Pretreatment 18F-FDG PET/CT Radiomics for Predicting Response to First-Line Targeted Therapy Plus Immunotherapy in Metastatic Renal Cell Carcinoma

Overview

This single-center retrospective observational study evaluates whether quantitative imaging features from pretreatment 18F-fluorodeoxyglucose positron emission tomography/computed tomography (18F-FDG PET/CT) can help predict response to first-line tyrosine kinase inhibitor (TKI) plus immune checkpoint inhibitor (ICI) therapy in patients with metastatic renal cell carcinoma. The study reviews existing medical records and imaging data from 228 patients treated at the First Affiliated Hospital of Fujian Medical University between January 2019 and December 2025. Treatment was selected as part of routine clinical care and was not assigned by the study. Imaging features are extracted separately from the primary kidney tumor and metastatic lesions and are then combined to develop a patient-level multi-lesion radiomics model. Treatment response is assessed according to Response Evaluation Criteria in Solid Tumors version 1.1 approximately 12 weeks after treatment initiation. Complete response or partial response is classified as response, whereas stable disease or progressive disease is classified as non-response. The study compares imaging-based, clinical, and combined prediction models and also evaluates their association with progression-free survival.

Patients with metastatic renal cell carcinoma may have different responses to first-line treatment combining a tyrosine kinase inhibitor with an immune checkpoint inhibitor. Differences may also occur among the primary renal tumor and metastatic lesions within the same patient. Pretreatment 18F-FDG PET/CT provides whole-body information about tumor glucose metabolism, while radiomics can extract quantitative imaging features that may reflect tumor heterogeneity. This is a single-center retrospective observational cohort study using existing clinical, pathological, treatment, follow-up, and pretreatment 18F-FDG PET/CT data. Eligible patients had pathologically confirmed metastatic renal cell carcinoma, underwent 18F-FDG PET/CT before starting first-line TKI plus ICI therapy, had at least one measurable lesion according to RECIST version 1.1, and had adequate treatment-response and follow-up information. The study did not assign treatment or alter routine clinical care. Primary renal tumors and eligible metastatic lesions are segmented separately on pretreatment PET/CT images. Conventional metabolic parameters and radiomic features are extracted from each lesion. For patients with multiple metastatic lesions, lesion-level features are aggregated to generate patient-level variables. Primary-tumor, metastatic-lesion, and combined multi-lesion radiomics models are developed. Clinical and metabolic variables are also evaluated, and a combined prediction model is constructed. Patients are divided into training and internal validation cohorts for model development and evaluation. The primary outcome is treatment response assessed approximately 12 weeks after treatment initiation according to RECIST version 1.1. Complete response and partial response are classified as responder status, while stable disease and progressive disease are classified as non-responder status. Progression-free survival is also evaluated to explore whether the prediction model is associated with longer-term treatment benefit.