OBJECTIVES: Recent studies suggest that standard uptake value on fluorodeoxyglucose positron emission tomography scan can predict mediastinal lymph node status in malignant pleural mesothelioma. Because mediastinal nodal metastasis is known to be associated with poor prognosis, we hypothesized that standard uptake value on fluorodeoxyglucose positron emission tomography might independently predict survival. METHODS: Patients with pathologically proven mesothelioma underwent fluorodeoxyglucose positron emission tomography scanning. Patients fasted and received a minimum of 10 mCi of F18-fluorodeoxyglucose. Whole-body emission studies were acquired, followed by whole-body transmission scans with iterative reconstruction. On the basis of the maximal chi-square method, a standard uptake value of 10 was chosen to classify patients as low versus high standard uptake value. Survival probabilities for both standard uptake value groups were estimated by the Kaplan-Meier method. A Cox proportional hazards model assessed the joint influence of standard uptake value, histology, and stage on survival. RESULTS: From 1998 to 2005, 137 patients with malignant pleural mesothelioma underwent positron emission tomography scans. The median follow-up for all surviving patients was 24 months. Median survivals were 9 and 21 months for the high and low standard uptake value groups, respectively (P = .02). In a multivariable analysis, high standard uptake value tumors were associated with a 1.9 times greater risk of death than low standard uptake value tumors (P < .01). Mixed histology carried a 2.9 times greater risk of death than epithelioid histology (P < .01), and stages III and IV had a 1.8 times greater risk of death than stages I and II (P = .05). CONCLUSIONS: Standard uptake value greater than 10, mixed histology, and stages III and IV are poor risk factors in malignant pleural mesothelioma. These findings suggest that fluorodeoxyglucose positron emission tomography can be used to stratify patients for treatment and clinical trials.
