The aim of this study was to test the validity of the hypothesis that patients in whom brain metastasis is the first indication of an undiagnosed primary tumor have a better chance of survival than similar patients with a known primary lesion.
Between January 1983 and December 1998, 342 patients with computed tomography–diagnosed brain metastases were treated at a single institution. Information on potential prognostic factors, including primary diagnosis status, was collected retrospectively. Univariate and multivariate analyses were performed to identify prognostic factors related to survival.
Survival was not statistically different between patients with an undiagnosed primary (UDP) lesion and those with a diagnosed primary (DP) tumor (6 and 4.5 months, respectively; p = 0.097). In the UDP group (122 patients [36%]), survival was not affected by the eventual identification of the primary disease (p = 0.905). The median survival for the entire population was 5.2 months, with 1-, 2-, and 3-year survival rates of 25, 11, and 4%, respectively. Prognostic factors for the overall population included treatment (p < 0.0001), an age less than 65 years (p = 0.004), discharge status (p < 0.001), absence of systemic metastasis (p = 0.036), and asymptomatic cerebral metastasis (p = 0.05).
Treatment modality was the most significant independent variable affecting survival in patients with brain metastases. Eventual identification of a primary tumor does not affect overall survival; therefore, delaying therapeutic intervention in pursuit of a primary diagnosis may not be appropriate. Data in this study failed to demonstrate a statistically significant difference in survival between patients with UDP and those with DP lesions, on first presenting with brain metastases.