A pathology-based model for postoperative recurrence or metastasis prediction and adjuvant immunotherapy implications of clear-cell renal cell carcinoma: A multi-center, retrospective study.

Abstract

e16539 Background: Approximately 30% of patients with clear cell renal cell carcinoma experience recurrence after surgery, making it crucial to improve the prognosis for these patients. Stratification of patients with non-metastatic clear cell renal cell carcinoma (ccRCC) based on the risk of postoperative recurrence helps guide adjuvant immunotherapy after surgery. Methods: We enrolled 2,154 patients from two centers. Only patients without postoperative adjuvant therapy were used to analyze risk factors and constructing models. A multivariable model was constructed to predict disease-free survival (DFS) and stratify patients. The log-rank test was used to examine the effects of immune checkpoint inhibitors (ICIs) and targeted therapy on DFS across different strata. Results: We identified seven independent risk factors including: sex, microvascular invasion (MVI), T stage, pathological grade, sarcomatoid differentiation, necrosis and capsular involvement. Using these seven characteristics, a prognostic model (MISNCST model) for non-metastatic clear cell renal cell carcinoma was constructed. Using the constructed model, we stratified the patients and validated the stratification efficacy with DFS, overall survival (OS), and cancer-specific survival (CSS) as endpoints (all p < 0.001). With our model, we stratified patients who either received or did not receive postoperative adjuvant therapy and found that ICIs significantly improved DFS in the high-risk group patients. Compared to the current medication criteria in clinical trials for ICIs, our model demonstrated superior overall performance. Conclusions: We identified 7 crucial prognostic features influencing the prognosis of non-metastatic ccRCC, and developed a prognostic model using these features. Based on our model, we stratified patients and discovered that high-risk patients could benefit from treatment with ICIs. Comparison of stratification approaches between the MISNCST model and other cohorts in terms of the five-year recurrence risk. Training Set (N=1605) Sensitivity Specificity Youden's Index MISNCST model 0.550 0.944 0.494 KEYNOTE-564 0.275 0.972 0.247 PROSPER RCC 0.438 0.932 0.370 CheckMate 914 0.388 0.951 0.339 IMmotion010 0.250 0.984 0.234