Predicting common solid renal tumors using machine learning models of classification of radiologist-assessed magnetic resonance characteristics

• Published in: Abdominal imaging 2020-09, Vol.45 (9), p.2797-2809
• Main Authors: Lopes Vendrami, Camila, McCarthy, Robert J., Villavicencio, Carolina Parada, Miller, Frank H.
• Format: Article
• Language: English
• Subjects: Accuracy; Bladder; Clear cell-type renal cell carcinoma; Datasets; Gastroenterology; Hepatology; Homogeneity; Image classification; Imaging; Kidney cancer; Kidneys; Learning algorithms; Lesions; Machine learning; Magnetic resonance imaging; Medicine; Medicine & Public Health; Model accuracy; Papillary renal cell carcinoma; Performance evaluation; Radiology; Recursive methods; Retroperitoneum; Tumors; Ureters
• ISSN: 2366-004X; 2366-0058
• DOI: 10.1007/s00261-020-02637-w

Purpose Solid renal masses (SRM) are difficult to differentiate based on standard MR features. The purpose of this study was to assess MR imaging features of SRM to evaluate performance of ensemble methods of classifying SRM subtypes. Materials and methods MR images of SRM ( n  = 330) were retrospectively evaluated for standard and multiparametric (mp) features. Models of MR features for predicting malignant and benign lesions as well as subtyping SRM were developed using a training dataset and performance was evaluated in a test data-set using recursive partitioning (RP), gradient booting machine (GBM), and random forest (RF) methods. Results In the test dataset, GBM and RF models demonstrated an accuracy of 86% (95% CI 75% to 93%) for predicting benign versus malignant SRM compared to 83% (95% CI 71% to 91%) for the RP model. RF had the greatest accuracy in predicting SRM subtypes, 81.2% (95% CI 69.5% to 89.9%) compared with GBM 73.4% (95% CI 60.9% to 83.7%) or RP 70.3% (95% CI 57.6% to 81.1%). Marginal homogeneity was reduced by the RF model compared with the RP model ( P