Pathologic practice standards for breast carcinoma: tumor size, reliable data, or miscues?

Abstract

Currently, when the majority of newly diagnosed breast carcinomas are increasingly node-negative, therapeutic decisions will necessarily be based on the characteristics of the primary tumor. Of those pathologic features, the most important for disease-free and cause-specific survival is tumor size. Tumor size is one of the essential elements of the TNM staging system, and for the small mammographically detected breast cancer without axillary nodal or distant disease, the only measurable variable of consequence. In this issue, Wilkinson and colleagues document a low level of compliance with the College of American Pathologists (CAP) guidelines for the reporting of invasive breast carcinomas in a comparison of pathology reports from the community and their own referral center. Notably, tumor sizing is one of the most common features that is not compliant with the CAP recommendations. The CAP guidelines require measurements of the gross tumor in three dimensions, and the maximum tumor size as determined microscopically, but the reported deficiencies in compliance with the guidelines do not begin to measure the problem. Tumor size as determined by a pathologist from gross examination of a specimen can misrepresent actual tumor size dramatically, both by underestimating the true extent of an invasive carcinoma and by overestimation of its size by inclusion of contiguous areas of either noninvasive carcinoma, proliferative breast disease, or postbiopsy reactive changes. Also, the size recorded by the pathologist, either estimated grossly or measured microscopically, is highly dependent on the manner of sectioning of the gross specimen. Depending on the plane of section, as much as a threefold difference in recorded size can result. A recent consensus conference has noted that pathologic tumor size should be correlated with available preoperative imaging to reduce the errors of underand overestimation. In my experience, there have been many occasions when size determined by imaging has been ignored and treatment decisions were based on an erroneous size reported by the pathologist. Such discrepancies in either direction should be resolved by reevaluation before determining therapy. Tumor size would seem to be a prognostic feature both easily established and very reproducible, but clearly in current practice it is not. Small differences in tumor size that straddle the arbitrary boundaries of the T-size classification, eg, a low-grade, nodenegative invasive ductal carcinoma of 10 mm (T1b N0) versus the identical tumor at 11 mm maximum size (T1c N0), can precipitate remarkably different therapeutic recommendations and estimates of prognosis. The authors note, “Not all pathology reports contain the information needed for clinical decision-making.” Efforts to correct the deficiencies in recording the prognostic features and improvement in the accuracy of the data recorded should both be priorities. Such deficiencies not only limit therapeutic decisions for individual patients, but also impact the interpretation of some clinical trials: the results and recommendations of the NSABP B20 trial were based on clinical size alone, and the pathology results of B17 were to a substantial degree based on clinical and gross pathologic estimates of the extent of the ductal carcinoma in situ. Unlike Wilkinson and colleagues, I am not so sanguine about the possibility that passive distribution of CAP guidelines to practicing pathologists will result in substantial change.