometrial proliferations analogous to the complex endometrial hyperplasia to atypical hyperplasia terminology that is still preferred by your blogger.
Wednesday, June 13, 2012
A 16-year-old boy presented with recurrent, often severe epistaxis. Clinical examination showed a nasopharyngeal polypoid mass. Radiographic studies documented the nasopharyngeal mass with extension into the posterior nasal cavity and bowing of the wall of the maxillary antrum. A presumptive diagnosis was made and the patient underwent operative resection. The resultant specimen is shown below.
What is your diagnosis?
The specimen consists of collagenized stroma in which the collagen fibers are haphazardly oriented and separated by fibroblastic cells with small, distinctly stellate nuclei. Stromal mitotic figures are not identified and there is no evidence of necrosis. Embedded within this stroma are prominent, irregularly shaped vascular channels lined by variably obvious endothelial cells. The latter can be more readily identified with endothelial markers such as Factor VIII-related antigen or CD31.
Interestingly, immunohistochemistry for smooth muscle actin (SMA) shows that many of the irregular vascular spaces have a pericytic cuff that is not apparent on H&E stains (see below).
Finally, immunohistochemistry for androgen receptor shows strong nuclear positivity in the stromal cells but not the endothelium or pericytes (see below).
This is a nice example of a so-called juvenile nasopharyngeal angiofibroma, a tumor that is sharply localized in location (nasopharynx), age (adolescence), and sex (male).
Purported examples in females have invariably represented other vascular lesions such as lobular capillary hemangioma. Well-documented examples have occasionally been described in older males. Resection (or biopsy) of these lesions is often associated with voluminous blood loss and, thus, should only be done where adequate supplies of blood/fluid replacement and high-volume vascular access are immediately available.
Although resections of these lesion can be difficult (and bloody!), complete removal is typically curative and the lesions only occasionally recur. The androgen receptor positivity of the stromal cells is of interest in light of the tumor's strong predilection for young males with peak testosterone levels. Studies have shown decrease in tumor vascularity when the patients were treated preoperatively with short courses of estrogen therapy and, conversely, testosterone has been shown to increase tumor size and vascularity. At the current time, most patients get pre-operative embolization (rather than estrogens) to help minimize blood loss.
Thursday, May 24, 2012
A 59-year-old woman presented with a polypoid gingival mass that was excised. The resection specimen is shown below at progressively higher magnifications.
The polyp consists of aggregates of multinucleated giant cells associated with foci of stromal hemorrhage and hemosiderin. The surrounding tissue is variably fibrotic and contains a mixed inflammatory infiltrate.
What is your diagnosis?
This is a straightforward example (they can't all be trick cases!) of a peripheral giant cell reparative granuloma or simply peripheral "giant cell granuloma" (GCG). This lesion is the soft tissue counterpart of giant cell granuloma of bone. For years it has been known that intraosseous GCG was closely related to aneurysmal bone cysts (ABC), and in fact, the so-called solid variant of aneurysmal bone cyst is histologically identical to GCG, although there is a spectrum of radiographic changes associated with the two entities.
Recently it has been shown that primary aneurysmal bone cysts often show a t(16;17) translocation with activation of ubiquitin protease USP6 on chromosome 17 in the spindled stromal cells but not the endothelium. This translocation has also been shown in GCG, further supporting their close relationship. Interestingly, the t(16;17) translocation is not seen in secondary aneurysmal bone cysts associated with other, presumably inciting, lesions. This strongly suggests that secondary aneurysmal bone cysts are fundamentally different from the ABC - solid ABC - GCG spectrum of lesions.
It is unclear at present if the t(16,17) translocation is seen in peripheral (soft tissue) GCG. Given that these are often traumatically induced, ie. reactive lesions, my guess would be that they do not carry this translocation, but, of course, we'll have to wait and see. It should be noted that other translocations have been described in ABC/GCG, but the t(16;17) seems to be the most common.
To return to the current case, these lesions are typically cured by simple excision and rarely recur.
Wednesday, May 16, 2012
A 66-year-old woman presented with an enlarged cervical lymph node and underwent an excisional biopsy. The resultant specimen was submitted to us for consultation. An H&E section and cytokeratin stain (AE1/AE3) from the outside pathologist are shown below.
The neoplastic cells exhibited marked nuclear variation, frequent mitotic figures and moderate amounts of eosinophilic cytoplasm. The nuclei were often eccentrically located, creating a rhabdoid or plasmacytoid appearance. As part of our workup a CD138 was performed and shown below:
In conjunction with the negative cytokeratin from the outside hospital, this raised the possibility of a plasmablastic lymphoma. However, the partial involvement of the node and the loosely cohesive nature of the cells was unusual for that entity. In addition the patient had no history of immunocompromise and, perhaps most importantly, in-situ hybridization for EBER was negative except for rare "passenger" lymphocytes. At this point a cytokeratin stain was repeated using our broad spectrum cytokeratin cocktail of antibodies. The result is shown below.
Based on the strong cytokeratin positivity shown above, the node was considered to contain metastatic carcinoma and a phone call indicated that the patient did have a history of invasive breast carcinoma. Additional immunohistochemical studies showed that the tumor was positive for ER, PR, mammaglobin and BRST-2. Accordingly, it was diagnosed as metastatic carcinoma of breast origin.
I illustrated this case to make two at least moderately important points. First, although AE1/AE3 is commonly seen in consultation material and often labelled by the submitting pathologist as "broad spectrum" cytokeratin, it misses many keratin subtypes. We use a cocktail of multiple cytokeratin antibodies including AE1, AE3, Cam5.2, MAK6, MNF116, OSCAR, and 34BH11. The result is highly sensitive and negativity can be interpreted as truly cytokeratin negative!
Second, although CD138 is a highly sensitive marker for plasmacytic/blastic differentiation, it is by no means specific. Among the carcinomas, CD138 is often strongly positive in cutaneous squamous cell carcinoma and basal cell carcinoma, colorectal adenocarcinoma, cholangiocarcinoma, transitional cell carcinoma, endometrial adenocarcinoma, ductal and lobular breast carcinoma, hepatocellular carcinoma, renal cell carcinoma, and lung adenocarcinoma. Many other carcinomas and sarcomas will less commonly be positive.
Thursday, May 03, 2012
A 72-year-old woman with abnormal uterine bleeding underwent an endometrial biopsy. The resultant specimen is shown below.
What is your diagnosis?
The specimen consists of a complex proliferation of epithelial cells forming sheets, glands and papillae. The cells vary from having clear vacuolated to eosinophilic cytoplasm. Cells lining the papillae have enlarged nuclei protruding into the lumina in a "hobnail" configuration.
This is a straightforward example of a uterine clear cell carcinoma. These "type II" uterine adenocarcinomas, like their close counterparts, serous papillary carcinomas, tend to occur in older postmenopausal women and do no typically arise in a setting of endometrial hyperplasia. There is no association with estrogen or progesterone therapy. These tumors are histologically identical to ovarian clear cell carcinomas. They are also identical in microscopic appearance to the DES-associated clear cell carcinomas of the cervicovaginal region, but there is no known association with DES for the uterine tumors. In the past, when DES-associated tumors were more common, distinction of uterine v. cervicovaginal origin could be problematic, but today DES-associated adenocarcinomas are virtually non-existent as the drug was discontinued many years ago.
Uterine clear cell carcinomas make occur in pure form or be associated with papillary serous carcinoma. Distinction between the two is occasionally problematic, but not of great clinical importance. When both tumors have a papillary pattern, the distinction seems to be most difficult as clear cells are less prominent on the papillae. The papillae of clear cell carcinoma typically have hyalinized, densely eosinophilic cores.
Uterine clear cell carcinomas are typically positive for ER and PR, but are less often p53 positive, as compared to serous carcinomas. Clear cell carcinomas should be distinguished from uncommon endometrial metaplasias, including clear cell, hobnail, and papillary syncytial metaplasia. This is usually (but not always) straightforward on the basis of the marked nuclear pleomorphism and mitotic activity in clear cell carcinoma.
It should be noted that the FIGO grading system for uterine endometrioid adenocarcinomas does not apply to clear cell or serous carcinomas. These tumors are, by definition, high-grade neoplasms. Their diagnosis in a curettage specimen will typically lead to retroperitoneal lymph node sampling or dissection, regardless of the depth of myometrial invasion.
Friday, April 20, 2012
A 65-year-old man underwent surveillance endoscopy for long-standing Barrett's esophagus. Prior biopsies had been said to show Barrett's epithelium (Goblet cell metaplasia) with mild dysplasia or indefinite for dysplasia changes. The images below depict the current findings at progressively higher magnifications.
At low power the mucosa closely resembles cardia-type gastric mucosa with foveolar cells and no evidence of goblet-cell metaplasia. However, the surface has a peculiar villous pattern that is not typical of normal gastric cardia and our GI colleagues assured us that this clearly came from the esophagus. Higher magnification shows some patchy stratification of the foveolar epithelial cells but no evidence of increased mitotic activity or appreciable nuclear atypia. Interestingly, Ki67 immunostaining, shown below showed considerable DNA synthesis well out onto the surface villi, a clearly abnormal pattern suggesting disorderly maturation.
What is your diagnosis?
This is an example of villous foveolar metaplasia with questionable low-grade dysplastic change. The relationship between this lesion and the much more common intestinal (Barrett's) metaplasia of the esophagus is not entirely clear. Some view this change as a precursor of Barrett's metaplasia and others view it as a separate, distinct pathway. Equally unclear is the risk this form of metaplasia has for the subsequent development of carcinoma. This is complicated by the fact that in many (?most) patients foveolar and Barrett's metaplasia co-exist in the same esophageal biopsies. To further complicate things, the presence and grading of dysplasia in foveolar metaplasia is also controversial; considerably more so than even in Barrett's metaplasia. This is in large part due to the fact that our recognition of this change as a distinct form of esophageal metaplasia is more recent and less well documented that the more common Barrett's change.
Given that the current case showed deranged DNA synthesis in the villous tips, as evidenced by the aberrant Ki67 staining, we interpreted this as at least indefinite for dysplasia, inspite of the fact that nuclear changes (other than focal pseudostratifiction) are minimal or completely absent. This patient had four biopsies during this endoscopy and all four showed virtually identical findings and were devoid of Barrett's metaplasia. Prior biopsy specimens DID show Barrett's metaplasia with, on several occasions, indefinite or mild dysplasia.
Our understanding of foveolar v. intestinal metaplasia in the development of esophageal adenocarcinomas will continue to develop and guidelines for recognizing and classifying dysplasia will continue to mature. Foveolar metaplasia and subsequent dysplasia has also been documented as a route to the development of gastric carcinoma, distinct from the intestinal metaplasia pathway.
Monday, April 09, 2012
A 25-year-old woman in her second trimester of pregnancy was noted on pelvic examination to have a 1.5 cm. polypoid mass protruding from her endocervix. An excisional biopsy was performed and yielded the specimen shown below in progressively higher magnifications.
The lesion consists of an extremely complex proliferation of mucin-producing glandular epithelium associated with a variable infiltrate of acute inflammatory cells. There is mild nuclear variability.
What is your diagnosis?
This is an example of microglandular hyperplasia, also referred to as microglandular change. This lesion used to be MUCH more common when oral contraceptives (OCPs) contained higher hormone levels, particularly progestogens. With the advent of newer versions of OCPs, the incidence of this lesion has dropped dramatically, and many pathology residents have not encountered it. It is, not surprisingly, still seen in the high progesterone environment of normal pregnancy.
The glandular complexity of this lesion can easily lead to confusion with adenocarcinoma and before its more widespread recognition, it was often overdiagnosed as such. However, awareness of this process and a high index of suspicion in the proper clinical setting should lead to the correct interpretation. There is certainly a high level of glandular complexity, and there may be scattered atypical cells, but the mitotic rate is very low and atypical mitotic figures are not present. Many of the spaces contain acute inflammatory cells with the epithelium "splayed out" around the spaces, rather than forming true glandular lumina. In other areas lumen formation may be prominent. It should be noted, however, that some cases of microglandular hyperplasia can form alarmingly cellular nests with enough nuclear atypia to be easily confused with carcinoma. Moreover, this lesion can also rarely be encountered in postmenopausal women.
It has been said that in longstanding examples of
microglandular hyperplasia, basal or myoepithelial cells may be demonstrated around the glandular nests with p63 immunohistochemistry. These cells were lacking in this case. More importantly, microglandular hyperplasia lacks staining for CEA, which is typically positive in endocervical adenocarcinoma.
Monday, March 26, 2012
A 43-year-old woman underwent a hysterectomy for multiple leiomyomas. One of the leiomyomas contained a 3 mm. nodule composed of uniform epithelioid cells with clear cytoplasm. The clear cell nodule was sharply demarcated from the surrounding leiomyoma. Mitotic figures were not identified. There was no evidence of necrosis. The 3 mm nodule is shown at increasing magnifications below.
What is your diagnosis?
The clear cell nodule showed focal staining for smooth muscle actin and was more diffusely positive for HMB-45. The surrounding leiomyoma was completely HMB-45 negative.
