Background: Despite the survival advantage, not all metastatic renal cell carcinoma (mRCC) patients achieve a long-term benefit from immunotherapy. Moreover, the identification of prognostic biomarkers is still an unmet clinical need.
Methods: This multicenter retrospective study investigated the prognostic role of peripheral-blood inflammatory indices and clinical factors to develop a novel prognostic score in mRCC patients receiving at least second-line nivolumab. The complete blood count before the first cycle of therapy was assessed by calculating neutrophil-to-lymphocyte ratio (NLR), derived NLR (dNLR), lymphocyte-to-monocyte ratio (LMR), platelet-to-lymphocyte ratio (PLR), systemic inflammation index (SII), and systemic inflammation response index (SIRI). Clinical factors included pre-treatment International Metastatic RCC Database Consortium (IMDC) score, line of therapy, and metastatic sites.
Results: From October 2015 to November 2019, 571 mRCC patients received nivolumab as second- and further-line treatment in 69% and 31% of cases. In univariable and multivariable analyses all inflammatory indices, IMDC score, and bone metastases significantly correlated with overall survival (OS). The multivariable model with NLR, IMDC score, and bone metastases had the highest c-index (0.697) and was chosen for the developing of the score (Schneeweiss scoring system). After internal validation (bootstrap re-sampling), the final index (Meet-URO score) composed by NLR, IMDC score, and bone metastases had a c-index of 0.691. It identified five categories with distinctive OSs: group 1 (median OS - mOS = not reached), group 2 (mOS = 43.9 months), group 3 (mOS = 22.4 months), group 4 (mOS = 10.3 months), and group 5 (mOS = 3.2 months). Moreover, the Meet-URO score allowed for a fine risk-stratification across all three IMDC groups.
Conclusion: The Meet-URO score allowed for the accurate stratification of pretreated mRCC patients receiving nivolumab and is easily applicable for clinical practice at no additional cost. Future steps include its external validation, the assessment of its predictivity, and its application to first-line combinations.