DUCTAL CARCINOMA IN-SITU OF THE BREAST - CORRELATION OF PATHOLOGICAL AND MAMMOGRAPHIC FEATURES WITH EXTENT OF DISEASE

Optimal treatment of ductal carcinoma in situ (DCIS) of the breast req uires an improved understanding of its pathologic extent and propensit y for local recurrence. This study was performed to analyze mammograph ic and pathologic features of DCIS that might predict the extent of di sease within the breast and facilitate treatment selection between lum pectomy alone, lumpectomy and radiotherapy, and mastectomy. At our ins titution, 60 cases of DCIS were diagnosed in 59 patients from June 198 5 to February 1995 and form the basis of this retrospective study. Dem ographic and treatment-related information was obtained from hospital and tumor registry records. Mammograms were reviewed and size estimate s of the abnormalities were determined. Pathologic slides from all cas es were reviewed and classified according to size group, focality, nuc lear grade, necrosis, and histologic subtype. DNA ploidy status and pr oliferation indices were available for 28 patients. Pathologically, 43 (72%) cases were <15 mm, 14 (23%) were 16 to 40 mm, and 3 (5%) were > 40 mm. Five (8%) of the lesions were multicentric, 28 (47%) focal, and 27 (45%) multifocal. Thirty-three (55%) patients were treated by mast ectomy, 16 (27%) by lumpectomy alone, and 11 (18%) by lumpectomy and r adiation therapy. Mammographic size, histologic grade, presence or abs ence of necrosis, histologic subtype, DNA ploidy, and proliferative in dex were compared with pathologic size and focality by chi(2) analysis . Mammographic size correlated significantly with pathologic size (chi (2) = 11.3; P = 0.02) but underestimated the extent of disease in 9 ca ses. Although focality correlated significantly with pathologic size ( chi(2) = 15.8; P = 0.003), the remaining histopathologic features did not significantly correlate with pathologic size or focality. Histopat hologic features, including DNA studies, do not reliably predict the p athologic extent of DCIS, but mammographic size and focality do signif icantly correlate with pathologic size. Nevertheless, most cases of DC IS are small focal or multifocal lesions that are amenable to breast c onservation approaches; further studies are needed to determine the ap propriate use of lumpectomy, radiation therapy, and mastectomy in the treatment of DCIS.