p53-expression in patients with renal cell carcinoma correlates with a higher probability of disease progression and increased cancer-specific mortality after surgery but does not enhance the predictive accuracy of robust outcome models.

OBJECTIVE: Due to lacking external validation, molecular biomarkers are currently not applied for risk-stratification of patients with localized renal cell carcinoma. The objective of this study was to externally validate a molecular multi-marker panel included in a previously proposed prognostic nomogram for the prediction of postoperative disease-free survival. METHODS AND MATERIALS: Besides pathologic tumor stage (pT) and ECOG-Performance Status, the nomogram includes 5 molecular markers (Ki-67, p53, VEGFR-1 endothelial or epithelial, and VEGF-D epithelial). The validation cohort comprised 343 renal cell carcinoma patients treated by radical nephrectomy or nephron-sparing surgery from 1999 to 2004 at a single academic center (median follow-up: 100 months). By multivariable Cox proportional-hazards regression models, the impact of clinical and molecular markers included in the nomogram on disease progression (DP) and cancer-specific mortality (CSM) was assessed; in addition, it was evaluated to what extent molecular markers added to the models' predictive accuracy (PA). RESULTS: Of all parameters included in the nomogram, ECOG-PS and pT-stage only revealed a significant impact on both endpoints. p53 (per 10% measures) showed a significant impact on DP (HR = 1.31; P = 0.008), albeit not on CSM, while all other molecular markers did not impact study endpoints. Using Martingale residuals, a cut-off value for p53-expression<20% (negative) vs. ≥20% (positive) yielded the highest impact on DP and CSM. In outcome-models including further well-established histo-pathological factors, p53-expression dichotomized at 20% independently impacted DP (HR = 4.13; P = 0.004) and CSM (HR = 3.74; P = 0.033), while no significant PA gain was achieved. CONCLUSIONS: p53 showed a statistically significant impact on DP, albeit not on CSM, when applying the 10% expression cut-off as used in the original nomogram, while the prognostic value of all other examined markers included in the nomogram could not be confirmed. When an alternative cut-off of 20% was applied in multivariable models, p53 independently impacted DP and CSM, while the PA was not significantly enhanced. Hence, the clinical significance of p53 is still to be determined. Based on the results of this study it is not recommendable to use p53-expression and the Klatte nomogram in routine clinical decision-making.