Margin status is an important prognostic factor for local recurrence after partial mastectomy for breast malignancy. Options for intraoperative evaluation of margin status include gross examination of the specimen, frozen section, and "touch preparation" cytology. This study evaluates the accuracy of gross examination without other intraoperative pathological analysis as a method of determining margin status. Records of 254 consecutive patients undergoing partial mastectomy for 255 breast malignancies (199 invasive, 56 DCIS) over 6 years were analyzed retrospectively. All women underwent en bloc excision of the primary lesion with gross examination of margin status by the surgeon and pathologist. All suspicious areas were reexcised, and the specimen was inked, serially sectioned at 2-3 mm intervals and examined with hematoxylin and eosin (H&E) stains. Specimens with tumor <2 mm from a margin were considered margin-positive and those with all tumor > or =2 mm from the margin were designated margin-negative. One hundred fourteen (45%) of the 255 segmental resections were considered to have grossly tumor-free margins, and intraoperative reexcision was not performed. Ninety-six (84%) of these specimens had histologically negative margins. Gross examination prompted intraoperative reexcision in 141 (55%) cases. Ninety-five (67%) of these 141 resections had tumor-free margins on histopathology. Overall, the final margin was involved in 64 of the 255 partial mastectomies. Seventeen (27%) women with initially margin-positive resections underwent mastectomy, while 46 (72%) underwent reexcision, which was margin-negative in 41 (89%). After a median follow-up of 42 months, there have been eight (3.5%) local recurrences. The initial margin-positive rate was similar in ductal carcinoma in situ (DCIS) (30%) and invasive carcinoma (24%). Margin status was correlated with nodal status; there was no correlation with age, tumor size, grade hormone receptor status, or type of diagnostic biopsy. Gross examination of the resection specimen does not reflect margin status in at least 25 per cent of women undergoing partial mastectomy for breast malignancy. Other techniques for evaluation of margin status should be considered to reduce the need for reexcision of involved margins. We are currently designing a prospective clinical trial to examine the efficacy of new techniques for intraoperative evaluation of margin status.