This is an example of an incidental PEComa arising in a uterine leiomyoma. PEComas (Perivascular Epitheloid Cell tumors) are being described at an ever-increasing number of anatomic sites. Several cases ago in this series, I presented a malignant PEComa of the lung with CNS metastases. In fact, the prototypical PEComa was the so-called "Sugar Tumor" of the lung, so named because the clear cytoplasm was due to abundant glycogen. A number of years ago we demonstrated that these tumors expressed HMB-45, as well as some other melanocytic markers but were typically S100 protein negative. Expression of smooth muscle markers was also demonstrated and some have referred to these tumors as myo-melanocytic neoplasms. The term "PEComa" is currently the favored designation.
The relationship between uterine PEComas and leiomyomas is far from clear. Several series have noted that these tumors are closely related to and perhaps part of a spectrum of epithelioid smooth muscle neoplasms involving the uterus. The current case is unusual in that the incidental PEComa was quite distinct from the surrounding leiomyoma, which was a typical benign spindle cell lesion with no evidence of an epithelioid component. In addition, many of the previous cases exhibited tongues of invasion into the surrounding myometrium in a pattern reminiscent of endometrial stromal sarcoma. The current case lacked any evidence of such invasion.
Several studies have noted an association between uterine PEComa and lymphangioleiomyomatosis (LAM), as well as tuberous sclerosis.
Monday, February 27, 2012
A 55-year-old woman with a history of carcinoma of the bladder was treated with several courses of preoperative chemotherapy and then underwent a radical cystectomy and lymph node dissection. The bladder revealed a pure small cell carcinoma diffusely invading the detrusor muscle with little or no evidence of chemotherapy effect. Images from one of the lymph nodes are shown below.
What is your diagnosis?
Multiple lymph nodes showed obvious involvement by small cell carcinoma. In several nodes there was a component of mucin-producing enteric-type adenocarcinoma in addition to the small cell component. The adenocarcinoma was always closely associated with the small cell carcinoma and was never seen separate from it. The junction between the two components was always abrupt with no evidence of a gradual transition. My initial reaction was that the adenocarcinoma represented metastases from a second primary tumor elsewhere in the body. However, thorough examination of the bladder turned up no primary adenocarcinoma component, and history, physical exam and radiographic studies demonstrated no evidence of a separate malignancy.
Immunohistochemistry can be of little if any help in this situation since tumors typically reflect their phenotype immunohistochemically. Thus, tumors that light microscopically intestinal adenocarcinoma will typically express CDX2 and other intestinal markers, clearly expressing differentiation, rather than "origin."
Given the above findings, particularly the invariably close association of the two elements and the lack of adenocarcinoma elsewhere in the patient, it seems most likely that the associated adenocarcinoma represents multi-focal divergent differentiation of the small cell carcinoma. Small cell carcinomas at other anatomic sites, particularly extra-pulmonary sites are well known to be associated with other carcinoma subtypes. In the salivary glands, for example, small cell carcinoma is almost always accompanied by a component of adenocarcinoma and this is often the case in pulmonary small cell carcinomas. In such locations, however, where the divergent patterns are present in the primary site, it is usually assumed that the better differentiated component came first, and subsequently gave rise to the small cell carcinoma. In the current case, assuming the two components are clonally related, it seems possible that the adenocarcinoma component came second, which is admittedly a bit hard to swallow since it implies a high-grade poorly differentiated tumor mutating into a better differentiated one. Alternately, and perhaps more realistically, there may have been a small component of adenocarcinoma present from the beginning and over run by the small cell component. The latter was more sensitive to chemotherapy and removal of a large portion of the small cell carcinoma made the adenocarcinoma component more obvious.
Whatever the mechanism, the heterogeneity of tumors, particularly high-grade tumors, is well understood by pathologists but often ignored by oncologists. They wish to tailor their chemotherapy to a single tumor type and have difficulty coping with divergent differentiation.
Thursday, February 16, 2012
A 46-year-old woman presented with a 3 cm. thyroid mass and underwent a lobectomy. Microscopic images from the resultant resection specimen are shown below.
This was a partially encapsulated, well-demarcated nodule composed of papillary and glandular structures. These were lined by pseudostratified columnar cells with mildly pleomorphic nuclei having punctate chromatin and one or more nucleoli. Mitotic figures were easily found. Eosinophilic colloid within the gland-like spaces was only focally present. There was no evidence of necrosis. Elsewhere in the thyroid lobe was a separate smaller nodule of papillary carcinoma, follicular variant type.
What is your diagnosis?
This is an example of columnar cell carcinoma of the thyroid gland. Although most pathologists consider this to be a rare variant of papillary carcinoma, a few consider it to be a completely distinct carcinoma subtype. The presence of more typical papillary carcinoma (follicular variant) elsewhere in the same lobe in the current case is at least circumstantial evidence for an association. The columnar cell variant differs from the more common tall cell variant of papillary carcinoma by having marked nuclear stratification and lacking the characteristic nuclear features of papillary carcinoma (vesicular, grooved, overlapping nuclei with pseudoinclusions).
Although initial descriptions of the columnar cell variant of papillary carcinoma suggested an aggressive behavior, this appears to have been mainly due to the presence of high stage tumors in these studies. More recent studies
suggest that small, encapsulated tumors have a prognosis similar to that of conventional papillary carcinoma. Given that the current lesion is small, confined to thyroid lobe and at least partially encapsulated, we would expect an outcome similar to that of conventional papillary carcinoma.
Tuesday, January 31, 2012
An 82-year-old woman with a history of breast carcinoma presented with back pain. Radiographs showed a lytic lesion in T6. A biopsy was performed. The resultant specimen is shown below.
The marrow spaces are filled with large epithelioid cells without any light microscopic evidence of glandular or squamous differentiation. A presumptive diagnosis of metastatic carcinoma of the breast was made, but cytokeratin stains were completely negative and the case was referred to us for consultation.
The images below show the tumor cells at progressively higher magnification.
What is your diagnosis?
We repeated the cytokeratin stains, as well as adding GCDFP15 and mammaglobin. All were negative. CD34 and CD31 (shown below) were strongly positive and the case was interpreted as an epithelioid angiosarcoma.
As is often the case with consultation material, additional clinical information is not readily available, and it is unclear whether this is a primary bone tumor or a metastasis from an unknown primary.
Although the diagnosis of this case is difficult, and in fact impossible on H&E-stained sections alone, the "take home" message from this case is simple and straightforward. Angiosarcomas may be composed of distinctly "carcinoma-like" endothelial cells with little or no light microscopic evidence of endothelial differentiation. This possibility, along with others such as melanoma and anaplastic lymphoma, should always be considered whenever an apparently obvious carcinoma lacks appropriate reactivity for epithelial markers.
There is, of course, a reverse and potentially even more treacherous pitfall in that epithelioid angiosarcomas may also express cytokeratins. The current case, fortunately, did not. If it had, it might well have been interpreted as a carcinoma and never studied for endothelial differentiation.
Epithelioid angiosarcoma lies at the most malignant end of the some what loose spectrum of epithelioid vascular tumors which also includes benign epithelioid hemangiomas, and less aggressive but definitely malignant epithelioid hemangioendotheliomas. The former tumors demonstrate clear-cut vascular differentiation with well-formed lumina, and the latter tumors typically show more abortive but definite vasoformative features, often manifest as intracytoplasmic lumina associated with a characteristic chondromyxoid-like background stroma.
Friday, January 20, 2012
A 59-year-old woman presented with a 3 cm. lung tumor and an intracranial lesion. There was a history of a renal cell carcinoma that was resected in the distant past and recurred in the renal bed 15 years ago. The time interval between the nephrectomy and the bed recurrence is not clear. Neither the original renal tissue or the recurrence is available for review. Both the lung lesion and the intracranial lesion were excised and were histologically identical. The pulmonary lesion is illustrated below.
The tumor consisted of epithelioid cells with predominantly clear but focally eosinophilic cytoplasm. Some of the latter cells had a distinctly rhabdoid appearance. There was considerable nuclear pleomorphism and there were large areas of geographic necrosis. The next two images show the tumor at progressively higher magifications.
The mitotic rate was high and atypical mitotic figures were easily identified.
What is your diagnosis?
This case was seen in consultation. Although the initial pathologist considered that this most likely represented a metastasis from the prior (or a new) renal cell carcinoma, the tumor did not label for the usual renal cell markers including RCC, CD10. It also did not label for cytokeratin or S100 protein. Vimentin was strongly positive. We performed immunohistochemical stains for HMB-45, Mel-A and smooth muscle actin (SMA), and repeated the S100 protein and broad spectrum cytokeratin stains. The HMB-45 labelling is shown below.
As can be seen the tumor was strongly positive for HMB-45. S100 protein labelled only scattered cells, presumably dendritic cells. Mel-A showed focal positivity. SMA and cytokeratin were negative As is probably obvious at this point, this is an example of a malignant PEComa, a tumor composed of cells thought to show the features of so-called perivascular epithelioid cells. They have also been called "myomelanocytic cells." In the lung, these tumors are almost always benign and were referred to as "sugar tumors" for years because the abundant clear cytoplasm was known to be due to the presence of glycogen. Malignant forms, have however, been described and we have seen at least one other example.
Tumors containing PEC-like cells have been described in multiple organ sites. In the lung, these cells are also associated with the diffuse process, lymphangioleiomyomatosis. In the kidney, they are part of angiomyolipomas. In that location, one variant, the epithelioid angiomyolipoma, is known to have malignant potential. PEComas have also been described in the uterus, where they may be confused with smooth muscle tumors, as well as other anatomic sites.
It is unfortunate that, at least at the moment, we do not have the original renal tumor or its local recurrence to review. This may well represent a true renal cell carcinoma and hence an entirely different process, although it is interesting that the patient would have only been in her early 40's at the time of that diagnosis, relatively young for a renal cell carcinoma (but certainly not unheard of). It is tempting to wonder if this patient might have actually had a renal PEComa that, although initially low grade, over time converted to the high grade tumor currently present in her lung and brain. The long time course and the fact that lung is a relatively common site for malignant PEComas suggests that this is probably a separate process, but we may never know for sure.
Wednesday, January 04, 2012
A 52-year-old woman underwent a left thyroid lobectomy for a small nodule that was noted to be "cold" on thyroid scan. The resultant lobectomy specimen demonstrated an encapsulated 8 mm. nodule seen at low magnification below.
There is a cellular nodule that is demarcated from the adjacent thyroid tissue by a well-formed fibrous capsule. Several follicles with colloid are noted peripherally. The next three images show progressively higher magnifications of the cells within the nodule. These cells have granular to vesicular nuclei and moderate amounts of amphiphilic cytoplasm. Prominent vascularity is also noted.
What is your diagnosis?
To start 2012, this is a relatively straightforward example of a medullary (neuroendocrine) thyroid carcinoma seen in consultation. The presence of some follicle-like structures located peripherally caused some diagnostic concern. However, the tumor cells, including those surrounding the islands of colloid, stained strongly for CEA, chromogranin (shown below), CD56, and calcitonin. The neoplastic cells were negative for thyroglobulin, including the cells surrounding the foci of colloid, suggesting that this was entrapment of pre-existent colloid by the tumor cells.
There was no evidence of C-cell hyperplasia in the surrounding thyroid tissue and the patient had no other personal or family evidence of a genetic syndrome. Given the small size of this lesion, its encapsulated nature, and the absence of MEN2b, the patient should do extremely well. Medullary carcinomas of the thyroid associated with MEN2b are known to be more aggressive than sporadic cases, and often result in the death of the patient. The sporadic tumors follow a somewhat more variable but generally better clinical course. It should be noted that the presence of calcitonin immunohistochemically is helpful diagnostically but by no means specific for the diagnosis of medullary carcinoma. Many neuroendocrine carcinomas will express this antigen.
Monday, December 19, 2011
An 81-year-old woman presented with a sublingual, base of tongue mass and underwent an excisional biopsy of the lesion. The resultant specimen, seen below, consisted of nests and cords of epithelioid cells with clear to eosinophilic cytoplasm, embedded in a fibrotic to hyalinized stroma. The neoplastic cells showed mild nuclear pleomorphism and scattered mitotic figures. There was no evidence of glandular or squamous differentiation. Necrosis, vascular and perineural invasion were also absent. The tumor cells were negative for S100 protein and SMA, and were positive for p63 and cytokeratins.
What is your diagnosis?
This is an example of a hyalinizing clear cell carcinoma, a tumor with a striking predilection to involve the base of the tongue. It has been suggested, based on ultrastructural studies and p63 positivity, that these tumors have squamous features. It should be noted, however, that occasional mucin-producing cells may be present in this neoplasm. The tumors typically have an indolent course, although local control in this anatomic region can be difficult.
Hyalinizing clear cell carcinoma should be distinguished from other head and neck neoplasms, particularly myoepithelial predominant epithelial-myoepithelial carcinoma and mucoepidermoid carcinoma. Both typically have a more aggressive clinical course. The lack of staining for myoepithelial markers and S100 protein is helpful in excluding the former tumors. The lack of any glandular differentation and the uniform appearance of the neoplastic cells with the prototypical hyalinized stroma is helpful for distinction from the latter tumors.
