PUK3 Synthesizing Evidence on Overall Survival and Assessing the Feasibility of Network Meta-Analyses for Previously Untreated Advanced/Metastatic Renal CELL Carcinoma Therapies

The treatment landscape for previously untreated advanced/metastatic renal cell carcinoma (aRCC) is rapidly evolving. With this, knowledge concerning the comparability of clinical efficacy across interventions is essential beyond the available head-to-head comparisons, which mostly only include sunitinib as a comparator. This study investigates the validity of conducting a network meta-analysis (NMA) to compare overall survival (OS) in nivolumab+cabozantinib versus all relevant interventions. A systematic literature review identified all available randomized controlled trials (RCTs) in aRCC. Available evidence was synthesized by evaluating whether the set of pre-defined relevant interventions (n=12) could form a connected network of evidence. Clinical heterogeneity was assessed for study and population characteristics as well as outcome and treatment characteristics. Finally, inconsistency in observed treatment effects was evaluated. The SLR identified 18 individual RCTs that met our inclusion criteria. Imbalances were found in the prognostic risk status at baseline. As this is a potential treatment effect modifier, networks of evidence were clustered by all-risk, intermediate-/poor-risk and favorable-risk. Non-relevant interventions were only included when required to connect the network. The all-risk network included 8 studies connecting 7 relevant interventions. The intermediate-/poor-risk network included 10 studies connecting 8 interventions. Only 3 studies were included in the favorable-risk network (4 relevant interventions). The majority of population characteristics were inconsistently reported; therefore, imbalances could not be assessed. However, the reported characteristics showed heterogeneity in ECOG score, nephrectomy, and study design in the all-risk and intermediate-/poor-risk networks. The few direct comparisons in the network supported by multiple studies did not show significant differences in their treatment effects. While it is feasible to perform a NMA to determine the comparative efficacy of relevant interventions on OS in aRCC, results must be interpreted with caution because unobservable heterogeneity may compromise the validity of the results.