Radiomics and Image-based Artificial Intelligence for Predicting Recurrence and Survival After Surgery in Localized Renal Cell Carcinoma: An APPRAISE-AI Systematic Review and Meta-analysis Author links open overlay panel

Comprehensive Summary

This systematic review and meta-analysis evaluated the use of radiomics and image-based artificial intelligence (AI) models to predict recurrence and survival after surgery in patients with localized renal cell carcinoma. Analyzing 30 studies involving over 17,000 patients, the authors found that these models, mainly based on preoperative CT scans, showed high predictive accuracy, with a pooled 5-year recurrence-free survival AUC of 0.87 and similarly strong performance in externally validated cohorts. Texture and shape radiomic features, especially those derived from gray-level co-occurrence matrices, were the most consistently predictive. Models that combined radiomic features with clinical data performed better than radiomics alone.

Outcomes and Implications

These findings suggest that radiomics and AI models could help doctors identify high-risk RCC patients before surgery, guiding follow-up intensity and adjuvant therapy decisions. By improving preoperative risk stratification, these tools may optimize patient outcomes and support personalized treatment planning.