The Pⁿ Blog is a forum for opinions, questions, controversies, and instructive discussions across the field of pathology and its relevant subspecialties.
Tuesday, February 25, 2014
In the past I have discussed the problems with overdiagnosis and overtreatment of both mammary carcinoma in situ and prostatic adenocarcinoma. A notable third example is the overdiagnosis and overtreatment of papillary thyroid carcinoma. A recent article from JAMA Otolaryngology - Head & Neck Surgery, entitled "Current Thryoid Cancer Trends in the United States
," and published online on February 20th details this problem.
To review some findings in the article:
Since 1975, the incidence of thyroid cancer has nearly tripled, from 4.9 to 14.3 per 100,000 individuals Virtually the entire increase is attributable to papillary thyroid cancer diagnoses. The absolute increase in thyroid cancer in women was almost 4 times greater than that of men. In spite of this dramatic increase, the mortality rate from thyroid cancer has remained stable between 1975 and 2009 (approximately 0.5 deaths per 100,000).
The authors conclude that there is an ongoing epidemic of thyroid cancer in the United States. The epidemiology of the increased incidence, however, suggests that it is not an epidemic of disease but rather an epidemic of diagnosis. The problem is particularly acute for women, who have lower autopsy prevalence of thyroid cancer than men but higher cancer detection rates by a 3:1 ratio.
That there is an epidemic of overdiagnosis and not of disease is based on the facts that: The increasingly detected cancers are highly unlikely to cause symptoms or death. Despite a more than 2-fold increase in the rate of thyroid cancer detection, thyroid cancer mortality had remained stable. Virtually all of the increase was confined to a type with the least aggressive histologic characteristics, namely papillary thyroid cancer. Finally, the bulk of the increase (87%) was explained by tumors 2 cm or smaller.
It is not surprising that women have bourne the brunt of this phenomenon. Benign thyroid disease is more common in women, leading to fine needle aspirations, and surgical resections that uncover occult and largely harmless lesions. Clinicians, the article suggests, are "looking too hard" for thyroid carcinomas. Because the article is a large review based on SEER data, there is no discussion regarding what culpability pathologists might bear for this "epidemic."
The article does not distinguish between conventional papillary carcinoma, a usually straightforward diagnosis, and the follicular variant, which, as any consultant reviewing such cases will attest, is often anything but straightforward. Indeed, the latter diagnosis is associated with considerable interobserver variability and given the overall excellent prognosis, refining diagnostic criteria would require a huge database with tens of thousands of years of patient follow up.
As mentioned above, a related issue that is discussed is the diagnosis of unequivocal but very small papillary carcinomas; lesions for which the term "microcarcinoma" has been used (and I also follow this practice). The authors raise the issue of whether such lesions should carry a designation other than "carcinoma," with all its associated connotations and emotional baggage. At the very least, clinicians should share with patients the extremely indolent behavior of small papillary carcinomas, and suggest the possibility of watchful waiting as appropriate therapy. After all, this is a tumor for which we have "magic bullet" therapy if needed. The current statistic of 85% total thyroidectomy following a diagnosis of papillary carcinoma (of any type or size) is far too high.
Monday, February 10, 2014
Pathologists and clinicians have been taught for decades that atypical lobular hyperplasia (ALH) and atypical ductal hyperplasia (ADH) of the breast are fundamentally different lesions. ALH has been viewed as a marker for subsequent carcinoma of lesser predictive power than lobular carcinoma in situ (LCIS), and associated with increased risk of carcinoma in either breast (although Page and colleagues did show that 2/3rds of subsequent carcinomas occur in the ipsilateral breast).
ADH has been viewed traditionally as a precursor lesion with more power than ALH to predict subsequent carcinoma, predominantly in the same breast. Surgeons will often do reexcisions when ADH is present at the margin but almost all will ignore margins positive for ALH (or even LCIS).
A recent article published in Cancer Prevention Research
based on a large study from the Mayo Clinic calls these assumptions into question. I recommend reading it and passing it on to your colleagues, both in pathology and the clinical specialties involved in breast disease management. There are a LOT more data in the text than are briefly discussed here.
The study involved almost 700 women with atypical hyperplasias, 330 with ADH, 327 with ALH, and 32 with both. This is by far the largest follow-up study in the literature on atypical hyperplasias of both types. To cut to the chase, the conclusion can be summarized in the following quote from the manuscript, "Our observations do not support present assumptions that ADH and ALH have substantively different behaviors. More DCIS may occur in women with ADH than ALH (25% vs. 13%, p=0.07), but numbers are small and this was not statistically significant."
