Abstract / Summary
BACKGROUND The optimal combination of surgical and/or medical therapy for patients with malignant pleural mesothelioma is controversial. The MARS1 and MARS2 trials found worse outcomes after surgery plus chemotherapy compared to chemotherapy alone. However, there are concerns about the validity of the findings because of the trials' small sample sizes and patient selection. METHODS We used a novel method for sensitivity analysis, the Robustness of Inferences to Replacement (RIR), to evaluate the main findings of MARS1 and MARS2. For a trial finding worse survival after surgery plus chemotherapy versus chemotherapy alone, the RIR calculates how many patients would need to be resampled/replaced from a hypothetical "null hypothesis" population without any difference in survival between treatment groups for the result to be reversed. RESULTS MARS1 found worse survival after chemotherapy plus extrapleural pneumonectomy versus chemotherapy alone. The RIR was 8, meaning that 8 of 50 patients (16%) would need to be replaced for the finding to be reversed so there was no survival difference. MARS2 found worse survival after chemotherapy plus extended pleurectomy and decortication versus chemotherapy alone. The RIR was 71, meaning that 71 of 335 patients (21%) would need to be replaced for the finding to be reversed. CONCLUSIONS The RIR analysis sheds doubt on the statistical robustness of the trial findings, since the results would be reversed if as few as 15-20% of the patients had better survival. Additional studies may come to opposite conclusions favoring surgery. More trials are needed to definitively guide treatment for this disease.