Abstract Introduction and Objective: The prognostic significance of bone metastasis (BM) distribution in renal cell carcinoma (RCC) is unclear, as commonly used tools such as the Memorial Sloan Kettering Cancer Center (MSKCC) /Motzer score don’t account for the spatial bone lesion patterns. This study aimed to quantify the association between bone metastasis distribution and prognosis in RCC-BM patients and to build a predictive model based on the random survival forest (RSF) algorithm. Materials and Methods: At first BM diagnosis, 122 patients were stratified by MSKCC/Motzer risk score and classified into locoregional (thoracic/lumbar vertebrae limited, 21. 3%), stochastic (random bone lesions without visceral involvement, 56. 6%) and extensive (bone plus visceral metastases, 22. 1%) groups based on bone lesion distribution. Spinal, pelvic, upper extremity and lower extremity involvement occurred in 39. 3%, 35. 2%, 18. 0% and 16. 4% of cases. Univariate analysis, logistic regression, and Kaplan-Meier survival analyses assessed associations between bone lesion distribution and clinical risk. A RSF model was developed for survival prediction. The performance was compared to Cox regression using time-dependent area under the curve (AUC) analyses with repeated random train-validation splits. Results: Locoregional spread, spinal involvement (odds ratio/OR 3. 30, 95% CI 1. 20–9. 09, p0. 05) and advanced age (OR 1. 04, 95% CI 1. 00–1. 08, p0. 05) predicted higher MSKCC/Motzer risk stratification. Pelvic metastasis yielded shorter median overall survival (32 vs 49 months; p0. 05). The RSF model was trained (70%) and validated (30%) with spatial lesion involvement (pelvic, spinal and upper extremity involvement), MSKCC/Motzer score and age. In single split validation, 1- and 3-year AUCs were 0. 90 and 0. 87. Consistent performance was observed across 100 repeated splits, with median AUCs of 0. 89, 0. 86 and 0. 89 for 1-, 3- and 5-year survival. A cutoff of 15. 03 distinguished high- and low-risk groups (p0. 05). RSF outperformed over Cox regression (median AUC 0. 89 vs 0. 59). Conclusion: This study identifies the distribution patterns as prognosis indicators. Locoregional and spinal metastasis predict higher risks in MSKCC/Motzer system. Incorporating these factors into RSF model with MSKCC/Motzer score improves risk stratification in RCC-BM patients and supports more precise care. Citation Format: Zixiong Huang, Xiaopeng Zhang, Shijun Liu, Srinivas Nallandhighal, Tao Xu, Simpa Salami. Lesion distribution and prognosis of renal cell carcinoma bone metastasis: A novel evaluation model based on random survival forests abstract. In: Proceedings of the AACR Special Conference in Cancer Research: Innovations in Kidney Cancer Research: From Molecular Insights to Therapeutic Breakthroughs; 2026 Mar 13-16; Philadelphia, PA. Philadelphia (PA): AACR; Cancer Res 2026;86 (5Suppl₂): Abstract nr B003.
Huang et al. (2026) studied this question.