A study was performed to determine if prognostic factors could be used preoperatively to predict outcome following resection of metastases. Sixty-seven soft tissue sarcoma (STS) patients (median follow-up, 36 months) and 39 osteogenic sarcoma patients (OGS) (median follow-up, 29 months) underwent thoracic exploration at the first indication of pulmonary metastases, and the results for each group were reviewed. The number of metastatic nodules, disease-free interval (DFI), and tumor doubling time (TDT) significantly correlated with postoperative survival for STS patients. Patients with four or fewer nodules on preoperative linear tomograms survived longer (median, 23 months) than patients with more than four nodules (median, 6 months; P less than 0.005). Patients with a DFI greater than 12 months had a longer survival (median, 30 months) than patients with a DFI less than or equal to 12 months (median, 10 months; P less than 0.005). Patients with a TDT greater than or equal to 20 days had a longer survival (median, 22 months) than patients with a TDT less than 20 days (median, 6 months; P less than 0.005). The only significant predictor of survival for OGS patients was the number of nodules on preoperative linear tomograms (less than or equal to 4, 37 months median survival; greater than 4, 10 months median survival; P less than 0.05). This was due to significant differences noted for the DFI and TDT distributions between OGS and STS patients, with most OGS patients having a short DFI (less than or equal to 12 months) and a rapid TDT (less than or equal to 20 days) whereas STS patients had a more heterogeneous distribution (P less than 0.01). Thus, the number of metastases visible on the preoperative tomogram was the best predictor of survival for both OGS and STS patients. However, the applicability of other prognostic factors could not be generalized for these two closely related groups of patients.