2018 Leibovich prognostic model for renal cell carcinoma: Performance in a large population with special consideration of Black race.

Abstract

The 2018 Leibovich prognostic model for nonmetastatic renal cell carcinoma (RCC) combines clinical, surgical, and pathologic factors to predict progression-free survival (PFS) and cancer-specific survival (CSS) for patients with clear cell (ccRCC), papillary (pRCC), and chromophobe (chRCC) histology. Despite high accuracy, <1% of the original cohort was Black. Here, the authors examined this model in a large population with greater Black patient representation. By using a prospectively maintained RCC institutional database, patients were assigned Leibovich model risk scores. Survival outcomes included 5-year and 10-year PFS and CSS. Prognostic accuracy was determined using area under the curve (AUC) analysis and calibration plots. Black patient subanalyses were conducted. In total, 657 (29%) of 2295 patients analyzed identified as Black. Declines in PFS and CSS were observed as scores increased. Discrimination for ccRCC was strong for PFS (AUC: 5-year PFS, 0.81; 10-year PFS, 0.78) and for CSS (AUC: 5-year CSS, 0.82; 10-year CSS, 0.74). The pRCC AUC for PFS was 0.74 at 5 years and 0.71 at 10 years; and the AUC for CSS was 0.74 at 5 years and 0.70 at 10 years. In chRCC, better performance was observed for CSS (AUC at 5 years, 0.75) than for PFS (AUC: 0.66 at 5 years; 0.55 at 10 years). Black patient subanalysis revealed similar-to-improved performance for ccRCC at 5 years (AUC: PFS, 0.79; CSS, 0.87). For pRCC, performance was lower for PFS (AUC at 5 years, 0.63) and was similar for CSS (AUC at 5 years, 0.77). Sample size limited Black patient 10-year and chRCC analyses. The authors externally validated the 2018 Leibovich RCC prognostic model and found optimal performance for ccRCC, followed by pRCC, and then chRCC. Importantly, the results were consistent in this large representation of Black patients. In 2018, a model to predict survival in patients with renal cell carcinoma (kidney cancer) was introduced by Leibovich etal. This model has performed well; however, Black patients have been under-represented in examination of its performance. In this study, 657 Black patients (29%) were included, and the results were consistent. This work is important for making sure the model can be applied to all patient populations.