Evaluation of an intensive strategy for follow-up and surveillance of primary breast cancer.

Abstract

Controversies over the frequency and intensity of the follow-up care of breast cancer patients exist. Some physicians have adopted an intensive approach to follow-up care that consists of frequent laboratory tests and routine imaging studies, including chest radiographs, bone scans, and CT scans, whereas others have established a minimalist approach consisting of only history, physical examinations, and mammograms. Our objective was to evaluate the role of intensive follow-up on detection of breast cancer recurrence and to examine the impact of follow-up on overall survival. During a 10-year period (1986-1996), 129 patients with recurrent disease were identified from a prospective database of 1898 breast cancer patients. The patients with recurrent disease were divided into minimalist or intensive groups according to method of detection. Twenty-seven of 126 (21%) patients were assigned to the intensive method of detection group (LFT, CEA, CA 15-3, chest radiograph, CT scan, and bone scan); 99 of 126 (79%) patients were assigned to the minimal detection group (history, physical examination, and mammography). Distant disease to the bone was the most common initial tumor recurrence, at 27%. History, physical examination, and mammography detected recurrent cancer in approximately the same amount of time as LFTs, tumor markers, CT scans, and chest radiographs (P = .960). When the recurrent patients were divided into intensive and minimalist groups and analyzed by time to detection of recurrence, there was no significant difference between the time to detection in those recurrences detected by intensive methods and those recurrences detected by minimalist methods (P = .95). The independent variables age, tumor size, type of surgery, number of positive nodes, time to recurrence, method of detection, and site of recurrence (regional or distant) were subject to univariate and multivariate analysis by the Cox proportional hazards model. Only two variables had an impact on survival by multivariate analysis: early timing of the recurrence (P = .0011) and the site of the recurrence (P = .02). Timing was defined as early (< or =365 days from the time of diagnosis to recurrence) or late (> or =365 days from the time of diagnosis to recurrence). Early recurrence was the first variable found to be significant on stepwise forward regression analysis. The primary site of recurrence was significant at step two. The method of detection--intensive or minimal--did not significantly affect survival (P = .18). There is no survival benefit to routine intensive follow-up regimens in detecting recurrent breast cancer. Expensive diagnostic tests such as bone scans, CT scans, and serial tumor markers are best used for detection of metastasis in symptomatic patients.