Novel genetic translocations continue to be found at a rapid rate for an ever-growing list of neoplasms. Mucoepidermoid carcinomas have been noted to frequently display translocations involving the MECT1 and MAML2 genes in a high percentage of cases. Recently, it was noted in a single study (Genes, Chromosomes & Cancer 50:559,2011) that hyalinizing clear cell carcinomas display a characteristic EWSR1-ATF1 gene fusion. FISH on our case with a break-apart EWS probe showed that the EWS gene was translocated, presumably representing an EWSR1-ATF1 fusion, as well.
Friday, December 09, 2011
A 59-year-old man presented with small bowel obstruction and underwent a resection of an ileal mass. As seen below, the overlying mucosa is intact, but the submucosa and muscularis are infiltrated by sharply demarcated nests of cells with no associated stromal reaction.
What is your diagnosis?
At sequentially higher magnifications seen below, the nests of cells can be seen to better advantage. There is no evidence of squamous or glandular differentiation. The cells show mild to focally moderate pleomorphism. Mitotic figures and scattered apoptotic cells are easily noted.
Although the low power image is strikingly suggestive of a typical carcinoid tumor and the higher power images suggest a higher grade neuroendocrine carcinoma, the tumor is completely negative for neuroendocrine markers, including chromogranin, synaptophysin and CD56. It is strongly positive for cytokeratins and p63.
This is an example of a carcinoma, with (non-keratinizing) squamoid features that has grown and invaded in a nested pattern, strikingly mimicking a carcinoid tumor or moderately differentiated neuroendocrine carcinoma. At least in part, the "nested" pattern appears to be due to invasion and dilation of lymphovascular spaces.
The differential diagnosis should include a high-grade urothelial carcinoma and a non-keratinizing squamous carcinoma from any of a wide variety of anatomic sites. This was a case seen in consultation and the patient is currently undergoing evaluation for an occult primary.
This is admittedly a bit of a trick case, but it points out the need to avoid jumping to a diagnosis based solely on a low-power image, without confirmatory higher magnification and immunohistochemistry. Although some would argue that the features of a typical carcinoid tumor are so sterotyped as to not warrant immunohistochemical verification, I've seen enough mimics over the years to recommend confirmatory immunohistochemistry on even "obvious" cases.
The current case on higher magnification, is clearly far too mitotically active to be a typical carcinoid tumor (well-differentiated neuroendocrine carcinoma), but its features are virtually identical to those of a higher grade neuroendocrine neoplasm, or so-called "atypical carcinoid tumor," a term I've tried to avoid using for several decades. Failure to document lack of neuroendocrine differentiation may well have led to the erroneous conclusion that this was a primary neoplasm, leaving the true primary undetected.
Thursday, December 01, 2011
A 35-year-old man with a history of HIV presented with difficulty breathing. Bronchoscopy reviewed a submucosal mass and a biopsy was performed.
Microscopically the mass consisted of fusiform spindled cells with relatively uniform nuclei, and prominent eosinophilic cytoplasm. Mitotic figures were present but rare and there was no evidence of tumor necrosis.
What is your diagnosis?
As seen above the neoplastic cells were strongly positive for smooth muscle actin.
In-situ hybridization for Epstein-Barr virus encoded RNA (EBER) was strongly positive as seen above.
This is an example of an EBV-associated smooth muscle tumor (EBV-SMT). The number of reported EBV-SMT's arising in severely immunosuppressed patients (usually with AIDS) has steadily increasing since the first description in the mid-1990s. A 2011 review by Purgina et al (Pathology Research International 2011) identified 53 articles in the English language literature, including 64 reported cases of EBV-SMT. The majority of these reports involved patients who were young, severely immunosuppressed, and had multifocal tumors. The central nervous system was the most common site to be involved, but the tumors are noted to arise in very unusual sites for smooth muscle neoplasms, including the vocal cords and orbit, and they are frequently multicentric.
Histologically, tumors had smooth muscle features and were immunoreactive for muscle markers and almost invariably demonstrated the presence of EBV by in situ hybridization or PCR. Importantly, while mitotic figures and necrosis are used to separate leiomyoma from leiomyosarcoma, when applied to smooth muscle tumors NOT associated with EBV, these features do not correlate with clinical outcome in the patients with EBV-SMT's.
Treatment consists primarily of resection, and less often radiotherapy, chemotherapy and highly active antiretroviral therapy. Overall, EBV-SMTs appear to have variable aggressiveness and clinical outcome and may exhibit a more favorable prognosis compared to conventional leiomyosarcoma. Tumor-related deaths from EBV-SMT occurred in only 4 of 51 patients.
The list of EBV-associated neoplasms continues to grow!
This wasn't one of my personal consult cases. I "borrowed" it from somewhere and placed the photographs in my teaching image file. For the life of me, I can't remember where I got it, so I apologize for not giving appropriate credit!
Monday, November 21, 2011
An 82-year-old woman presented with a mass involving the upper outer quadrant of her right breast. A biopsy was performed, resulting in the specimen seen at low-power magnification below.
As seen above, this is a complex mixture of epithelial elements and stromal sclerosis with an infiltrative-appearing growth pattern. At higher magnification, the epithelial elements are variable in appearance, though cytologically they are uniformly bland. Mitotic figures were not seen.
Some of the epithelial elements (above) form small nests and cords of cells with scattered lumen formation.
In other areas, seen above, the epitheloid cells form larger solid, "squamoid" nests which occasionally exhibit overt keratinization (not shown).
What is your diagnosis?
This is an example of so-called low-grade adenosquamous carcinoma of the breast, a lesion that is quite similar (if not identical) to "syringomatous adenoma" of the nipple. The issue of the difference in terminology (benign v. malignant) for a similar or identical lesion in two different locations of the breast is one that ultimately needs resolution. Both lesions are noted for locally infiltrative growth and a tendency for local recurrence, with little or no capability for distant spread. The nipple lesions are undoubtedly encountered at a smaller, easier to resect stage, and this may account in large part for the nomenclatural difference.
Immunohistochemistry was performed for myoepithelial markers and the results for smooth muscle actin (SMA) are shown below.
As can be seen above, the nests of infiltrative epithelial cells are invariably surrounded by a well-formed SMA-positive layer of myoepithelial cells. This led the contributing pathologist to consider whether this was a radial scar or complex sclerosing lesion instead of a carcinoma.
In fact this is one of multiple exceptions to the rule that malignant lesions don't have an outer layer of myoepithelial cells. Rosen noted in his text that 80% of adenosquamous carcinomas were positive for at least two of the three commonly used myoepithelial markers (CD10, p63, SMA). I'll have a little more to say about the presence of myoepithelial cells in some malignant lesions, and their absence in some benign lesions, in a future "blog" on Pathology Network.
Tuesday, November 08, 2011
A 46-year-old woman with a distant history of malignant melanoma and no known history of asbestos exposure presented with a unilateral pleural effusion. A biopsy was performed and the resultant specimen is seen at low-power below.
There are two distinct populations of epithelioid cells. The cells on the left are smaller with smaller, more dense nuclei. Those on the right are clearly larger with larger "powdery" nuclei and small nucleoli. The latter population was the dominant one in this biopsy and these cells can be seen to better advantage in the image below.
What is your diagnosis?
Immunohistochemical stains were performed. The results of a cytokeratin cocktail are shown below.
Although somewhat difficult to appreciate at this magnification, the smaller epithelioid cells and some spindled sub-pleural fibroblasts reacted strongly for cytokeratins. The larger epithelioid cells were virtually all cytokeratin negative. The cytokeratin positive cells also reacted for a variety of mesothelial markers and these cells were interpreted as a component of reactive mesothelial hyperplasia. The larger epithelioid cells were negative for melanoma markers but reacted strongly for endothelial markers including CD34 (not shown), and the more specific marker CD31, shown below.
Although this tumor lacks virtually any obvious evidence of vasoformation, I believe that it represents and epithelioid vascular neoplasm. Because of it cytologic uniformity, low mitotic rate, and involvement of the pleura, it was classified as most likely representing an epithelioid hemangioendothelioma, rather than an overt angiosarcoma.
Like malignant melanomas, epithelioid hemangioendotheliomas are great pathologic mimickers at many sites throughout the body. In the lungs they were initially considered by Dail and Liebow to be intravascular epithelial neoplasms (IVBAT), in the liver they have been confused with cholangiocarcinomas, in the pleura as in this case they can quite easily be confused with malignant mesotheliomas.
Epitheloid hemangioendotheliomas occupy the middle position in a histologic spectrum of epithelioid vascular neoplasms. At the benign end of the spectrum are epithelioid hemangiomas (also known as histiocytoid hemangiomas, pseudopyogenic granulomas, and angiolymphoid hyperplasia with eosinophilia, as well as other terms). At the more malignant end are epithelioid angiosarcomas. All three of these tumors can be very diagnostically difficult. Like other epithelioid stromal tumors they may occasionally express cytokeratins which can add to the diagnostic confusion. The pleural lesions are especially treacherous and this diagnosis should always be considered.
The clinical course for these tumors is variable and dependent on the anatomic location. Unfortunately, many of those with pleural lesions pursue a course similar to that of mesothelioma.
Thursday, October 27, 2011
A 65-year-old man presented with signs and symptoms of oropharyngeal obstruction and an enlarging mass involving much of his right palatine tonsil, with probable extension into adjacent structures. A biopsy was performed and yielded the tissue seen below.
The neoplastic cells can be seen to better advantage in the higher power image seen below.
The tumor is composed of small blue cells with angulated, molded nuclei and no nucleoli. There is little or no visible cytoplasm. Mitotic figures, although somewhat difficult to see at this magnification, were very numerous. There was no evidence of glandular or squamous differentiation. Immunhistochemical stains showed punctate perinuclear positivity for cytokeratin, patchy staining for synaptophysin, and more diffuse staining for CD56
What is your diagnosis?
This is an example of a small cell (neuroendocrine) carcinoma involving the palatine tonsil. These tumors are rare in this location or elsewhere in the head and neck and most upper aerodigestive tract neuroendocrine-related neoplasms are either olfactory neuroblastomas arising high in the nasal cavity or sinonasal undifferentiated carcinomas (SNUC's) composed of considerably larger cells and showing some neuroendocrine features by immmunohistochemistry. As seen below, this tumor exhibited strong cytoplasmic staining for p16.
In situ hybridization for high-risk HPV subtypes showed scattered unequivocally positive cells as seen below.
It is not surprising that, as in the uterine cervix, HPV occasionally may be associated with adenocarcinomas or neuroendocrine carcinomas, instead of its much more common positivity in squamous cell carcinomas.
There is emerging recognition of these HPV-variant tumors in the oropharynx and for the small cell carcinomas there are important distinctions from the more common HPV-related squamous cell neoplasms. First, many small cell carcinomas are strongly p16 positivity and this reactivity CANNOT be used as a highly specific surrogate marker for HPV, as it can with squamous cell neoplasms. Second, although HPV-related squamous cell carcinomas of the oropharynx often present at high stage and have a paradoxically better prognosis than HPV-negative squamous carcinomas, there is no evidence, at least so far, that HPV-positivity has any affect on the highly aggressive nature of small cell carcinoma.
As in the lower respiratory tract, small cell carcinomas in the upper aerodigestive tract retain a strong association with smoking, regardless of HPV status.
Monday, October 17, 2011
A 33-year-old male smoker was noted to have bilateral reticulonodular pulmonary infiltrates on a routine chest radiograph. An open lung biopsy was performed and a low-power image is shown below.
Microscopically, the biopsy specimen showed multiple interstitial nodules that, at their periphery, extend into and expand alveolar septa.
As seen above, the nodules are composed of aggregates of histiocyte-like cells and clusters of eosinophils.
At still higher magnification, seen above, the larger cells have abundant. somewhat granular or "foamy" eosinophilic cytoplasm with vesicular nuclei having distinctly folded or convoluted nuclear membranes.
What is your diagnosis?
The larger histiocyte-like cells were strongly S100 positive (shown above).
The same cells were also positive for Langerin (above) and for CD1a (not shown).
This is a nice example of pulmonary eosinophilic granuloma (EG), also known as Langerhan's cell histocytosis and several other equivalent designations. In the lung, this process is almost exclusively seen in smokers (current or former). Many patients are asymptomatic in spite of the often prominent radiographic findings. Others may experience dyspnea, hemoptysis, pleuritic chest pain, or spontaneous pneumothorax.
In extra-pulmonary sites, EG has been shown to almost always be a clonal process, whereas in the lung clonality via HUMARA assay is seen in about half of cases. Whether clonality indicates neoplasia is a subject of considerable debate.
In the lung, the prognosis of EG is generally excellent. Many patients will undergo apparent spontaneous regression of their disease without treatment. Others may have persistent radiographic findings. Steroids and chemotherapy have been applied to symptomatic cases with good results.
Thursday, October 06, 2011
A 69-year-old woman with a history of colon cancer presented to her local physician with progressive neck pain. Metastatic disease was suspected and radiographs showed a lytic lesion in the left pedicle of the C3 vertebra.
An open biopsy was performed.
Microscopically, there were large fragments of acellular fibrocartilage consistent with disc material. Embedded within these fragments were aggregates of cells with small, rather uniform nuclei and coarsely vacuolated cytoplasm.
What is your diagnosis?
The above higher magnification image better documents the bland cell population with coarsely vacolated cytoplasm. Clearly, this is not the appearance of metastatic colonic adenocarcinoma.
As seen above, the vacuolated cells were strongly positive for cytokeratin.
