A three-gene expression score is equivalent to a 25-gene panel as an independent prognostic factor in clear cell renal cell carcinoma

Abstract Purpose. Whether a cell-death-derived signature is prognostic in clear cell renal cell carcinoma (ccRCC) and survives parsimony, transportability and size-matched benchmarks is untested. Mitoxyperilysis is a lytic cell-death modality driven by innate immune stimulation plus metabolic disruption. Methods. From GSE235046 we assembled a 101-gene mouse panel, 87 evaluable in TCGA-KIRC (n = 508; 168 deaths); "MRG" denotes that selection history, not effectors. A 25-gene ridge index and a three-gene LASSO-Cox score were benchmarked against size-matched random panels, adjusted for clinical variables, frozen in two cohorts, and compared at 36 months by reclassification and decision curves. Results. Forty-five of 87 genes tracked survival at P < 0.05, 32 after stage and grade adjustment. The 25-gene index reached 0.711 with no significant coefficient, its 1-SE compression 0.705 against 0.706. The score was independently prognostic (HR 1.609 per SD; P = 1.02 × 10⁻⁹), recovered 26.4% of the room above its null, and held in E-MTAB-1980 but inverted in GSE29609. At 36 months it raised the C-index + 0.026 and the IDI + 0.041 beyond a size-matched null, while the NRI (+ 0.34) did not and the net-benefit gain included zero. High-risk tumours showed SETD2/BAP1 enrichment. Conclusions. Equivalence, not superiority: 22 further genes bought no measurable discrimination from 25 terms at 168 events. It marks a proliferation-adjacent, SETD2/BAP1-mutant state, not the mitoxyperilysis pathway; its clinical gain is modest and transportability rests on one positive and one negative cohort.