BACKGROUND:The authors validated the Royal Marsden Hospital (RMH) prognostic score in patients with advanced lung, pancreatic, and head and neck cancers who were enrolled on phase 1 trials in the MD Anderson Cancer Center Phase I Clinical Trials Program. METHODS: The RMH score uses albumin ( 3.5 g/dL vs <3.5 g/dL), lactate dehydrogenase (less than or equal to the upper limit of normal [ ULN] vs >ULN), and the number of metastatic sites ( 2 sites vs 3 sites) to predict patient survival in phase 1 trials. The authors of this report retrospectively reviewed the outcomes of 229 consecutive patients with lung, pancreatic, and head and neck tumors who were treated on 57 phase 1 trials. RESULTS: Two hundred twenty-nine consecutive patients with lung cancer (N ¼ 85), pancreatic cancer (N ¼ 83), and head and neck tumors (N ¼ 61) were treated. The median patient age was 60 years (range, 26-85 years), and 63% of the patients were men. Patients with a good RMH prognostic score (0-1) at baseline had a longer median survival than patients with a poor prognostic score (2-3; 33.9 weeks vs 21.1 weeks; P < .0001). The RMH score was an independent variable that predicted survival in multivariate analysis. Other independent variables that predicted better survival were hemoglobin level ( 10.5 g/dL), Eastern Cooperative Oncology Group performance status (0-1), and tumor type. Patients who were treated on first-in-human trials did not fare worse compared with those who were not treated on first-in-human trials. CONCLUSIONS: For patients with lung, pancreatic, and head and neck tumors who were treated on phase 1 trials, survival was predicted accurately by the RMH prognostic score.