As also seen above the vacuolated cells showed both nuclear and cytoplasmic positivity for S100 protein.
This is an example of a "chordoma wanna-be." There are multiple studies dealing with this family of small notochordal-related tumors under a variety of names including, "ecchordosis physaliphora," "notochordal remnants, "giant vertebral notochordal rests," and "benign notochordal cell tumor." Notochordal remmants, by definition, should be incidental findings with no associated radiographic abnormality. The other entities are typically associated with small sclerotic or, less commonly, lytic lesions without extraosseous extension. Benign notochordal cell tumors have been noted to be associated with chordomas and it has been suggested that chordomas may arise from these precursor lesions.
In the current case, the biopsy material clearly contains disc-like fibrocartilage, but the radiographs show apparent extension of the process into the adjacent osseous pedicle, though there is no extraosseous extension. This finding, coupled with the history of progressive pain led us to interpret this as simply a "notochordal cell tumor," eliminating "benign" from the diagnosis. Indeed, others have noted that in individual cases, it may not be possible to clearly distinguish "benign notochordal cell tumor" from a small chordoma. Accordingly, we believe that this patient should be followed for possible recurrent or progressive disease.
Monday, September 26, 2011
A 75-year-old man presented with an enlarging cervical lymph node. An excisional biopsy was performed and yielded the specimen seen at low-power magnification below.
At low magnification, haphazard cords and bundles of epithelioid to spindled cells with vesicular nuclei infiltate the lymph node parenchyma.
At higher magnification, seen above, cords of more spindled cells with vesicular nuclei are seen on the right side of the image. Other areas showed more prominent nuclear pleomorphism.
What is your diagnosis?
Immunohistochemical stains are, not surprisingly, necessary to arrive at a definitive diagnosis of this case.
The figure above shows strong S100 positivity in the neoplastic cells.
The figure above shows strong positivity for fascin in the neoplastic cells. The tumor cells were negative for epithelial markers, HMB-45, Melan-A, CD56, CD21, CD23. This is an example of an interdigitating dendritic cell tumor.
At least once a month we consider this diagnosis in a consultation case, typically representing a cervical lymph node biopsy. In many instances the suspicion turns out to be correct. Dendritic cell tumors are relatively new in the lexicon of surgical pathology diagnoses and should be considered whenever unusual spindle cell proliferations are encountered associated with lymphoid tissue.
Proliferations of dendritic cells are currently divided into follicular dendritic cell tumors (FDCT), interdigitating dendritic cell tumors (IDCT), and Langerhans cell tumors (histiocytosis). The last is clinicopathologically distinct from the remaining two.
FDCT’s tend to occur in young to middle-aged adults usually as solitary enlarged lymph nodes in the neck or axilla. Extranodal sites, including the tonsils, have also been described. Males and females appear to be equally affected. About half of patients with FDCT develop metastatic disease but, interestingly, deaths from disease are extremely rare.
IDCT’s like the current case tend to occur in slightly older adults, but the presenting features are otherwise identical to those of patients with FDCT. Approximately 40% of patients develop metastases and occasional patients have rapidly progressed and died from disease. One study noted CD68 positivity in patients with progressive disease. The apparently worse prognosis of patients with IDCT, as compared to FDCT supports the distinction of these entities.
FDCT’s typically consist of spindled cells with indistinct cell borders, ovoid to elongated, vesicular nuclei and variably prominent nucleoli. These cells are intermixed with lymphocytes, plasma cells, neutrophils and eosinophils. Nuclear pleomorphism is usually minimal, but may occasionally be marked. The lesion is typically centered in the cortex of a lymph node and shows a whorling, fascicular, or storiform growth pattern. In one study of 12 cases, mitotic rate averaged 11 mitotic figures per 10 HPF. At least focal necrosis is noted in about 1/3 of cases.
IDCT’s typically display more nuclear pleomorphism and are composed of more polygonal cells with more abundant eosinophilic cytoplasm, when compared to FDCT’s. Mitotic activity and frequency of necrosis are similar to FDCT.
Immunohistochemically, FDCT’s are typically recognized by their positivity for CD21 and CD23. More recently, they have also been shown to be strongly reactive for clusterin. IDCT’s are typically negative for CD21 and CD23, but show positivity for S100 protein and for fascin (which is also present in FDCT’s). The often strong positivity of IDCT’s for S100 protein may easily lead to confusion with a mucosal malignant melanoma, a tumor noted for a wide variety of morphologic appearances including a spindle cell pattern.
Because of the marked difference in prognosis, this distinction should be made based on other more specific immunohistochemical markers for melanoma (Melan-A, HMB-45, tyrosinase, etc.).
Friday, September 16, 2011
A 41-year-old woman with a long history of Sjogren's syndrome presented with a slowly enlarging parotid mass. A superficial parotidectomy was performed. The resultant specimen is shown below. The normal salivary gland tissue has been virtually completely replaced by a mononuclear infiltrate surrounding cords, strands and nests of somewhat basaloid epithelial cells.
What is your diagnosis?
This consultation case from this week is a very nice example of a marginal zone lymphoma (MALToma) of the parotid gland. The presence of large aggregates of monocytoid B-cells surrounding nests of basaloid epithelial cells (so-called lymphoepithelial lesions or islands) is quite characteristic of this entity.
The two higher power images above show the nests of epithelial cells being infiltrated by the neoplastic monocytoid cells with their prominent pale cytoplasm. These cells clearly contrast with the non-neoplastic lymphocytes at the periphery of the monocytoid nests. Recognition of these aggregates of monocytoid cells surrounding lymphoepithelial islands usually allows for ready recognition of this entity.
The neoplastic cells of marginal zone lymphoma will typically express monotypic IgM or IgG. They are typically positive for BCL-2 and lack expression of CD5, CD10, CD25, and cyclin D1. Genotypic studies will show clonal immunoglobin gene rearrangements but no specific abnormalities of oncogenes.
As is often the case, the current patient had a protracted history of Sjogren's syndrome. Presumably, the longstanding chronic salivary inflammation is the soil in which the MALToma develops. This is analogous to the association of gastric MALTomas with longstanding Helicobacter pylorii infection and associated inflammation.
The prognosis for these patients is extremely good, although they have been noted to occasionally transform into high-grade lymphomas.
Wednesday, September 07, 2011
A 70-year-old man had a several year history of a relatively stable right parotid mass. A month prior to admission, the mass was noted to rapidly increase in size. A superficial parotidectomy was performed, yielding the tissue shown below.
Note the difference in appearance between the majority of the image, and the upper right-hand portion of the field.
What is your diagnosis?
(admittedly a guess at this magnification!)
A portion of his patient's mass had the classical appearance of a Warthin's tumor as seen on the left-sided two-thirds of the above image. However, something different is happening in the upper right portion of the screen. This component accounted for the majority of the mass and is seen at higher power below.
The higher power image shown above shows a distinctly different appearing high-grade carcinoma with no obvious evidence of glandular or squamous differentiation, and no evidence of the oncocytic features typical of the Warthins tumor epithelium. The mitotic rate in this component was very high, and there were numerous cells with grotesquely enlarged nuclei.
This is a rare example of a Warthin's tumor gone bad, a high-grade carcinoma arising in a pre-existent Warthin's tumor. The term, "dedifferentiated Warthins tumor" may also be applied to this lesion.
Follow-up is short for this case, but the high-grade component was surrounded by a thin rim of typical Warthins tumor. This may be analogous to a carcinoma arising in but still confined to a pre-existent mixed tumor, a lesion that has been shown to have a very good prognosis. Given the high-grade nature of the current lesion, close follow-up is certainly recommended.
Wednesday, August 24, 2011
A newborn infant was noted to have a 3 cm. cystic mass attached to the midline of his external nose. The lesion was excised and the image above shows the resultant microscopic features.
What is your diagnosis?
The inside of the cyst was lined with prominent capillaries surrounded by small, almost inapparent cuboidal cells. The latter somewhat resemble choroid plexus on the surface of the brain. Beneath this vascular surface was densely eosinophilic fibrillar material with a distinctly glial appearance.
This is an example of a nasal encephalocele, aka "glial heterotopia." These are non-neoplastic malformations that represent eversions of glial tissue outside of the cranial cavity during embryonic development. There may or may not be a residual cranial defect. Some suggest referring to lesions WITH a cranial defect as "nasal encephaloceles" and those without a cranial defect as "glial heterotopias." They are clearly very closely related lesions. Importantly, the term "nasal glioma" should be avoided for these lesions as this erroneously suggests that these are true neoplasms.
Although there have been isolated reports of glial neoplasms arising in nasal encephaloceles these are vanishingly rare, reportable cases. The glial tissue may, however, show marked gemistocytic change potentially leading to confusion with a gemistocytic astrocytoma. Neurons and axons may also be present. The differential diagnosis is primarily with a mature teratoma. Radiologic correlation may be extremely helpful in arriving at the correct diagnosis.
Simple excision is curative and cosmetic defects are typically minimal.
Monday, August 15, 2011
A 55-year-old woman underwent routine surveillance colonoscopy. Several diminutive polyps were removed and submitted together in a single jar for histologic examination. All showed features of hyperplastic polyps, as shown below.
One of the hyperplastic polyps had, in addition, the microscopic features seen below.
What is your microscopic diagnosis?
At medium and higher magnifications seen above this otherwise typical hyperplastic polyp contains a less than one millimeter superficial focus of intramucosal poorly differentiated carcinoma with no evidence of glandular or squamous differentiation. Although this focus was lost when attempts were made to obtain additional tissue for immunohistochemical stains, it would not be surprising if this poorly differentiated focus exhibited some (or considerable) neuroendocrine differentiation.
I have encountered small foci of high-grade neuroendocrine carcinomas involving the bases of villous and tubulovillous adenomas. These lesions appeared to invade the base and although only 1-2 millimeters in size exhibited vascular invasion and hepatic metastases. This is the first example of a poorly differentiated, probably neuroendocrine, carcinoma arising in a hyperplastic polyp that I have encountered. The fact that this focus is confined to the tip of the polyp and may well be entirely "in situ" suggests that this patient may do well (and is INCREDIBLY LUCKY!).
This case raises the question of whether this is actually a serrated sessile adenoma (SSA), a diagnosis I freely admit to having considerable difficulty reproducibly making, and whether the patient (or lesion) has microsatellite instability. Prior polyps from this patient, one of which WAS considered to be an SSA did not show microsatellite instability.
Although I'm not sure that any additional therapy other than repeat colonoscopy after a few months, is warranted for this patient, this case emphasizes the pitfall of the common endoscopic practice of throwing all of the colon polyps from a patient into a single specimen jar. Several times a year we invariably encounter problems with this approach, usually due to an unsuspected carcinoid tumor or occasionally a typical adenocarcinoma involving the base of an endoscopically benign-appearing polyp. Although submitting EACH polyp in a separate specimen jar would be costly and time-consuming overkill, it would be far better if endoscopists would at least divide their polyps into ascending, transverse, descending, and rectosigmoid specimens. These areas would at least correspond to the surgical variants of partial colectomy in the likely event that the base of a malignant polyp cannot be found on repeat colonoscopy. Many, but not all endoscopists do follow this approach. Of course, ANY clinically suspicious polyp should be submitted separately, and the base can be tattooed to aid in finding the area of the polyp on subsequent examination.
Friday, August 05, 2011
A 35-year-old woman presented with abnormal vaginal bleeding. Ultrasound showed a large endometrial polyp. A hysterectomy was performed yielding the specimen seen above.
What is your diagnosis?
The polypoid mass consists of a mixture of benign-appearing endometrial glands with some infolding and a bland, variably cellular stroma. The stroma shows a strong tendency to cuff around the glandular structures, as can clearly be seen in the figure above and at higher magnification below.
Within these hypercellular areas the mitotic rate is greater than 5 mitotic figures per 10 high power fields (see below). Overtly atypical mitotic figures were not seen. The lesion did not invade the myometrium and there were no areas of stromal overgrowth, high grade nuclear features, or heterologous differentiation. This is a nice example of a mullerian adenosarcoma seen in consultation last week.
As the name indicates, adenosarcomas are proliferations consisting of a mixture of benign glands and sarcomatous stroma, with the latter component often forming a distinctive cuff or "cambium layer" around the glands. The resemblance to phyllodes tumor of the breast is often quite striking. Although uterine adenosarcomas more typically occur in postmenopausal women, unlike carcinosarcomas, they not uncommonly also occur in younger patients as well. Features associated with a poorer prognosis include myometrial invasion, high-grade sarcoma, often with heterologous (typically rhabdomyosarcomatous) differentiation, and stromal overgrowth. Tumors lacking these features, as in this case, are usually cured by simple hysterectomy and the prognosis for this patient is quite good.
It should be noted that both adenosarcomas and unusual endometrial polyps have been associated with long-term administration of tamoxifen. Given the widespread use of tamoxifen, the risk of this occurrence must be quite low, but exact numbers are not currently available.
Wednesday, July 27, 2011
A 75-year-old man presented with an several nasal "polyps" and a slightly larger nasal mass. The polyps were excised and interpreted as typical inflammatory polyps. The mass lesion was biopsied and the specimen demonstrated a light microscopically undifferentiated neoplasm as seen below.
