A Web-Based Prognostic Model for Pediatric Genitourinary Rhabdomyosarcoma: Analysis of Population-Based Cohort With External Validation

BACKGROUND: We conduct an analysis of data from the Surveillance, Epidemiology, and End Results (SEER) database, intending to identify prognostic factors of pediatric genitourinary rhabdomyosarcoma (PGU–RMS). Prognostic nomogram and web-based calculator were developed for potential clinical use. METHODS: Data of PGU–RMS patients were extracted from the SEER database as training and internal validation cohort, patients diagnosed as PGU–RMS from 2001 to 2015 in Beijing Children's Hospital were collected as an external validation cohort. We used log-rank tests to seek risk factors on the overall survival (OS) in the overall SEER cohort, tumor site subgroups, radiation subgroups, and metastasis subgroups. The univariable and multivariate Cox regression analyses were applied to establish the prognosis model. RESULTS: A total of 372 PGU-RMS patients in SEER and 84 patients from our center were included. 1-, 3-, and 5-year OS of the overall SEER cohort were 95.8, 82.1, and 78.8%. Subgroup analysis indicated that tumors located in the prostate/bladder were associated with a worse prognosis than the paratesticular, female genital system, and other sites (P < 0.001). Tumors of the T1/T2 stage, without regional lymph node, involvement or metastasis, can benefit from radiotherapy (P < 0.05). For patients without metastasis, younger age, T1/T2 stage, and undergoing radiation were associated with better prognosis (P < 0.05). The prognosis nomogram was well-calibrated, the concordance index (C-index) for the OS prediction was 0.823, 0.803, and 0.768 in training, internal and external validation cohort, the area under the receiver operating characteristic curve for 3-, and 5-year OS were 0.84, 0.84 in the training cohort, 0.90, 0.84 in internal validation cohort and 0.75, 0.80 in the external validation cohort. Decision curve analysis showed good clinical utility. The predictive performance of the nomogram was higher than the Intergroup Rhabdomyosarcoma Study Group (IRSG) pretreatment stage system based on the comparison of overtime C-index, net reclassification index, and integrated discriminatory index (P < 0.001). CONCLUSION: A comprehensive analysis of OS for PGU–RMS patients was conducted based on population cohort. The established prognosis nomogram has been fully validated and evaluated, exhibits better performance than the IRSG pretreatment stage system. Furthermore, a web-based risk calculator was developed to optimize clinical decisions.