Clinical genitourinary cancer · 2026
Multidistance Peritumoral Radiomics With Intratumoral Radiomics Integration Nomogram for Differentiating Fat-Poor Angiomyolipoma From Clear Cell Renal Cell Carcinoma in Discordant and Indeterminate Cases.
СтатьяНазвание и аннотация: Перевод в очереди
Обработка статьи · этапы, ошибки и управление
Загрузка состояния…
Ошибки и ограничения
История · было → стало
Состояние обновляется каждые 5 секунд, пока блок открыт. Процент показывается только там, где известен объём работы. Пауза останавливает продолжение этапа; текущий запрос к источнику может завершиться.
Xiao Yue, Ye Xupeng, Mao Huchao, Zheng Runfei, Chen Shuyao, Du Xiaoying, Bao Rendang, Chen Yuhan, Yang Mei, Zhang Yan, Cao Guoquan
Аннотация
INTRODUCTION: Accurate preoperative differentiation between fat-poor angiomyolipoma (fp-AML) and clear cell renal cell carcinoma (ccRCC) is clinically critical but remains challenging due to overlapping imaging features. Existing radiomics studies primarily focus on intratumoral features with limited validation on diagnostically ambiguous cases.
PATIENTS: This retrospective study included 623 patients with pathologically confirmed renal tumors. The training cohort comprised 429 cases with concordant imaging-pathology diagnoses, while the independent test cohort was enriched with 194 clinically challenging cases characterized by indeterminate imaging reports or initial imaging-pathology discordance.
METHODS: Intratumoral and peritumoral regions of interest were manually segmented, with peritumoral boundaries systematically expanded by 1mm, 2mm, 3mm, 4mm, and 5mm. Sixteen radiomics models were constructed and compared using seven machine learning algorithms. The optimal radiomics model was integrated with clinical features to develop a nomogram, with performance evaluated by AUC, calibration curves, and decision curve analysis.
RESULTS: The combined intratumoral-2mm peritumoral model using XGBoost demonstrated superior discriminative performance, achieving an AUC of 0.779 in the independent test cohort. Integration with clinical features (age and gender) yielded a nomogram with significantly improved performance (AUC 0.829), outperforming both clinical-only (AUC 0.733) and radiomics-only models. Calibration and decision curve analyses confirmed satisfactory goodness-of-fit and clinical net benefit.
CONCLUSIONS: The nomogram integrating intratumoral and peritumoral radiomics with clinical data provides a robust noninvasive decision-support tool for improving preoperative differentiation between fp-AML and ccRCC in diagnostically challenging cases. This approach has the potential to guide personalized management and reduce unnecessary surgical interventions.
PATIENTS: This retrospective study included 623 patients with pathologically confirmed renal tumors. The training cohort comprised 429 cases with concordant imaging-pathology diagnoses, while the independent test cohort was enriched with 194 clinically challenging cases characterized by indeterminate imaging reports or initial imaging-pathology discordance.
METHODS: Intratumoral and peritumoral regions of interest were manually segmented, with peritumoral boundaries systematically expanded by 1mm, 2mm, 3mm, 4mm, and 5mm. Sixteen radiomics models were constructed and compared using seven machine learning algorithms. The optimal radiomics model was integrated with clinical features to develop a nomogram, with performance evaluated by AUC, calibration curves, and decision curve analysis.
RESULTS: The combined intratumoral-2mm peritumoral model using XGBoost demonstrated superior discriminative performance, achieving an AUC of 0.779 in the independent test cohort. Integration with clinical features (age and gender) yielded a nomogram with significantly improved performance (AUC 0.829), outperforming both clinical-only (AUC 0.733) and radiomics-only models. Calibration and decision curve analyses confirmed satisfactory goodness-of-fit and clinical net benefit.
CONCLUSIONS: The nomogram integrating intratumoral and peritumoral radiomics with clinical data provides a robust noninvasive decision-support tool for improving preoperative differentiation between fp-AML and ccRCC in diagnostically challenging cases. This approach has the potential to guide personalized management and reduce unnecessary surgical interventions.