The neoplasm consists of patternless sheets of medium-sized cells with high nuclear to cytoplasmic ratios, pleomorphic, nuclei, a high mitotic rate and numerous apoptotic bodies. The cells have small amounts of eosinophilic, often bubbly appearing cytoplasm which tends to be eccentrically oriented with respect to the nucleus.
The higher magnification view above better illustrates the cytologic features. Immunohistochemical stains for cytokeratin, CD20, and CD45 were negative. Epithelial membrane antigen (EMA) was focally positive.
What is your diagnosis?
Additional immunohistochemical studies were performed. An immunohistochemical stain for CD138 was srongly positive (above), and an in-situ hybridization for EBV-RNA (EBER) was also strongly positive (below).
This is an example of a plasmablastic lymphoma, a somewhat heterogeneous group of aggressive large B-cell lymphomas with a variably plasmacytic phenotype. Some examples have a much more myeloma-like microscopic appearance and their plasmacytic differentiation is easily appreciated by light microscopy alone. Other examples, like this one, are much more undifferentiated in appearance and likely to be confused with carcinoma, rhabdomyosarcoma, rhabdoid tumor, malignant melanoma, and other high-grade neoplasms.
These tumors typically show a strong association with EBV, based on in-situ hybridization for EBER, as in this case. They are typically negative for HHV-8. Plasmablastic lymphomas were initially described in the oral mucosa of HIV+ patients, but they may arise in other extranodal sites in the head and neck region, and elsewhere. They also occur in patients with other causes of immunodeficiency and, as in this case, in older patients.
Like anaplastic lymphomas, plasmablastic lymphomas typically show staining patterns (CD45-, EMA+) that may lead one away from the diagnosis of a lymphoproliferative lesion. Having been fooled by these before, I always include CD138 in my staining panel for light microscopically undifferentiated "epithelioid" malignancies in the head and neck. Other stains such as MUM1 will confirm the plasmacytic nature of the lesion, and EBER should be strongly positive.
Unfortunately, these are highly aggressive neoplasms and most patients succumb within a year of diagnosis.
Monday, July 18, 2011
A myocardial biopsy was performed and yielded the specimen seen above, and at higher magnification below.
What is your diagnosis?
The biopsy specimen shows a dense inflammatory infiltrate composed mainly of lymphocytes, histiocytes, eosinophils, and occasional plasma cells, with associated myocyte necrosis. Residual myocytes are disrupted and separated by the dense inflammatory infiltrate. Multinucleated giant cells are easily identified, but well-formed granulomas are not present. Immunohistochemical stains for CD68 are positive in the giant cells, documenting their monocyte/macrophage lineage. They are negative for muscle markers and thus do not represent regenerating myocytes. The lymphoid infiltrate labels predominantly with T-cell markers. B-cells are rare.
This is an uncommon but typically easy to diagnose entity, idiopathic giant cell myocarditis. The lesion is characterized by usually abrupt onset in previously healthy young to middle-aged adults. The prognosis is grim without high-dose immunosuppressive therapy or cardiac transplantation. The exact cause is unknown but certainly suggestive of an autoimmune process. About 20% of patients will have other autoimmune diseases including ulcerative colitis, rheumatoid arthritis, myasthenia gravis, and drug hypersensitivity reactions.
The distinction of idiopathic giant cell myocarditis from sarcoidosis can often be difficult. As noted above, well-formed granulomas are rare in giant cell myocarditis and are classic for sarcoidosis. Conversely, eosinophils are numerous in giant cell myocarditis and rare in sarcoidosis. The latter also typically shows more fibrosis, less mononuclear inflammation, less evidence of myocyte necrosis, and more of a mass-like appearance.
Wednesday, July 06, 2011
A 37-year-old man presented with a firm, several cm. mass in the region of his right shoulder. An excision was performed and the lesion was interpreted as nodular fasciitis. The mass recurred over the ensuing few months and a re-excision was performed, yielding the specimen seen above.
The figure shows a modestly cellular proliferation of bland spindled cells which form broad, sweeping fascicles and engulf adjacent skeletal muscle fibers. Although not seen in the figure, the margin of the lesion was infiltrative and somewhat ill defined. There was no nuclear pleomorphism or necrosis. Mitotic figures were vanishingly rare.
This was a case received in consultation last week. Both the original and recurrent lesions were submitted and both were identical spindle cell proliferations. The figure below shows the neoplastic cells and their uniform, markedly elongated nuclei at higher magnification.
Immunohistochemical stains for smooth muscle actin showed focal weak positivity (not shown). The tumor was negative for desmin and S100 protein. Antibodies to beta-catenin showed prominent nuclear positivity in the majority of neoplastic cells, seen below.
This is a prototypical example of an extra-abdominal ( aggressive, desmoid-type ) fibromatosis. These lesions are noted for their infiltrative but often non-destructive growth and their strong tendency to local recurrence. If properly defined, extra-abdominal fibromatoses do not metastasize. Local control, however, may be difficult. The strong nuclear staining for beta-catenin is an extremely helpful finding present in virtually all cases. The immunohistochemical stain can be somewhat tricky to perform, however.
Although originally interpreted as nodular fasciitis, this lesion lacks the haphazard, tissue culture-like appearance and sharp circumscription of that entity. Moreover, nodular fasciitis virtually never recurs and recurrence following that diagnosis should lead to rapid re-appraisal of the original interpretation.
Monday, June 27, 2011
A 75-year-old woman was noted to have a suspicious breast lesion on routine mammography. An excisional biopsy was performed, yielding the tissue seen below.
By routine H&E stain, this appears to be a clearly intraductal lesion. Cytologic features of malignancy were noted and the lesion was interpreted as apocrine DCIS. Diagnosing apocrine DCIS is non-trivial, but is not the primary subject of this case and the photos do not show the full range of cytologic atypia encountered in this lesion. However, a higher power image of one focus of atypia is seen below.
The contributing pathologist performed immunohistochemistry for myoepithelial cells using SMA, CD10, CK5/6, and p63, looking for occult areas of microinvasion. The result, which was the same for all markers, is shown below.
He was startled when there was complete absence of myoepithelial cells around these large duct-like structures with all of the markers employed. The case was sent for consultation with the question of whether this could be an invasive carcinoma with a pushing border.
I have seen several cases in the last month raising exactly this issue. Recent studies have demonstrated that apocrine lesions of the breast, whether metaplastic or neoplastic, typically show a marked diminution or complete lack of an outer myoepithelial layer (AJSP 2011;35:202-11).
The frequency of this finding in apocrine intraductal lesions is not entirely clear and will require further study. It is also unclear why this phenomenon should accompany apocrine proliferations in particular.
It is somewhat better understood that expansile intraductal lesions, regardless of cytologic type and regardless of organ, will often produce "stretching" and some attenuation of the outer myoepithelial layer such that it may well appear discontinuous in some planes of section. In the breast, complete absence, in our experience, has only been encountered with apocrine intraductal lesions. Negative immunohistochemical markers are always problematic when used to support a "positive" result (invasion in this case). Until this phenomenon is much better understood, it is important to remember that the presence of myoepithlial cells in breast proliferations is strong evidence for an in-situ process, but lack of myoepithelial cells does not fully equate with an invasive lesion, esp. when the light microscopic features are typical of an intraductal lesion.
Friday, June 17, 2011
An 81-year-old man underwent upper GI endoscopy for investigation of chronic diarrhea. A several millimeter nodule was noted in the region of the duodenal bulb and was biopsied. The resultant specimen is shown below.
In the submucosa there is a sheet-like infiltrate of epithelioid cells with a signet-ring-like appearance seen better in the insert at the left. The overlying mucosa was not ulcerated and appeared normal. The neoplastic cells had relatively uniform nuclei virtually absent mitotic figures. There was no evidence of necrosis or vascular invasion.
What is your diagnosis?
In some areas, the neoplastic cells were discohesive and the individual cells had a distinctly signet-ring appearance as seen above. The pale eosinophilic material within these cells was completely negative on a PAS stain (not shown).
As seen above, immunohistochemistry for cytokeratin showed strong positivity, particularly in the region of the cytoplasmic "signet rings." Central "palor" in these aggregates may relate to the difficulty of antibody penetration into the crystalline arrays of filaments.
The neoplastic cells were also strongly positive for chromogranin (not shown) and for synaptophysin (seen above.)
This is an example of a duodenal carcinoid tumor with prominent signet-ring-like cells. The latter cells are due to the accumulation of cytoplasmic proteins, predominantly cytokeratins, rather than mucin as is seen in gastrointestinal adenocarcinomas. Signet-ring carcinoid tumors containing aggregates of cytokeratin filaments have been described in the gastrointestinal tract. There is no evidence that these tumors behave differently than conventional carcinoid tumors, although the number of cases is very small. Duodenal carcinoid tumors in general are rare and fully capable of developing metastatic disease, which may be present at the time of initial diagnosis.
It is important to distinguish this morphologic variant of a typical carcinoid tumor from mixed lineage tumors variously labelled as mixed carcinoid-adenocarcinoma, goblet cell carcinoid, adenocarcinoid, or amphicrine carcinomas. The latter tumors contain cytoplasmic mucin within their signet-ring or goblet-shaped cells and appear to behave in a more aggressive fashion, typical of an adenocarcinoma. They are particularly common in the appendix.
Wednesday, June 08, 2011
A 59-year-old man (your humble blogger) presented with an approximately 1-year history of a 2-3 mm firm dermal mass involving the right side of his chin. The mass was non-tender, and appeared stable in size. The patient sought removal primarily because the mass was repeatedly traumatized during shaving. An excision was performed.
The biopsy specimen showed intact surface epithelium underlain by a mixed inflammatory infiltrate consisting primarily of histiocytes, many of which were multinucleated with nuclear "wreaths" (Touton-type giant cells).
What is your diagnosis?
At higher magnification, the Touton-type giant cells, some with foamy cytoplasm, are readily apparent, along with background mononuclear giant cells and scattered lymphocytes. Mitotic figures are not identified and there is no necrosis. Although not shown, the infiltrate closely "hugs" hair follicles and has a relatively sharp, triangular base in the deep dermis. The neoplastic cells stain strongly for CD68, also not shown.
This is a nice example of a juvenile xanthogranuloma. This is a relatively common lesion of infancy and early childhood but occasionally arises in patients of any age. The lesions vary considerably from patient to patient in terms of the amount of lipidization and the number of characteristic Touton-type giant cells.
Clinically, the head and neck region is the most common site, followed by the trunk. Lesions may remain stable in size for years or regress spontaneously. Surgery is the treatment of choice and the prognosis is excellent with only occasional recurrences.
The main differential diagnosis is with Langerhans cell histiocytosis. Unlike the latter tumors, the cells of JXG are negative for CD1a and generally negative for S100 protein and (presumably) Langerin. Touton giant cells are rare in Langerhans histocytosis and eosinophils are usually not numerous in juvenile xanthogranuloma, though they may occasionally be plentiful.
Thursday, May 26, 2011
A 64-year-old man presented with a painful left femur. Radiographs showed a lytic, destructive lesion involving the distal femoral shaft. The lesion had a somewhat biphasic radiographic appearance, with one part of the lesion containing annular "popcorn" calcifications, and the remainder being devoid of matrix and extending through the cortex into the surrounding soft tissues.
The low-power image of the tumor (above) shows areas of very low-grade hyaline cartilage immediately adjacent to a pleomorphic neoplasm composed of spindled and epithelioid cells with no evidence of differentiation. Mitotic figures including overtly atypical forms were easily found in the second component.
At higher magnification, the sharp interface between the low-grade cartilage and the pleomorphic stromal component is more easily visualized. An atypical mitotic figure is present at left-center field, just above the chondroid component. The sharp juxtaposition of the low-grade cartilaginous component and the pleomorphic sarcomatous component defines this lesion as a so-called dedifferentiated chondrosarcoma. Although many authors have noted that this is much more likely due to a failure of stem cells to differentiate as opposed to dedifferentiation of more mature chondroid cells, the term is well established for both this lesion and other forms of "dedifferentiation."
Dedifferentiated chondrosarcoma can often be strongly suspected radiographically when there is, as in this case, the sharp juxtaposition of a low-grade chondroid lesion producing characteristic "popcorn" calcifications, and a completely lytic, highly destructive secondary component. Although biopsies from both elements are ideal for making the diagnosis, the chondroid component can be inferred radiographically, and if the high-grade component is confirmed by biopsy, the diagnosis is secure.
Of all forms of dedifferentiated sarcoma, dedifferentiated chondrosarcoma appears to have the worst prognosis with 90+% tumor-related mortality. This differs from, for example, dedifferentiated liposarcoma, which is not the most malignant form of liposarcoma.
Dedifferentiated chondrosarcoma can be distinguished relatively easily from chondroblastic osteosarcoma. The latter tumor occurs predominantly during the adolescent growth phase, whereas dedifferentiated chondrosarcoma is a tumor of later life. More importantly, microscopically, dedifferentiated chondrosarcoma has distinct and discrete low-grade cartilaginous and high-grade pleomorphic elements, whereas in chondroblastic osteosarcoma, the cartilage is higher grade, and much more diffusely admixed with the remainder of the sarcoma. Although the high-grade component of dedifferentiated chondrosarcoma most often resembles, as in this case, an undifferentiated pleomorphic sarcoma (aka,"malignant fibrous histiocytoma"), it may occasionally consist of a high-grade osteosarcoma, or show evidence of other stromal lineages.
