Intratumoral and Peritumoral Fat CT‑Based Radiomics for Predicting Recurrence Risk in Non-Muscle-Invasive Bladder Cancer: A Two-Center Study.
Authors
Affiliations (5)
Affiliations (5)
- Department of Nuclear Medicine, Shengjing Hospital of China Medical University, Shenyang, China.
- Department of Radiology, People's Hospital of China Medical University, Shenyang, China.
- Department of Radiology, The People's Hospital of Liaoning Province, Shenyang, China.
- Department of Urology, Shengjing Hospital of China Medical University, Shenyang, China.
- Department of Nuclear Medicine, Shengjing Hospital of China Medical University, Shenyang, China. [email protected].
Abstract
Non-muscle-invasive bladder cancer (NMIBC) has a high risk of recurrence, and multiple surgeries increase the disease burden on patients. Using computed tomography (CT)-based machine learning, this study established a pre-treatment recurrence prediction model incorporating tumor and peritumoral fat characteristics. This approach may guide early clinical interventions. In this retrospective study, 208 NMIBC patients who underwent enhanced CT before transurethral resection of bladder tumor (TURBT) with intravesical chemotherapy were collected from two hospitals. The radiomics features were extracted from the intratumoral region and peritumoral fat region (5 mm), followed by least absolute shrinkage and selection operator (LASSO) selection. Three radiomics models were developed: intratumoral, peritumoral-fat, and combined intratumoral-peritumoral model. Kaplan-Meier analysis assessed the association between radiomics features and recurrence-free survival. Cox analyses identified clinical risk factors integrated with radiomics score into a clinical-radiomics nomogram. Time-dependent ROC assessed the nomogram's predictive performance, and decision curve analysis evaluated its clinical utility. The combined model based on logistic regression demonstrated superior discrimination, with area under the curve (AUC) values of 0.88 in the test set and 0.82 in the external validation set. The clinical-radiomics nomogram exhibited optimal performance in predicting early recurrence for NMIBC, with time-AUC values of 0.86 and 0.84 in the test and external validation sets, respectively. The nomogram showed better calibration and reclassification than the clinical model (net reclassification improvement, 0.736; p < 0.05). The integrated radiomics-clinical model enhances the predictive capability compared with individual models, demonstrating significant value in predicting early recurrence of NMIBC. This approach offers a novel predictive strategy for assessing NMIBC recurrence risk.