It is well known that breast pathology is one of the areas within surgical pathology that often yields differences of opinion. This may in part be a result of the numerous factors that now need to be routinely analyzed in patients with breast lesions. As other fields of medicine have become increasingly subspecialized, so too has surgical pathology, and there is an increasing trend toward subspecialty diagnosis of surgical pathology cases in academic settings; that is, breast cancer cases are often signed out only by pathologists with expertise in breast pathology. However, this arrangement is not feasible in all academic centers, let alone community hospitals. Because of the sheer volume of breast surgery and the prevalence of breast cancer, it is not feasible to relegate all breast pathology interpretation to subspecialists. Therefore, second opinions may become critical for a significant number of patients. In fact, many medical centers require (and, we believe, rightly so) pathology review of outside material before surgery. In the article that accompanies this editorial, Kennecke et al examine the impact of routine second-opinion breast pathology review in a group of patients with breast cancer for whom a change in diagnosis would likely impact treatment. Some changes in pathology were not surprising. For example, it is known that histologic grade is among the parameters most discrepant when viewed by multiple pathologists. Nodal and margin status changes are less expected, however, particularly the former. It is surprising that of the 15 upstages from node negative, six were to N1, not just N0 (i ); these upstages occurred, for the most part, without the use of immunohistochemical stains. These patients alone should serve as a wake-up call for treating physicians that certain diagnoses in breast pathology are intrinsically more difficult, occasionally subtle, and prone to differences of opinion. As the authors additionally point out, the use of molecular testing such as Oncotype DX (Genomic Health, Redwood City, CA) has been used adjunctively to complement accurate histologic interpretation; that is, as a way of leveling the playing field vis-a-vis grading and subtyping of invasive carcinoma. Clearly, however, as this study shows, use of such testing is not enough. Parameters such as node status, lymphatic invasion, and margins are not directly measured by such tests, and therefore these sorts of discrepancies would not be accounted for by different test values, yet they remain influential to treatment decisions. In most institutions, negative margins are a requisite for deciding when to proceed with the radiation therapy phase of breast cancer treatment. It is disturbing, but not surprising, that eleven of the twelve cases that involved margin changes were from negative to positive. Depending on individual circumstances, in some practices this might require the patient to be referred back to the surgeon for reexcision. Two patients in the study who were initially diagnosed with invasion were found, on review, to only have ductal carcinoma in situ. As these patients and some of the margin data begin to indicate, discrepancies can go in either direction. In fact, not all initially reported patients with node-positive disease are truly node positive. One of us (I.J.B.) previously reported in Journal of Clinical Oncology a series of patients whose sentinel lymph nodes contained benign epithelial cells that were physiologically transported as a result of biopsyinduced displacement. We subsequently received numerous cases for second opinion in which the nodal status was unexpected, given the pathologic features of the primary tumor. Some of the cases were truly node positive and some were not; however, the reasons for the discrepancies were variable. The study by Kennecke et al specifically selected patients who were initially node negative, not node positive, but such diagnostic issues exist for a certain number of patients with nodepositive disease, as well, with equally important clinical implications. The practice of surgical pathology of the breast is not limited to cancer risk, typing, and prognostication. Pathologists routinely evaluate benign entities that constitute risk factors, such as atypical duct hyperplasia, lobular neoplasia, intraductal papilloma, and radial scar, to name a few. Differences of diagnostic opinion abound for such lesions, perhaps even more so than when grading invasive ductal carcinoma. So, too, have medical oncologists branched out. Management of risk with antihormonal therapy is, for the most part, in the domain of the medical oncologist. Therefore, proper pathologic classification has clinical oncologic implications even in patients without carcinoma. Clearly this study by Kennecke et al has a direct implication for patient care. A correct diagnosis may determine the use of effective treatment (or show particular treatments to be unnecessary), whereas an incorrect diagnosis may lead to treatment-related adverse effects without known benefit. Yet the study by Kennecke et al does not go far enough. Central routine review of a targeted subset of patients with breast cancer in a relatively controlled, closed system of health care is a laudable effort, even if incomplete. Without specifically addressing the JOURNAL OF CLINICAL ONCOLOGY E D I T O R I A L VOLUME 30 NUMBER 18 JUNE 2