Tuesday, May 17, 2011
A 30-month-old boy presented with a 6 cm. mass involving his left kidney. A nephrectomy was performed. Grossly, there was a lobulated, yellow-tan mass replacing much of the renal parenchyma.
Microscopically, the mass consisted of sheets of medium-sized cells separated by a prominent capillary network. The neoplastic cells had predominantly clear cytoplasm and indistinct cell borders. The nuclei were variable in size and shape but relatively uniform chromatically with faintly stippled chromatin and indistinct nucleoli. Mitotic figures were easily found, but there was no obvious necrosis.
What is your diagnosis?
Higher magnification shows the chromatically uniform nuclei with absent or small nucleoli, clear cytoplasm and capillary rich stroma. This is the classic pattern of a clear cell sarcoma of the kidney. These tumors, like rhabdoid tumors, are distinct from Wilms' tumor and considered to represent an "unfavorable histology" for pediatric renal neoplasia. They have also been known as "bone metastasizing renal tumor of childhood." In the past, prognosis has been poor, but has been improved considerably with the addition of adriamycin to the chemotherapy regimen. It is therefore important to distinguish these tumors from Wilms tumor.
Although the classic pattern of clear cell sarcoma, as seen here, is relatively easy to recognize, there are several variant patterns that can be much more difficult to properly classify. The variant patterns include epithelioid, spindle cell, myxoid, sclerosing, palisading, and anaplastic.
Immunohistochemistry has not been of great value in the diagnosis of clear cell sarcoma. The tumors are strongly vimentin positive, and negative for markers of epithelial, neuroendocrine, neural, and muscle differentiation.
Tuesday, May 10, 2011
A 58-year-old man presented with an enlarging, polypoid mass involving one vocal cord. A smoking history was not available. A biopsy was performed. The resultant specimen was a polypoid, partially ulcerated lesion. The intact squamous mucosa does not demonstrate overt dysplasia. The underlying stroma contains markedly pleomorphic stellate to spindled cells in a fibrous to myxoid background.
What is your diagnosis?
At higher magnification, the atypical cells closely approach the surface epithelium and in some areas appear to "drop off" of the mucosa. Focally, pleomorphic basal-layer cells with atypical mitotic figures are also noted, although the overyling more superficial mucosa, as mentioned above, does not show features of typical squamous dysplasia.
A cytokeratin cocktail immunohistochemical stain shows that the atypical stromal cells are strongly cytokeratin positive. This is an example of a sarcomatoid (spindle cell) carcinoma. In the head and neck, the vocal cords are a relatively common site for these tumors. Sarcomatoid carcinomas often produce distinctly polypoid, often pedunculated masses. A number of years ago, Vince Hyams at the AFIP noted that these lesions should be staged based on the size of the base of attachment, rather than the diameter of the polypoid mass.
The distinction of sarcomatoid carcinoma from a true sarcoma may be straightforward, as in this instance when there is strong cytokeratin reactivity and surface mucosal changes are also noted. However, only about 50% of sarcomatoid laryngeal carcinomas will be cytokeratin positive and some may even show evidence of myogenous differentiation. Interestingly, in the vocal cords, the behavior of such cytokeratin-negative tumors is identical to that of their CK-positive counterparts. It has been suggested that p63 is slightly more sensitive than cytokeratin for recognizing sarcomatoid carcinomas.
The above findings call into question the diagnosis of a true sarcoma arising in the vocal cords, even ones showing overt myogenous, osseous or other stromal differentiation. With the exception of low-grade chondrosarcomas, The vast majority of such tumors are better interpreted as sarcomatoid carcinomas.
Monday, May 02, 2011
Unfortunately, radiographs were not available for review, but were said to show an expansile, lytic lesion with no evidence of calcified matrix. A core biopsy of the humerus lesion was performed and this material was submitted to us for consultation.
The biopsy shows vacuolated epithelioid cells with a background of myxoid and fibrous stroma, as well as islands of reactive bone.
What is your diagnosis?
At higher magnification, the tumor can be seen to be composed of highly vacuolated, "bubbly" epithelioid cells with a minor component of cells having more uniform, eosinophilic cytoplasm. There is mild to focally moderate nuclear variation. Mitotic figures are rare, and there is no evidence of necrosis or vascular invasion.
The presence of the highly vacuolated "physaliphorous" cells should immediately lead to consideration of a metastatic chordoma as the diagnosis. The tumor cells were strongly positive for S100, EMA, and a cytokeratin cocktail, supporting this diagnosis. Additional clinical history was obtained from the referring pathologist, and it was noted that the patient had a "distant" history of a resection for a sacral chordoma.
Chordomas are uncommon neoplasms that with vanishingly rare exception, arise along the route of the fetal notochord, show notochordal differentiation and presumably arise from notochord or perhaps more likely from "chordally predisposed" stem cells that inhabit this region. Recently, brachyury has been shown to be a highly specific marker for these neoplasms, although in most instances S100 protein positivity, combined with positivity for epithelial markers (CK, EMA), is sufficient for the diagnosis. Even in major cancer centers, an antibody to brachyury purchased for clinical diagnosis is likely to expire long before it is used up.
The current case is an example of a classic chordoma, but several subtypes have been described including chondroid chordomas, and dedifferentiated chordomas. The latter are often recurrences of prior classic chordomas. In the current case, no evidence of dedifferentiation was found.
Metastases from chordomas are known to go to a wide variety of anatomic sites, including skin where they may easily be confused with cutaneous mixed tumors (chondroid syringomas), chondroid lipomas (see an earlier image of the week) or the rare parachordoma. In the bone, metastatic chordoma (away from the notochord axis), is likely to be confused with a high-grade chondrosarcoma. Positivity for epithelial markers (and brachyury if available), will be extremely helpful in the diagnosis. S100 protein positivity is, of course, of no value.
Wednesday, April 20, 2011
A 31-year-old woman presented with a mass in the superficial lobe of her right parotid gland. A superficial parotidectomy was performed and yielded the specimen seen above, and at higher magnification below.
The specimen consists of epithelioid cells with eccentric nuclei and prominent "glassy" eosinophilic cytoplasm. The cells are cytokeratin positive and the glassy cytoplasm also stains strongly for vimentin. Smooth muscle actin staining is present, often around the periphery of cytoplasmic aggregates of vimentin, but can be sparse in some cells.
This is an example of a plasmacytoid myoepithelioma. Salivary myoepitheliomas are more typically composed of spindled cells, but the plasmacytoid variant is well recognized. Myoepitheliomas are part of the spectrum of monomorphic salivary gland adenomas including oncocytomas, basal cell adenomas, canalicular adenomas, and others. Mixed tumors ("pleomorphic adenomas") also contain large numbers of myoepithelial cells and the plasmacytoid form may be present as well. In fact, it is not uncommon to encounter lesions with the typical features of a myoepithelioma with only one focus on one slide showing evidence of stromal differentiation, often in the form of cartilaginous tissue.
As their name implies, plasmacytoid myoepitheliomas are composed of cells that closely resemble plasma cells, though they lack the characteristic "clock face" nuclear chromatin and perinuclear "hoff" of the latter cells. The resemblance can be striking, however. I once received a case from a former resident who was setting up an immunohistochemistry lab in his hospital and wanted to know why his kappa and lambda light chain stains did not work on this "obvious plasmacytoma."
It should be noted that the plasmacytoid phenotype can be seen in a wide variety of other neoplasms including malignant melanoma, rhabdomyosarcoma, and neuroendocrine carcinomas.
Friday, April 08, 2011
A 42-year-old woman underwent an endometrial curettage for dysfunctional bleeding. The curettings contained the tissue shown in this figure. Within the endometrium were sharply defined, eosinophilic nodules containing variably pleomorphic cells. Mitotic figures were not present and there was no necrosis.
What is your diagnosis?
This is an example of a placental site nodule or plaque, also known as a subinvoluted implantation site. The nodule consists of chorionic-type intermediate trophoblasts embedded in an eosinophilic fibrillar background matrix. The nodule is typically sharply demarcated from the surrounding endometrium and superficial myometrium. The trophoblastic cells with stain strongly for low molecular weight cytokeratins (CK 8/18). Staining for human placental lactogen (HPL) is typically focal and staining for beta-HCG is usually absent. In contrast to proliferative lesions of intermediate trophoblasts, the Ki-67 labelling index should be essentially zero.
Placental site nodule should be distinguished from proliferative trophoblastic lesions including exaggerated placental site reaction, placental site trophoblastic tumor, and epithelioid trophoblastic tumor. The small size and sharp circumscription usually allow ready distinction and, if necessary, the low (zero) Ki-67 index will also be helpful. Placental site nodule may also be confused with squamous cell carcinoma and this confusion may be aided by the cytologic atypia and cytokeratin positivity of placental site nodule. Again, the small size and lack of mitotic activity / Ki-67 index should allow ready distinction.
Placental site nodules have been thought to represent residua of prior pregnancy. They may persist for years. Occasionally they are found in patients with no known history of pregnancy and may well represent a blighted gestation that never fully developed and was unrecognized by the patient. Regardless, the lesion requires no further therapy and is adequately treated by the curettage (or hysterectomy) that led to its diagnosis.
Wednesday, March 30, 2011
A 45-year-old woman presented with some alteration in her voice. Laryngoscopy revealed a 1 cm. mass in the supraglottic region of the larynx. The mass was covered by intact, normal-appearing mucosa. A local excision was performed.
The resected lesion, shown above, consists of irregular but well-demarcated nests of epithelioid cells. The cells have uniform nuclei and prominent, eosinophilic cytoplasm. Mitotic figures are absent and there is no evidence of necrosis. The overlying mucosa, not shown, is intact and separated from the tumor by a narrow zone of normal stroma.
What's your diagnosis?
The nested arrangement of the cells should immediately suggest the possibilities of well-differentiated neuroendocrine carcinoma, metastatic medullary thyroid carcinoma, paraganglioma, and (much less likely) malignant melanoma. A couple of immunohistochemical stains will greatly aid in this distinction.
A cytokeratin stain (above) is completely negative. This effectively eliminates a diagnosis of well-differentiated neuroendocrine carcinoma, including metastatic medullary thyroid carcinoma.
An S100 stain (above) shows rimming of the cell nests by S100-positive cells. The lack of diffuse staining eliminates the unlikely possibility of a malignant melanoma and the presence of peripheral S100-positive cells indicates the presence of sustentactular cells around the cell nests, a feature typical of paragangliomas.
This is an example of a laryngeal paraganglioma. Paragangliomas are relatively common in the head and neck, but the larynx is a distinctly rare location for these lesions. In fact, in the larynx they are greatly outnumbered by and often confused with well-differentiated neuroendocrine carcinomas, aka "typical carcinoid tumors." In some of the older literature on paragangliomas, including the 2nd series AFIP fascicle, it is noted that "paragangliomas" of the larynx have a higher rate of metastatic disease than paragangliomas arising elsewhere in the body. This is due to the misdiagnosis of a substantial number of well-differentiated (and probably moderately differentiated) neuroendocrine carcinomas as paragangliomas.
The normal larynx has four paraganglia, two slightly larger ones located symmetrically in the supraglotic portion of the larynx and two symmetrically located smaller ones in the subglottic larynx. Presumably, these are the source for the paragangliomas that rarely arise in this region.
Wednesday, March 23, 2011
Resection specimen of shoulder mass.
This mass has a distinctly triphasic appearance consisting of mature fat, well-formed fascicles of mature myofibroblasts, and nests of smaller, more immature cells in a somewhat myxoid matrix. The three elements vary considerably in amount from lesion to lesion but are always intimately admixed, often in a vaguely organoid fashion.
The resultant appearance, once seen, is easily identified in subsequent cases. This is an example of a fibrous hamartoma of infancy. Although usually a straightforward diagnosis (once considered), the lesion should not be confused with more aggressive forms of fibromatosis including infantile fibromatosis and diffuse myofibromatosis. The generally small size of the lesion and its triphasic "organoid" appearance should allow for distinction.
This lesion typically occurs in the first two years of life and is often rapidly growing. Although it can occur in any subcutaneous site, the region of the shoulder, back and axilla are common locations. The lesions are solitary and there is no evidence of a genetic association.
Fibrous hamartoma of infancy is cured by local excision. A few cases may recur and necessitate limited re-excision.
Wednesday, March 16, 2011
Plain films of the distal right phalangeal lesion. The bone is expanded and the cortex is thinned by a lytic intramedullary lesion with thin intralesional septations. A soft tissue component is not appreciated and the external contour of the finger appears virtually normal.
A curettage of the lesion was performed. Microscopically it consists of hyaline cartilage exhibiting areas of rather striking hypercellularity, with considerable nuclear pleomorphism. Nonetheless, this is an example of an enchondroma.
This is a great illustration of a circumstance in which the location of the lesion and, to a lesser degree, the radiographic features override the microscopic appearance. There is no doubt that this degree of hypercellurity in a long bone cartilaginous tumor would strongly support a diagnosis of grade I or even grade II chondrosarcoma. However, in the bones and soft tissues of the fingers, this degree of pleomorphism (and more!) is routinely encountered in benign cartilaginous neoplasms.
