Forty-nine women with ductal carcinoma in situ (DCIS) treated with lumpectomy and irradiation were studied retrospectively. The median age was 50 years (range, 29 to 73 years) and the median follow-up time from initiation of therapy was 86 months (range, 17 to 230 months). Twelve patients presented with palpable masses (0.4 to 4 cm), three with breast thickening, and three with nipple discharge. In 31 patients the tumors were detected by mammography. Intraoperatively, excision of lesions was confirmed by specimen xray (38 specimens) or gross inspection (five specimens) and was recorded to be complete. No record was available in the other six patients. Margins of excision free of DCIS were microscopically confirmed in 25 specimens. The size of impalpable DCIS lesions recorded in 25 patients ranged from 0.4 to 5.0 cm (mean, 1.5 cm). Using Lagios' classification system, there were 18 classic comedocarcinomas, high nuclear grade (NG) with necrosis; seven cribriform/ papillary, high NG with necrosis; 17 cribriform/micropapillary, intermediate NG with or without necrosis; and seven cribriform/micropapillary, low NG without necrosis. In two patients residual malignant calcifications were present on the postoperative mammogram. Disease recurred in the treated breast at the site of incision in five patients at 18 months and 8, 11, and 12 (two patients) years from initial therapy. The rate of local disease recurrence was 2% at 5 years and 6% at 10 years; three recurrences showed invasive ductal carcinoma and two were DCIS. To evaluate risk factors the following characteristics were considered: necrosis, NG, histological type, periductal fibrosis, periductal lymphoid infiltrate, margin status, age, and method of tumor detection. The end points chosen were recurrence and death from any cause (because only one patient died of disease). Although the recurrences were attributed to residual disease in two patients, of the clinical and pathological parameters evaluated, only periductal fibrosis showed a significant relationship with outcome, with a P value ≤ .05 by the Wilcoxon test. On the other hand, using the proportional hazards model, necrosis was a significant predictor for recurrence ( P = .02), as was the pair fibrosis and tumor detection when taken together ( P = .05). Fibrosis significantly associated with high NG, Lagios' histological subtypes I and II, periductal lymphoid infiltrate, and necrosis ( P ≤ .0006). Our study indicates that (1) complete excision, based on specimen radiological/mammographic correlation, microscopic evaluation of resection margins, and postexcision mammogram, is an important determinant of local control; (2) determination of high risk factors (ie, periductal fibrosis, necrosis, and tumor detection) is important in predicting local failure; and (3) delayed tumor recurrence may be seen in high risk DCIS after radiation therapy (XRT).
Read full abstract