Both ALH and ADH have features suggestive of precursor lesions and risk indicators. When carcinoma arises (either in-situ or invasive) both are predominantly associated with ductal as opposed to lobular carcinomas (ALH -> 77% ductal, 17% lobular; ADH -> 78% ductal, 8% lobular). Two-thirds of subsequent carcinomas arose in the ipsilateral breast in both ALH and ADH patients. Both ALH and ADH showed a tendency for contralateral carcinomas to develop later in life.
This is an valuable study that is packed with information, presented in a well-written format. Table 2 is especially data dense.
The authors conclude with the following:
"In summary, these findings underscore the importance of both ADH and ALH as premalignant lesions arising in an
altered tissue bed. The affected breast is at especially high risk for breast cancer in the first 5 years after diagnosis of breast cancer, with risk remaining elevated in both breasts long term. Both ADH and ALH portend risk for DCIS and invasive breast cancers, predominantly ductal, with two thirds moderate or high grade. These longitudinal data can help to inform clinical management strategies."
Although I think this is a very good study, I do have concerns that clinicians will interpret it to mean that ALH should treated more aggressively. I would argue that it rather suggests that ADH be treated more like ALH. The other problem(s), of course, are the distinctions between ALH & LCIS, and ADH & DCIS. A topic we've dealt with before.
Monday, January 27, 2014
In the past I have discoursed (ranted) in this blog space about pathologists' sins, primarily relating to the overuse of immunohistochemistry. The result of this practice has been across the board, "draconian" cuts in 83342 reimbursements, following on the heals of last year's cuts in 88305's. It could (and has) been argued that we brought this on ourselves.
However, we can take some perhaps perverse comfort in the fact that pathologists are real pikers when it comes to the occasional practice of procedure padding. I refer you to the following article in a recent issue of the New York Times
dealing with unnecessary procedures and charges in the world of dermatology.
The story focuses on a history professor at the University of Central Arkansas who saw a dermatologist for a small lesion on her cheek which turned out (I think) to be a very small basal cell carcinoma. By the time she had undergone resection with Mohs surgery and a plastic surgery closure the resultant bill topped $25,000! The charge for the plastic surgery closure alone (two stitches) was over $14,000! To say that the article takes a dim view of this approach and the physicians who follow such therapeutic overkill would be a major understatement.
As the article points out, procedures no matter how small are the big money makers in the medical world today. The easier they are and the faster you can perform them, the more likely you are to make large amounts of money. A good cardiac surgeon can perform, at most, two bypass surgeries a day but other specialists can do dozens of small procedures (biopsies, excisions, endoscopies, etc) in the same amount of time, often generating even more revenue than a busy heart surgeon.
Although the article deals in part with the rise and unnecessary use in many cases of Mohs surgery and the dramatic rise in dermatologist incomes in the last decade, they are not the only ones in the "cross hairs." The article notes, for example that oncologists make a substantial part of their income based on their ability to mark up chemotherapy charges. Urologists, the article mentions, often make 50% or more of their income from radiation administering equipment that they own, or from laboratories on their premises.
Those of us who are salaried academic physicians have been rather insulated from this phenomenon. It's undoubtedly true that patients in academic hospitals sometimes get more studies and procedures than they absolutely need due to a variety of causes including system inefficiencies, obsessive compulsive academic clinicians, trainees ordering unnecessary tests, intellectual curiosity, etc. At least phycician greed isn't a motive in the salaried physician model. Hospital administrators may be another story!
The vast majority of physicians of all types are dedicated, ethical, honest individuals who have devoted their life, often at considerable personal sacrifice, to help the sick. It's unfortunate that some really "bad apples" seem to be getting most (or all) of the press these days.
Friday, January 03, 2014
I hope that everyone had an enjoyable holiday season and has a happy new year.
Genetic testing is becoming quite a rage, at least in certain demographic segments of the population. Magazines and even television advertisements market these products. Take a buccal swab from yourself or your dog and you can find out your (or your dog's) lineage. Take another one from yourself, mail it off to a the right lab and, at least according to the advertising hype, you can find out your risk for all sorts of diseases. Sound too good to be true? Well, at least for now it IS too good to be true. I would refer you to a very interesting article in the December 30th issue of the New York Times entitled, "I had my DNA picture taken with varying results."
The author, Kira Peikoff is a 28-year-old writer who decided to send DNA specimens to three labs that claimed to provide disease risk analysis on the basis of a DNA sample. The specimens went to 23andMe, Genetic Testing Laboratories, and Pathway Genomics. The results were quite enlightening, in an unexpected way, and perhaps not too dissimilar from visiting several fortune tellers or palm readers with diverse crystal balls or tea leaves. As just one example of the divergence of findings, 23andMe indicated elevated risks compared to the general population for psoriasis and rheumatoid arthritis, while Genetic Testing Laboratories indicated the lowest risk for exactly these two diseases!