True chondrosarcomas of the digits are vanishingly rare, reportable neoplasms and are characterized by extensive osseous destruction with a LARGE soft tissue component. Even so, I am not aware of an example that metastasized from this location.
The radiographic differential diagnosis for this lesion includes other benign lesions involving the distal phalanges such as giant cell (reparative) granuloma, epidermal inclusion cyst, and vascular glomus tumor.
Friday, March 04, 2011
A 44-year-old man presented with a nasopharyngeal mass. The slides of the mass were waiting for me when I returned from the USCAP meeting. Microscopically, it consists of rounded nests of larger cells with uniform, vesicular nuclei and prominent nucleoli (inset), surrounded by a predominantly lymphocytic infiltrate. The larger cells show no evidence of glandular or squamous differentiation. Cell borders are indistinct and the nuclei appear to be floating in a syncytium of cytoplasm.
What is your diagnosis?
The differential diagnosis includes a mixed cell lymphoma and some form of carcinoma. The results of a broad spectrum cytokeratin "cocktail" are shown below.
The larger cells in the nests are strongly cytokeratin positive. The surrounding background is CK negative and shows features of a reactive lymphoid infiltrate (immunohistochemistry not shown). Given the cytologic uniformity and the location of the lesion, this was considered to represent a nasopharyngeal undifferentiated carcinoma of lymphoepithelioma type. To confirm this diagnosis, in-situ hybridization for EBV RNA (EBER) was performed and the results are shown below.
The larger cells in the nests are strongly reactive for EBER. In the nasopharynx, if tumors with this histologic appearance do not react for EBER, other diagnostic possibilities should be strongly considered.
Other carcinomas to be considered in the face of EBER negativity would include sinonasal undifferentiated carcinoma (SNUC), a tumor not associated with EBV or HPV and usually, as the name implies, arising anterior to the nasopharynx in the paranasal sinuses. The cells of SNUC are typically more plemorphic than those of lymphoepithelioma, and often have more prominent cytoplasm, even larger nucleoli, and typically abundant necrosis.
Small cell carcinomas rarely arise in this location and may show some neuroendocrine differentation. They are composed of smaller cells with hyperchromatic nuclei and often prominent nuclear molding. CK immunohistochemistry may show dot-like perinuclear positivity, a surrogate for neuroendocrine differentiation.
Non-keratinizing squamous carcinomas in the nasopharynx may also be associated with EBV infection, but manifest more cytologic variability with distinct cell borders.
Nasopharyngeal lymphoepitheliomas, although often metastatic at presentation, often show an excellent response to radiation therapy, unlike more conventional kertinizing squamous cell carcinomas.
Monday, February 14, 2011
A 50-year-old woman presented with a left upper lobe lung mass. A wedge excision was performed and the resultant sections were referred for consultation. Smoking history was not immediately available.
Microscopically, the specimen consisted of sheets of large cells with prominent eosinophilic cytoplasm and chromatically uniform, vesicular nuclei with single small nucleoli. There was no evidence of gland formation or keratinization. The neoplastic cells exhibited numerous mitotic figures. Associated with the tumor was a mixed inflammatory infiltrate consisting of lymphocytes, scattered neutrophils, and prominent eosinophils.
What is your diagnosis?
A reasonable differential diagnosis for this case would include large cell undifferentiated carcinoma, large cell or histiocytic lymphoma, dendritic cell tumor and, of course, malignant melanoma. Markers for lymphoid, dendritic, and melanocytic lesions were negative and the neoplastic cells were strongly positive for cytokeratin (figure below).
Given the lack of squamous or glandular differentiation, a diagnosis of large cell undifferentiated carcinoma is accurate for this case. However, I believe that the diagnosis can be further refined. This is a specific sub-type of large cell undifferentiated carcinoma characterized by chromatically uniform cells with vesicular nuclei in an inflammatory background. The resultant appearance is strikingly reminiscent of nasopharyngeal "lymphoepithelioma." Tumors of this type have been described in the lungs in multiple studies. Interestingly, in studies from Asia, where these tumors account for over 5% of lung carcinomas, there has been a clear association with Epstein-Barr virus (EBV) infection. The EBV+ tumors do not appear to be associated with smoking and often occur in younger, non-smoking individuals. In American or other western patients, this association has not been convincingly or consistently documented in histologically identical tumors, and in this patient population, the tumor appears to be smoking related.
The current tumor, from a patient who was born in the United States, did not show EBV RNA (EBER) by in-situ hybridization and the patient was a smoker.
Because of the relative rarity of this tumor variant, biologic behavior is not well documented. It has been suggested that these neoplasms have a better prognosis than other lung carcinomas, but it is not clear if this applies only to the EBV-related tumors.
The next image of the week will be slightly delayed as I have back-to-back meetings for the next two weeks, the second being the USCAP meeting in San Antonio. I hope to see you there!
Thursday, February 03, 2011
A 33-year-old woman had a destructive mass involving the maxillary and ethmoid sinuses. A biopsy was performed and the resultant specimen is illustrated here. Beneath mainly intact respiratory-type surface mucosa is a neoplasm consisting of small to intermediate sized cells.
What is your diagnosis?
The tumor cells primarily form sharply demarcated nests. In many of the nests, as in the center of this field, the cells appear dyscohesive and free-floating within the resultant spaces. Scattered tumor giant cells are present and at higher magnification many of the tumor cells have irregular, eccentric nuclei with more prominent eosinophilic cytoplasm.
A reasonable differential diagnosis for this case would include
olfactory neuroblastoma (ONB), small cell carcinoma, sinonasal undifferentiated carcinoma (SNUC), and rhabdomyosarcoma. It is unclear if the tumor involves olfactory mucosa and the cells are more pleomorphic than is typical of ONB, but the nesting pattern is common in this neoplasm. The cells are too large for typical small cell carcinoma and they are more pleomorphic than is typical for SNUC. In addition, SNUC typically grows as patternless sheets of cells with prominent nucleoli and does not typically form well-demarcated cell nests. On H&E-stained sections, the presence of considerable nuclear pleomorphism with eccentric nuclei, cells with prominent eosinophilic cytoplasm, tumor dyscohesion, and tumor giant cells led me to suspect that this was going to be an alveolar rhabdomyosarcoma.
As shown above, this was confirmed by strong nuclear staining for myogenin and strong cytoplasmic staining for desmin. Cytokeratins and neuroendocrine markers were negative.
It is important to keep this diagnostic possibility in mind when dealing with high-grade, light microscopically undifferentiated neoplasms involving the head and neck region. This is a relatively common site for rhabdomyosarcomas and these tumors may arise in adults at virtually any age. Distinction of alveolar from embryonal rhabdomyosarcoma has important prognostic significance, given the poorer prognosis of the former tumors.
In addition, FISH studies were performed on this case using a "break apart" probe to evaluate for the (t1:13) or (t2:13) FKHR gene rearrangement. The cell below shows one copy of normal FKHR gene indicated by the close proximity of the red-green signal and one copy of rearranged FKHR with the red and green signals located on different chromosomes.
Wednesday, January 26, 2011
A 72-year-old man presented for evaluation of a small intra-abdominal mass. There was a history of malignant melanoma involving the scalp. No further details were available and slides from that cutaneous lesion had not been reviewed. The adominal mass was biopsied and submitted for frozen section.
What is your diagnosis?
The lesion consists of amorphous sheets of foamy (xanthomatous) and histiocyte-like cells, intermixed with aggregates of lymphocytes and plasma cells. Most of the histiocyte-like cells have rather small uniform nucleoli, but a few have larger nuclei with nucleoli. Mitotic figures are not readily identified. I got this case for frozen section and interpreted it as chronic inflammation with prominent xanthomatous histiocytes. I suspected that this was a reaction to something else nearby, not included in the biopsy specimen. However, although there are quite a few histiocytes in the lesion, most of the xanthomatous and histiocyte-like cells stain strongly positive for S100 protein and HMB-45.
As you've probably guessed by now, this is an example of a metastatic balloon cell melanoma.
Needless to say, it would have certainly helped to know the "balloon cell" part of the prior melanoma diagnosis, but even in retrospect it's virtually impossible to make a definitive diagnosis of melanoma on this frozen section. As we all know, melanomas can mimic almost anything including large cell carcinomas, small cell carcinomas, plasmacytomas, myoepitheliomas, lymphomas of all types, spindle cell sarcomas, and to that list should be added xanthomatous inflammation.
You humble blogger is, after all, only human. It's tough to feel too badly about this one, though. I'd probably make the same mistake again.
Monday, January 17, 2011
A 43-year old woman presented with a long history of pain involving her left shin. Over time a small bone-hard mass had developed in this region. Radiographs showed a mixed cystic and sclerotic lesion involving both the cortex and medullary cavity of the tibial diaphysis. A biopsy was performed.
What is your diagnosis?
The specimen consists of fragments of normal lamellar medullary bone separated by a loose fibrotic stroma. Embedded within the stroma are cords and nests of bland epithelioid cells with scattered slit-like lumina. The epithelioid cells are strongly positive for cytokeratin, confirming their epithelial nature.
This is an example of a osseous adamantinoma. These tumors show a strong predilection for the tibia (80%-85%), with most of the remainder arising in the fibula. The epithelial elements in adamantiomas vary considerably. Some may be basaloid, spindled, squamous or more tubular as in the current case. In the past the tubular structures have been confused with vessels leading to misdiagnosis as a vascular neoplasm.
How a partially epithelial tumor arises intraosseously has been a subject of considerable debate. It has been suggested that because of the close apposition of the tibia and overlying skin epithelial "rests" become incorporated into the tibia during embryonic development. Interestingly, we have seen one example of a soft-tissue adamantinoma-like neoplasm involving the pre-tibial dermis and subcutis.
Equally confusing and contentious is the relationship between adamantioma and osteofibrous dysplasia. It has been known for many years that adamantinomas often contain areas resembling osteofibrous dysplasia and, conversely, adamantinoma-like areas may be present in osteofibrous dysplasia-like tumors. However, longitudinal follow-up with conversion of one tumor to the other has not been documented. More recent studies have shown that adamantinoma and osteofibrous dysplasia share similar genetic anomalies (trisomies 7,8,12,21) further supporting a strong (though mechanistically unclear) association.
Adamantinomas are typically treated by en bloc excision. Rare cases have metastasized, sometimes many years later. Distinction from metastatic carcinoma is based on the bland nature of the epithelial cells, the characteristic location in the tibia or fibula, and the often young age of the patient.
Monday, January 03, 2011
A 65-year-old woman with a prior history of invasive ductal carcinoma of the right breast presented with a small nodule in the same breast 15 years later.
What is your diagnosis?
The image shows somewhat atrophic breast parenchyma embedded in fibrous stroma. The latter component created the "mass" in the otherwise fatty breast tissue. Within the atrophic ductules are scattered, markedly atypical cells with enlarged, hyperchromatic nuclei. The most florid example in this field is shown at higher magnification in the lower left inset. Notably, the atypical cells are widely scattered and have increased cytoplasm as well as nuclear enlargement, such that the nuclear to cytoplasmic ratio is not greatly different from the surrounding more normal cells. A "clone" of atypical cells is not present, the lobules are not expanded and there is no mitotic activity.
This is an example of radiation atypia. Changes such as this are common in breast tissue from women with a history of prior radiation therapy, though in many instances breast tissue with prior radiation shows little or no evidence of this change, suggesting variable patient susceptibility. Changes of this type, once present, likely persist for the remainder of the patient's life.
Distinction from adenocarcinoma is based on the scattered nature of the atypical cells, the lack of mitotic activity, and the overall atrophic appearance of the lobules. Paradoxically, breast carcinomas that survive radiation therapy typically show little or no morphologic change from their pre-radiation appearance.
Monday, December 20, 2010
This elderly man presented with a dermal nodule involving the skin of his right temple. An excisional biopsy was performed. Microscopic examination showed nodules of small blue cells consistent with normal lymphocytes and larger cells with uniform vesicular nuclei.
What is your diagnosis?
Although this was initially thought to be either an exhuberant reaction or perhaps a nodular lymphoma, immunohistochemical studies showed that the larger cells with vesicular nuclei were strongly cytokeratin positive. This is an example of a lymphoepithelioma-like cutaneous carcinoma.
These tumors favor the sun exposed skin of the facial region and typically occur in middle-aged and older individuals. This lesion is distinct from Merkel cell carcinoma, which also occurs in this setting. The latter tumors have smaller more hyperchromatic nuclei, neuroendocine differentiation, and punctate perinuclear cytokeratin positivity, in contrast to the diffuse cytokeratin positivity seen in this lesion.
Although histologically identical to "lymphoepithelioma", ie. undifferentiated nasopharyngeal carcinoma, these tumors in the skin have NO association with Epstein-Barr virus. To my knowledge these tumors have not been studied for a possible association with polyomavirus.
In spite of there high-grade, undifferentiated appearance, in the skin lymphoepithelioma-like carcinoma has a rather good prognosis, with local recurrence, rather than metastasis, being the main clinical complication.
Friday, December 10, 2010
This 1.5 cm intrabronchial polypoid mass was removed with a rim of normal bronchial mucosa. There was no invasion of the surrounding tissues. The tumor has a distinctive papillocystic pattern and consists of bland cuboidal to columnar cells with eosinophilic cytoplasm and no discernible mitotic activity. The luminal spaces contain a pink, proteinaceous fluid.
