Therapies for metastatic renal cell carcinoma (mRCC) have evolved significantly, making treatment decisions more complex. We used machine learning (ML) to identify subgroups of patients who have a high probability of response to first line systemic treatment.
Patients from the International mRCC Database Consortium (IMDC) with mRCC and treatment response measured to first-line were identified, and a ML classification and regression tree analysis was conducted, in which we grew a complex tree up to a depth of 30 with a minimum node split size of 2 with no constraints on the cost-complexity parameter. The resulting tree was pruned according to the cost-complexity parameter that minimized the leave one out cross-validated error rate and had a minimum bucket size of 25 patients.
2,549 patients were included, 73.2% male, 13.5% non-clear cell histology, 70.3% nephrectomy. 19.4%, 54.2%, and 26.4% had favorable, intermediate and poor IMDC risk respectively. First line treatment regimens consisted of VEGF inhibitors (51.5%), IO-IO combinations (32.3%), and IO-TKI combinations (16.2%). The ORR was 36.0% overall, with 29.6% for VEGF inhibitors, 39.1% for IO-IO, and 50.2% for IO-TKI combinations. ML identified 5 hierarchal variables, therapy type, nephrectomy, lung metastasis, other sites of metastasis, and age, that divided patients into 7 different categories with different response probabilities. VEGF therapy showed the poorest response, with no additional variables able to predict response. The highest ORR was observed in patients treated with IO-TKI and nephrectomy (54.9%); and in those treated with IO-IO, nephrectomy, and only lung metastasis (59.8%). Factors associated with poorer ORR included non-clear cell histology, older age, bone and liver metastases, poor performance status, elevated neutrophils, and poor IMDC risk score.
This large-scale ML analysis identified five key clinical variables that predict treatment response in mRCC, with treatment type emerging as the primary determinant. These results suggest that treatment selection for mRCC could potentially be optimized by considering these hierarchical variables, though further validation is needed.
Frontiers in oncology. 2026 May 04*** epublish ***
Martin Zarba, Dylan O'Sullivan, David Maj, Winson Y Cheung, Lisa Ludwig, Connor Wells, Evan Ferrier, Razane El Hajj Chehade, Frede Donskov, Marc Eid, Sumanta Kumar Pal, Benoit Beuselinck, Rana R McKay, Lori Wood, Jae Lyun Lee, Cristina Suárez, Kosuke Takemura, Ignacio Duran, Toni K Choueiri, Daniel Yc Heng
Department of Medical Oncology, Arthur J.E. Child Comprehensive Cancer Centre, Calgary, AB, Canada., Cumming School of Medicine, Department of Oncology, University of Calgary, Calgary, AB, Canada., Department of Medical Oncology, IPSEN, Edmonton, AB, Canada., Dana-Farber Cancer Institute, Harvard Medical School, Boston, MA, United States., Department of Medical Oncology, University Hospital of Southern Denmark, Esbjerg, Denmark., Department of Medical Oncology, City of Hope Comprehensive Cancer Center, Duarte, Canada., Department of Medical Oncology, UZ Leuven, Leuven, Belgium., Department of Medical Oncology, UC San Diego School of Medicine, San Diego, CA, United States., Queen Elizabeth II Health Sciences Centre, Dalhousie University, Halifax, NS, Canada., Asan Medical Center, University of Ulsan College of Medicine, Seoul, Republic of Korea., Medical Oncology, Vall d'Hebron Institute of Oncology, Vall d'Hebron University Hospital, Barcelona, Spain., Department of Genitourinary Oncology, Cancer Institute Hospital, Japanese Foundation for Cancer Research, Tokyo, Japan., Hospital Universitario Marqués de Valdecilla, IDIVAL Santander, Santander, Spain.