THE TREATMENT OF PATIENTS WITH RECURRENT BRAIN METASTASES - A RETROSPECTIVE ANALYSIS OF 109 PATIENTS WITH NONSMALL CELL LUNG-CANCER

Background. Brain metastases represent a major source of morbidity in patients with cancer. Methods. Treatment outcomes were analyzed retros pectively in 214 patients with brain metastases from nonsmall cell lun g cancer (NSCLC) who underwent resection at Memorial Hospital (New Yor k, NY) between January, 1976, and December, 1990. Results. The study g roup included 109 patients (51%) with symptomatic recurrent brain tumo rs (median, 5.0 months after complete resection). Recurrence in the br ain was at the original site in 62% of patients and at other sites in 38%. The median survival (MS) was 11.3 months in the recurrence group compared with 9.5 months (P < 0.5) in the nonrecurrence group (N = 105 ), Thirty-two patients had further surgery after recurrence; their med ian relapse time was 5.2 months, In these patients, survival (MS, 15.0 months) calculated from the time of their first operation, was signif icantly different (P < 0.001) from that of patients who did not underg o a second procedure (N = 77) (MS, 10.0 months), In the 32 patients wh o underwent reoperation, MS from the time of the second operation was 10 months, whereas the median interval from the first operation was 5 months (average, 5.7 months). Eight of these 32 patients had a third o peration, after a median relapse time of 4 months; their MS was 42 mon ths. There was a significant difference (P < 0.02) between the MS of 3 9 patients synchronously diagnosed with lung cancer and brain metastas is (MS, 9.0 months) and 70 metachronously diagnosed patients (MS, 14.6 months). Women (N = 55) survived longer than men (N = 54) (14.4 month s vs. 9.7 months, P < 0.01). Univariate analysis showed that histology , disease stage, and completeness of resection of the primary tumor al so affected survival (P < 0.02, P < 0.014, and P < 0.001, respectively ), Although no significant difference was found between survival of pa tients with recurrence in the supratentorial space and patients with r ecurrence in the posterior fossa (MS, 11.4 months vs. 11.2 months, P < 0.13), no one from the latter subgroup survived 3 years. Conclusions. If technically feasible, further surgery is effective in prolonging t he survival of patients with NSCLC who have recurring brain metastases .