This is an example of a true bronchial adenoma. Before the early 1960's this term was used for virtually any intrabronchial low-grade neoplasm. We now know that many tumors included under that rubric were either carcinoid tumors or salivary-type (seromucinous gland) neoplasms including adenoid cystic carcinomas, epithelial-myoepithelial carcinomas, low-grade mucoepidermoid carcinomas, mixed tumors, and even acinic cell carcinomas.
Nonetheless, true benign bronchial adenomas do exist and were nicely documented in a study by England and Hochholzer (AJSP 19:887-899,1995) from 15 years ago. The tumors in that study were predominantly mucin-rich, but other cells types were present and cases such as the current one may be mucin poor.
Wednesday, December 01, 2010
An 85-year-old woman, living in a nursing home fractured her left femur at its "neck" just distal to the femoral head. Surgery to insert a prosthetic femoral head was delayed for two weeks. At the time of surgery a bulky, gritty soft tissue mass was noted to be engulfing the ends of the fracture. A biopsy was performed for frozen section to exclude a sarcoma.
What's your diagnosis?
This is an example of an extremely exhuberant but non-neoplastic process, a fracture callus. The body is attempting to repair the break by re-uniting the two ends of the fractured bone. Fracture callus begins as a proliferation of highly cellular fibroblast-like cells that first undergo cartilaginous metaplasia and then progressively mature to woven osteoid, lamellar osteoid, and fully calcified mature bone. The process often shows "geographic" maturation such that in single, properly oriented sections transformation from cellular fibrous tissue to cellular cartilage and then to well-formed bone can be seen on a single slide.
Taken out of context the cellular fibroblastic stroma (not well seen in this single photomicrograph) can occasionally be mistaken for a sarcoma. The cartilage represents an even greater pitfall in that its cellularity and nuclear pleomorphism are at least as great as that of a grade II chondrosarcoma of long bones.
The clinical features are highly helpful in arriving at the correct diagnosis and even with the exhuberant fracture callus, the radiographic appearance will remain that of a displaced non-pathologic fracture. This case emphasizes the importance of radiographic correlation with virtually every osseous lesion.
The zonal or geographic maturation seen microscopically is the key diagnostic feature and the intense osteoblastic rimming present in the maturing osteoid is not a feature encountered in osteosarcoma.
Fracture callus is microscopically identical to several other exhuberant but pseudoneoplastic processes including myositis ossificans and florid reactive periostitis.
Tuesday, November 23, 2010
The tumor consists of sharply defined nests of cells separated by prominent vascular channels. The cell nests often have free-floating or degenerating cells in the centers of the nests, with more viable-appearing cells clinging to the periphery of the nests. The individual cells have abundant clear to eosinophilic cytoplasm with relatively uniform nuclei and small nucleoli.
This is an example of a rare but microscopically well-defined entity, an alveolar soft-part sarcoma. This tumor was first defined (and named) in 1952 by William Christopherson. The exact differentiation of this tumor has been contentious from the beginning but recent molecular studies have greatly improved our understanding.
Alveolar soft-part sarcomas (ASPS) occur in a wide variety of anatomic sites, but predominantly involve the lower extremity. Most patients are adolescents or young adults and there is a female predominance.
Histochemically, ASPS's contain variable amounts of diastase sensitive glycogen as well as PAS-positive, diastase resistant rhomboid to rod-shaped crystals. Recently the crystals have been shown to consist of monocarboxylate transporter 1 (MCT1) and CD147.
Molecular studies have shown that ASPS has an unbalanced t(X;17) which creates an ASPL-TFE3 translocation protein. This translocation is also seen in so-called "translocation renal cell carcinomas" that often occur in children, among soft tissue tumors, it is unique to ASPS. Interestingly, ASPS and translocation renal cell carcinomas have quite similar histologic features.
Although, ASPS is a slow-growing tumor with an often protracted clinical course, the overall prognosis is poor with death due to metastatic disease. The tumor does not respond well to radiation therapy or current modes of chemotherapy.
Tuesday, November 16, 2010
A left-sided superficial parotidectomy was performed. Microscopically, there was a prominent plasmacytic infiltrate associated with periductular, periacinar, and perivascular sclerosis, as seen in the image above. Immunohistochemistry showed that the majority of the plasma cells labelled for IgG4. The salivary tissue engulfed by the inflammatory and fibrotic tissue shows atrophic changes.
This is an example of chronic sclerosing sialadenitis
(CSS), also known as "Kuttner's tumor" because of its tendency to produce tumor-like enlargement of the salivary gland. CSS is a member of the ever-growing group of IgG4-mediated sclerosing lesions. So-called autoimmune pancreatitis was one of the first IgG4-associated lesions and remains one of the most common, but the disease is amazing widespread throughout the body.
Immunohistochemical stains for CD138 (to identify all plasma cells) and IgG4 are key to making the diagnosis as there are other causes of sclerosing sialadenitis. Typically, more than 50% of the plasma cells will be IgG4+, whereas this IgG variant is much less common (~4%) in other plasmacytic infiltrates.
As evidence of how widespread (and popular) IgG4-related sclerosing disease has become, try a search for "IgG4" on the Pathology Network main page..... you'll get 400 "hits!"
Monday, November 08, 2010
The tumor is composed of small, uniform cells without nucleoli and with modest amounts of eosinophilic cytoplasm. The neoplastic cells form tubules, papillary structures and occasional glomeruloid-like bodies. There are scattered psammoma bodies. In some areas, the growth pattern appears solid at low-power, but at higher magnification, closely packed tubules can be seen in these areas. Mitotic figures are not identified. Foamy histiocytes are notably absent. There is no tumor capsule but the margin with the adjacent kidney is sharply demarcated.
This is an example of a metanephric adenoma, also referred to as an embryonal adenoma and a nephrogenic nephroma. Signs and symptoms are generally nonspecific, although polycythemia has been reported in a minority of patients, apparently due to direct production of erythropoietin by the tumor cells. Cytogenetic studies have not shown consistent abnormalities in these tumors. Although some have suggested that these represent end-stage differentiation of Wilms' tumors, evidence in support of this is weak at best.
The neoplastic cells of metanephric adenoma are often positive for cytokeratins, WT-1, and CD57. In contrast to papillary renal cell carcinomas, with which they are occasionally confused, metanephric adenomas are notably negative for AMACR.
Tuesday, October 26, 2010
At the time of surgery the 2 cm. mass was sharply demarcated, partialy encapsulated, and easily shelled out from the surrounding tissues. Microscopically, the tumor consists of a mixture of mature adipocytes with large solitary lipid vacuoles, sclerotic to myxoid and chondroid stroma, and lobular nests of uniform cells with central nuclei and clear finely vacuolated cytoplasm. Both the adipocytes and the clear cells stain strongly for S100 protein.
This is an example of a chondroid lipoma, and the abundant clear cells have ultrastructural and immunohistochemical features of chondrocytes. These uncommon lipoma variants show a tendency to involve the proximal extremities of middle-aged women, but can occur in a wide variety of anatomic sites and ages. The tumors are thought to be characterized by a balanced translocation involving t(11;16)(q13;p12-13).
The microscopic appearance of a chondroid lipoma can be quite worrisome if one has not encountered this tumor before and confusion with myxoid liposarcoma, myxoid chondrosarcoma, and myoepithelial neoplasms is common. In my experience, the tumor is most likely to be misdiagnosed as a myxoid chondrosarcoma and S100 positivity will be of no value in this distinction. Myxoid chondrosarcomas are larger, grossly lobulated lesions that are composed of cells with more non-vacuolated, eosinophilic cytoplasm. The integral adipocyte component will not be present, although the tumor may invade fat at its periphery.
Monday, October 25, 2010
A surgical resection specimen demonstrated an ill-defined multinodular mass composed of zones of necrotic appearing collagen-like material, surrounded by bands of polygonal cells with mild to at most moderate cytologic atypia and scattered mitotic figures.
This is an example of an epithelioid sarcoma, peripheral type. The neoplastic polygonal (epithelioid) cells stained positively for CD34, vimentin, and cytokeratin. They were focally positive for S100 protein. The tumor cell nuclei showed a characteristic loss of staining for INI-1. The latter is a relatively specific marker for epithelioid sarcoma and the closely related tumors, proximal epithelioid sarcoma, extrarenal rhabdoid tumor, rhabdoid tumor of the kidney, and atypical teratoid/rhabdoid tumor of the CNS.
The natural course of epithelioid sarcoma involving the extremities is one of multiple recurrences, often in a lymphovascular pattern of spread, progressively extending proximally along the extremity and often necessitating amputation. Distant metastases may also develop.
Because the degree of cytologic pleomorphism may be rather mild in these tumors and the lesion often consists of a band of such cells surrounding a central zone of necrosis, the critical differential diagnosis is with deep granuloma annulare or rheumatoid nodule. The "epithelioid" cells of granuloma annulare and rheumatoid nodule are histiocytic and will not express cytokeratin or CD34. They will be positive for INI-1 and usually for CD68.
Thursday, October 14, 2010
This case appeared in consultation over concerns about a possible sarcoma. There is a fibrinous exudate toward the surface (right) associated with fat necrosis and below that (left) are enlarged, somewhat atypical cells, partially resembling ganglion cells.
This is an example of so-called ischemic fasciitis
, also known as atypical decubital fibroplasia. The zonal appearance of the lesion with a band of fibrinous exudate and fat necrosis and an underlying zone of atypical cells is characteristic. The ganglion-like cells are similar to the cells seen in proliferative myositis and proliferative fasciitis. These cells are not ganglionic and stain strongly with vimentin and focally for actin and CD68, suggesting that they are myofibroblastic in differentiation.
The nodular pattern of the lesion may suggest the possibility of an epithelioid sarcoma, but the overall cellularity is quite low and the cells lack the immunohistochemical features of epithelioid sarcoma (CD34+, CK+, EMA+).
It is believed that this reactive lesion results from longstanding low-grade trauma to the region leading to repetitive tissue damage and regeneration.
Thursday, October 07, 2010
Diagnosing bone lesions without the accompanying radiograph is never a good idea, but we'll make an exception here, since I did get to see the films for this case. This was a recent consultation case and the radiographs showed a septated "soap bubble" lytic lesion eccentrically involving the metaphyseal region of the proximal tibia. The margins were well demarcated and there was no cortical bone destruction, in essence a lytic but radiographically benign-appearing lesion.
Microscopically, there is a prominent background of eosinophils and, less obviously, mast cells. Within this background are cords and nests of epithelioid cells with uniform vesicular nuclei and prominent eosinophilic cytoplasm. The cells closely resemble histocytes, cytologically, but are cohesive and in many areas form obvious lumina. Cytoplasmic strands often bridge the lumina.
This is an example of an epithelioid hemangioma. These lesions have also been known as "histiocytoid hemangiomas, " and in the skin as "angiolymphoid hyperplasia with eosinophilia" and even "pseudopyogenic granuloma." It is likely that older literature interprets these lesions, when in the bone, as low-grade angiosarcomas; a very unfortunate designation for a benign but occasionally multifocal osseous (and soft tissue) neoplasm. Epithelioid hemangioma is the benign end of a spectrum of epithelioid endothelial neoplasms including epithelioid hemangioendothelioma and epithelioid angiosarcoma. The latter two, unlike epithelioid hemangioma, are fully capable of malignant behavior, with epithelioid angiosarcoma, not surprisingly, being by far the most aggressive.
Distinction of epithelioid hemangioma from epithelioid angiosarcoma is easy given the overt cytologic pleomorphism and lack of well formed vessels in the latter tumors. Epithelioid hemangioendothelioma is characterized by a myxoid stroma with narrow strands of neoplastic cells with intracytoplasmic lumina, but not the well-formed vessels of hemangioma.
Tuesday, September 28, 2010
This was my evening frozen section on Monday night. A middle-aged woman had a 2 cm. sharply demarcated peripheral pulmonary nodule that was mildly PET positive. She had a long-term smoking history. The gross specimen showed a sharply circumscribed yellow-gold nodule just beneath the pleural surface. The microscopic appearance was somewhat variable and included distinctly papillary areas with the papillae lined by bland cuboidal cells, as well as somewhat more solid areas as shown here with small sclerotic nodules surrounded by bland epithelioid cells with uniform vesicular nuclei. This is an example of so-called sclerosing hemangioma. This entity, like many in pulmonary pathology was first described by Averill Leibow in the 1970's. The name derived from the fact that the epithelioid cells were though by Leibow to be endothelial. Subsequently, they were shown to be epithelial and considered to be derived from mesothelium. More recent and more reliable studies support that the epithelioid cells resemble type II pneumocytes and the alternate name of pneumocytoma has been suggested for this entity, but has yet to be widely accepted. The epithelial cells will label for cytokeratin and EMA, and show strong nuclear staining for TTF-1. They also show frequent intranuclear pseudoinclusions typical of type II pneumocytes. It is important to distinguish this lesion from a bronchioloalveolar adenocarcinoma as it is virtually always a benign process. The sharp demarcation, peripheral location, yellow-gold color, and occurrence in a middle-aged woman are typical features and were very helpful for suggesting the correct diagnosis at the time of frozen section.