The July 19th issue of the New York Times
contains an article that should be must reading for surgical pathologists. It deals with a woman who was originally diagnosed as having low-grade (solid, cribriform) DCIS of the breast and underwent a quadrantectomy and radiation therapy. Subsequent review changed the diagnosis to something less, presumably usual-type ductal hyperplasia.
The article deals in large part with the difficulties in diagnosing these low-grade in-situ lesions but suggests, or at least implies that the problem is largely one of expertise. Enter the College of American Pathologists with its highly controversial proposal for "voluntary" certification in breast pathology, and enter the federal government with a funded study on diagnosis of breast lesions, suggesting a 17% error rate for DCIS.
Few would deny that there are occasional inept pathologists out there and the pathologist in this case was not board certified at the time of this (mis)diagnosis, but the available data suggest that "experts" in breast pathology disagree with regard to these borderline in-situ lesions at almost the same rate as qualified community pathologists. It is true that education, well-articulated criteria and experience will help lower the error rate, but they will certainly not eliminate it. Indeed, defining "error" in these gray-zone cases may be difficult or impossible. I believe that the CAP's "voluntary certification" program is misdirected and will not solve the problem. Moreover, with help from lawyers and insurance companies it is likely to start us down the wrong road to multi-organ certification mandates.
To paraphrase my friend and colleague, Dr. Mark Stoler, the current president of the ASCP, although subspecialty education and more importantly second opinions can help minimize these diagnostic variations, education or certification will not eliminate them from the subspecialty of breast pathology or any other subspecialty. Rather, clinical medicine must understand and acknowledge the issues and realities of some borderline (subjective) pathologic diagnoses and modify their approaches to patient therapy that have diagnoses in these gray-zones. There is no biologic reason to believe that ADH and DCIS are distinctly different entities and, to put it plainly, the idea that clinicians should radically treat low-grade DCIS as opposed to ADH is just as much or more a part of the problem raised by the NY Times article than any perception of pathologist error.
We are of course at least partly responsible for this situation because we have given these two very closely related lesions, ADH and DCIS, such radically different names. Patients and reporters are often ill-equipped to distinguish carcinoma in-situ from any form of invasive carcinoma. The power of that word should never be under estimated, and we should have devised a more judicious term for these non-invasive, non-life-threatening but pre-cancerous lesions.