The selection of objective criteria that can reliably predict survival in patients undergoing resection of pulmonary metastases remains controversial. Between 1974 and 1982, 487 patients with soft tissue sarcomas presented to the National Cancer Institute. Eighty patients underwent thoracotomy for putative metastases and 67 patients had histologically proved pulmonary metastases. The 3 year tumor-free survival rate was 30% by actuarial analysis. Patients with resectable metastases had significantly prolonged post-thoracotomy survival compared to those patients with unresectable metastases. The most significant preoperative predictors of survival were the tumor doubling time, the number of metastases on preoperative linear chest tomograms, and the disease-free interval. Patients with a tumor doubling time of 20 days or more had a significantly longer post-thoracotomy survival (22 months median) than patients with a tumor doubling time of less than 20 days (6 months median). Those patients with four nodules or less on preoperative tomograms had significantly longer post-thoracotomy survival times (23 months median) than those patients with more than four nodules (6 months median). Patients with a disease-free interval of more than 12 months had a longer post-thoracotomy survival (32 months median) than patients with a disease-free interval of 12 months or less (10 months median). Combining these three prognostic factors significantly increased the predictive ability of this model. These criteria provide an accurate and rapid method to identify preoperatively those patients who will maximally benefit from resection of pulmonary metastases from soft tissue sarcomas.