Development of a prognostic risk model for clear cell renal cell carcinoma by systematic evaluation of DNA...

Although several prognostic models have been developed for clear cell renal cell carcinoma (ccRCC), these are still suboptimal and there is a need to identify additional prognostic biomarkers. Previously, we developed a prognostic model containing five DNA methylation markers (NEFH, NEURL, GATA5, GREM1, and LAD1) and clinicopathological characteristics based on the TNM 3rd edition and Fuhrman grading system. Here, we evaluated the effect of the recent ISUP/ 2022 WHO revisions on our previous prognostic model by incorporating the new ISUP/WHO standards, TNM 8th edition, and several novel prognostic factors (necrosis, lymphovascular invasion, sarcomatoid and rhabdoid features). Data from 308 ccRCC cases from the Netherlands Cohort Study were included for this study. Clinicopathological factors, novel prognostic factors, and the five methylation markers were analyzed for their individual and combined prognostic value using Kaplan-Meier analyses and Cox proportional hazard models. To compare models, the Akaike information criterion (AIC) and c-statistic were used. All evaluated factors were statistically significantly associated with cause-specific survival. The clinical model using the ISUP and TNM 8th edition performed similarly when compared to the Fuhrman/TNM 3rd edition model (AIC 592, c-statistic 0.63 and AIC 595, c-statistic 0.62, respectively). After addition of the five DNA methylation markers to the ISUP/TNM 8th model, this model was slightly improved (AIC 584, c-statistic 0.70). The addition of necrosis and lymphovascular invasion (LVI) did not further improve these results (AIC 586, c-statistic 0.71 and AIC 588, c-statistic 0.71, respectively). Despite the individual prognostic significance of necrosis, LVI, the presence of sarcomatoid and/or rhabdoid differentiation, ISUP, and TNM 8th edition, these factors did not influence the performance of our prognostic model. The model including the five DNA methylation markers, age at diagnosis, sex, TNM stage (8th edition), ISUP grading, and tumor size was the best performing model, thereby highlighting the potential importance of molecular markers.