The value of breast lumpectomy margin assessment as a predictor of residual tumor burden

Purpose: Margin assessment is commonly used as a guide to the relative aggressiveness of therapy for breast conserving treatment (BCT), though its value as a predictor of the presence, type, or extent of residual tumor has not been conclusively studied. Controversy continues to exist as to what constitutes a margin that is ''positive,'' ''close,'' or ''negative.'' We attempt to address these issues through an analysis of re-excision specimens. Patients and Methods: As part of an institutional prospective practice approach for BCT, 265 cases with ATCC Stage I/II carcinoma with an initial excision margin that was less than or equal to 2 mm or indeterminate were subjected to re-excision. The probability of residual tumor (+RE) was evaluated with respect to tumor size, histopathologic subtype, relative closeness of the measured margin, the extent of margin positivity graded as focal, minimal, moderate, or extensive, and the extent of specimen processing as reflected in the number of cut sections per specimen volume (S:V ratio). The amount of residual tumor was graded as microscopic, small, medium, or large. The histopathologic subtype of tumor in the re-excision specimen was classified as having an invasive component (ICa) or pure DCIS (DCIS). Results: The primary excision margin was positive, >0 less than or equal to 1 mm, 1.1-2 mm, and indeterminate in 60%, 18%, 5%, and 17%, respectively. The predominant histopathologies in the initial excision specimens were invasive ductal (IDC) (50%) and tumors with an extensive intraductal component (EIC) (43%). The histopathology of the initial excision specimen was highly predictive of the histopathology of tumor found on re-excision, as residual DCIS was found in 60% of +RE specimens with initial histopathology of EIC compared to 26% for IDC (p = 0.001). Neither the extent of margin positivity nor the extent of tumor in the re-excision were significantly related to the initial histopathologic subtype; however, a +RE was seen in 59% of EIC, 43% of IDC, and 32% of invasive lobular ILC cases (p = 0.01). The extent of margin positivity was significantly related to the size of the tumor such that tumor size less than or equal to 20 mm was associated with a greater probability of focal or minimal margin involvement. Positive margins graded as focal, minimal, moderate/extensive were associated with a +RE in 26%, 58%, and 84%, respectively (p = 0.001). Further, the extent of positivity was significantly correlated with the extent of residual tumor such that focal/minimal positivity was more commonly associated with micro/small +RE, whereas moderate/extensive positivity was associated with medium/large +RE. When the closest margin of the initial excision specimen was positive, >0 less than or equal to 1 mm, or 1.1-2 mm, a +RE was found in 56%, 41%, and 17%, respectively (p = 0.01) but did not correlate with the amount of residual tumor. The extent of specimen processing as reflected in the S:V ratio did not correlate with the probability of defining a measured margin as positive nor the probability of a +RE. In a univariate model, the extent of tumor in the re-excision and the histologic type of tumor in the re-excision were significantly associated with margin status and initial histopathology, respectively. The probability of finding a +RE, based on a multivariate model, was associated with the closeness and extent of margin involvement and initial histopathology of an EIC. Conclusion: The relative closeness of tumor to the specimen edge and the extent of margin positivity are predictive for residual tumor, though with an error consistent with its limitations as a sampling procedure. The histopathology of tumor in the initial excision is predictive of the type of residual tumor and the extent of margin positivity was correlated with the amount of residual tumor. (C) 1997 Elsevier Science Inc.