OBJECTIVES: To systematically review the published literature to identify effectiveness data of first-line (1L) treatments for patients with advanced or metastatic renal cell carcinoma (a/mRCC) from real-world evidence (RWE) studies.

METHODS: A systematic literature review (SLR) was conducted. Publications of interest were restricted to RWE studies on naïve patients with a/mRCC, focusing on intermediate to poor risk patients according to MSKCC/IMDC criteria. Relevant databases including MEDLINE, Embase, registries and conference proceedings were searched up to May 22, 2017.

RESULTS: A total of 31 relevant publications were included, of which 27 were retrospective studies and 4 prospective studies. Retrospective studies were based on registries (5 studies), institutional database (4 studies) or patients’ medical records review (18 studies).Treatments assessed were reported either as a specific treatment (e.g. sunitinib, pazopanib, sorafenib, everolimus) or as treatments classes (e.g. immunotherapy, targeted therapy). Sample size ranged between 21 to 744 patients, seven studies included more than 300 patients. Median OS (mOS) in MSKCC intermediate risk (IR) patients ranged between 9 and 41 months. In MSKCC poor risk (PR) patients, mOS ranged between 4 and 24 months. Median progression-free survival (PFS) in MSKCC IR patients ranged between 3 and 25 months, whilst PFS in MSKCC PR patients ranged between 2 and 38 months. Results stratified by IMDC scores were also reported. Heterogeneity across studies was due mainly to varied patient characteristics.

CONCLUSIONS: This SLR provides a comprehensive synthesis of RWE sources assessing the efficacy of 1L treatments for a/mRCC in intermediate to poor risk patients according to MSKCC/IMDC criteria in the RWE setting. The evidence suggests that there is considerable unmet need for improved survival outcomes in this population, hence new treatments providing superior outcomes are warranted. This strong dataset of real-world efficacy results was used in a cost-effectiveness model in order to validate survival extrapolations.