It's clear that whole genome sequencing could give anyone a LOT of information about genetic predispositions, although even there good controls do not exist and the role of nature v. nuture is complex for many disease entities. However, these marketed tests are looking at only single-nucleotide polymorphisms (SNPs), and often not the same SNPs for each disease predilection. As Dr. Robert Klitzman points out in the article, this is like trying to understand a book by reading the first letter on each page.
This is an area that isn't going to go away and is one that the FDA is increasingly concerned about. 23andMe received a strong warning from the FDA regarding concerns about the accuracy of its testing system. Pathway was blocked by the FDA from selling its DNA kits in Wahlgrens. Class action suits have been filed against several companies for allegedly meaningless results.
Ms. Peikoff's article is a short and enjoyable read. I recommend it. By the way, even though human disease risk prediction based on SNPs is a total crap shoot, at least at the moment, the pet lineage testing works rather well if you want to know what went into the makeup of your mongrel. Human lineage testing is also reported to work reasonably well.
Thursday, December 19, 2013
Intraoperative consultations (frozen sections, gross interpretations, cytologic specimens) are made in a time sensitive (ie. hurried) fashion, often with less than complete information, less than adequate material and without the aid of ancillary studies. They are then subjected to "Monday morning quarterbacking" by the individual dealing with the subsequent permanents. Under these circumstances it is surprising that we do as well as we do. A large part of resident training at my own institution is devoted to "thinking like a surgeon" in this setting and providing what the surgeon needs to know, no more and no less. Residents often want to hit a home run when only a single is needed. As an example, frozen sections on salivary gland tumors often yield tissue that cannot be definitively interpreted, but this is really not necessary. The surgeon really wants to know if the tissue is overtly malignant and/or high grade (adenoid cystic carcinoma, salivary duct carcinoma, etc), or is it a low-grade neoplasm. The former diagnosis is likely to lead to sacrifice of the facial nerve and the latter will lead to nerve sparing if at all possible. It isn't necessary to sort out acinic cell carcinoma, low-grade mucoepidermoid carcinoma, mixed tumor, etc. on a frozen.
There is a related issue, however, that is nicely reviewed in an article and editorial in a recent issue of Am. J. Clin. Pathol. The article, by Talmon and colleagues (AJCP 2013;140:651-7) discusses miscommunication between pathologists and surgeons at the time of frozen section, based on what the pathologist wrote on the frozen section report, and what the surgeon wrote in the operative note. There is an accompanying editorial by Andy Renshaw (AJCP 2013;140:608-9). One of the findings, dealing with the use of the word "no" in diagnoses is one with which I am personally familiar. It is all to easy for the phrase, "no carcinoma" to be miscommunicated as "carcinoma." Conversely, "adenocarcinoma" can be miscommunciated as "no carcinoma." ...they sound almost the same.
One of my colleagues at U.Va. once had the experience of relaying a frozen diagnosis over the phone, through a nurse, of "adenocarcinoma" and then listening with a mixture of horror and amusement as the nurse said, "ain't no carcinoma." to the surgeon! Following this and several other minor miscommunications, we instituted the rule that frozens could only be called back to a physican and the physician must give his name for the record and repeat the diagnosis to the pathologist, along with the patient's name. Although we got minor grumbling about this for a few weeks, all surgery staff rather quickly saw what we were trying to accomplish and came onboard.
I agree with the article authors and editorial that mistakes centered around the word "no" should be avoided by eliminating the use of "no" in frozen section or other intraoperative communications. "Negative for malignancy" is less likely to be miscommunicated than "no malignancy."
As a second finding the authors noted that pathologists were very reluctant to use the word "benign" in intraoperative communications and yet this powerful word is well understood by all physicians and should be generously used in appropriate circumstances. Other words including "suspicous" and "deferred" are perfectly acceptable and should be used in place of long descriptions likely to create further confusion.
The article and editorial are an excellent review of miscommunication in one direction, from pathologist to surgeon. Of course, miscommunications happen in the other direction as well and these can lead to frozen section errors based on faulty clinical or intraoperative information. I'm always delighted when the surgeon comes into the pathology suite during or, even better, before the interpretation of the slide. This is a fantastic time to ask them questions and review the clinical and operative findings. I can't count the number of times a mass submitted as "colonic" or "uterine" on the requisition form turned out to be "ovarian" (or vice versa) when the surgeon was present to describe the operative findings. Surgeons who routinely do this gain a better understanding of pathology and its intraoperative limits, we gain a better understanding of the surgical issues, and the patient inevitably benefits.