ABSTRACT
BACKGROUND: A prognostic model for patients with metastatic urothelial carcinoma (UC) progressing after platinum-based therapy was constructed from data from the phase III vinflunine trial. However, prognostic information for patients treated with other regimens is limited. MATERIALS AND METHODS: We pooled individual patient data from 7 second-line studies and analyzed the influence of factors of interest on overall survival (OS) through univariate and multivariate analysis. A prognostic model was constructed, and data from an independent series were used for validation. RESULTS: The data from 193 patients were pooled. The second-line chemotherapy regimen was single-agent taxane in 54 patients (28%), a platinum-based combination in 47 (24%), and a non-platinum combination in 92 (48%). On multivariate analysis, Eastern Cooperative Oncology Group performance status ≥ 1, hemoglobin < 10 g/dL, and metastatic patterns other than lymph node-only metastasis emerged as independent adverse prognostic factors. Patients with all 3 factors (poor risk), 1 to 2 factors (intermediate risk), and no factors (good risk) had a median OS of 3.1, 8.7, and 16.5 months, respectively (P < .0001). The corresponding median OS for the validation series (n = 44) was 3.3, 8.1, and 13.3 months (P = .023). Furthermore, platinum-based regimens were independently associated with an OS benefit compared with other regimens (hazard ratio, 0.31; 95% confidence interval, 0.18-0.53; P < .0001). CONCLUSION: We have proposed and validated a prognostic model for patients with metastatic UC who were eligible for second-line therapy. The proposed model could prove helpful for risk stratification. Furthermore, our data suggest that testing second-line platinum-based regimens in randomized trials is warranted.
