Ji Whae Choi1,2, Rong Hu3,4,5, Yijun Zhao6, Subhanik Purkayastha7,8, Jing Wu8,6, Aidan J McGirr9, S William Stavropoulos10, Alvin C Silva9, Michael C Soulen10, Matthew B Palmer11, Paul J L Zhang11, Chengzhang Zhu5,12, Sun Ho Ahn7,8, Harrison X Bai7,8. 1. Warren Alpert Medical School, Brown University, Providence, RI, 02903, USA. ji_whae_choi@brown.edu. 2. Department of Diagnostic Imaging, Rhode Island Hospital, Providence, RI, 02903, USA. ji_whae_choi@brown.edu. 3. School of Computer Science and Engineering, Central South University, Changsha, 410083, China. 4. Hunan Province Engineering Technology Research Center of Computer Vision and Intelligent Medical Treatment, Changsha, 410083, China. 5. Joint Laboratory of Mobile Health, Ministry of Education and China Mobile, Hunan, 410083, China. 6. Department of Radiology, Second Xiangya Hospital of Central South University, Changsha, Hunan, 410011, China. 7. Warren Alpert Medical School, Brown University, Providence, RI, 02903, USA. 8. Department of Diagnostic Imaging, Rhode Island Hospital, Providence, RI, 02903, USA. 9. Department of Radiology, Mayo Clinic Hospital, Scottsdale, AZ, 85054, USA. 10. Department of Radiology, Hospital of the University of Pennsylvania, Philadelphia, 19104, PA, USA. 11. Department of Pathology and Laboratory Medicine, Hospital of the University of Pennsylvania, Philadelphia, 19104, PA, USA. 12. College of Literature and Journalism, Central South University, Changsha, 410083, China.
Abstract
PURPOSE: Clear cell renal cell carcinoma (ccRCC) is the most common subtype of renal cell carcinoma. Currently, there is a lack of noninvasive methods to stratify ccRCC prognosis prior to any invasive therapies. The purpose of this study was to preoperatively predict the tumor stage, size, grade, and necrosis (SSIGN) score of ccRCC using MRI-based radiomics. METHODS: A multicenter cohort of 364 histopathologically confirmed ccRCC patients (272 low [< 4] and 92 high [≥ 4] SSIGN score) with preoperative T2-weighted and T1-contrast-enhanced MRI were retrospectively identified and divided into training (254 patients) and testing sets (110 patients). The performance of a manually optimized radiomics model was assessed by measuring accuracy, sensitivity, specificity, area under receiver operating characteristic curve (AUROC), and area under precision-recall curve (AUPRC) on an independent test set, which was not included in model training. Lastly, its performance was compared to that of a machine learning pipeline, Tree-Based Pipeline Optimization Tool (TPOT). RESULTS: The manually optimized radiomics model using Random Forest classification and Analysis of Variance feature selection methods achieved an AUROC of 0.89, AUPRC of 0.81, accuracy of 0.89 (95% CI 0.816-0.937), specificity of 0.95 (95% CI 0.875-0.984), and sensitivity of 0.72 (95% CI 0.537-0.852) on the test set. The TPOT using Extra Trees Classifier achieved an AUROC of 0.94, AUPRC of 0.83, accuracy of 0.89 (95% CI 0.816-0.937), specificity of 0.95 (95% CI 0.875-0.984), and sensitivity of 0.72 (95% CI 0.537-0.852) on the test set. CONCLUSION: Preoperative MR radiomics can accurately predict SSIGN score of ccRCC, suggesting its promise as a prognostic tool that can be used in conjunction with diagnostic markers.
PURPOSE:Clear cell renal cell carcinoma (ccRCC) is the most common subtype of renal cell carcinoma. Currently, there is a lack of noninvasive methods to stratify ccRCC prognosis prior to any invasive therapies. The purpose of this study was to preoperatively predict the tumor stage, size, grade, and necrosis (SSIGN) score of ccRCC using MRI-based radiomics. METHODS: A multicenter cohort of 364 histopathologically confirmed ccRCC patients (272 low [< 4] and 92 high [≥ 4] SSIGN score) with preoperative T2-weighted and T1-contrast-enhanced MRI were retrospectively identified and divided into training (254 patients) and testing sets (110 patients). The performance of a manually optimized radiomics model was assessed by measuring accuracy, sensitivity, specificity, area under receiver operating characteristic curve (AUROC), and area under precision-recall curve (AUPRC) on an independent test set, which was not included in model training. Lastly, its performance was compared to that of a machine learning pipeline, Tree-Based Pipeline Optimization Tool (TPOT). RESULTS: The manually optimized radiomics model using Random Forest classification and Analysis of Variance feature selection methods achieved an AUROC of 0.89, AUPRC of 0.81, accuracy of 0.89 (95% CI 0.816-0.937), specificity of 0.95 (95% CI 0.875-0.984), and sensitivity of 0.72 (95% CI 0.537-0.852) on the test set. The TPOT using Extra Trees Classifier achieved an AUROC of 0.94, AUPRC of 0.83, accuracy of 0.89 (95% CI 0.816-0.937), specificity of 0.95 (95% CI 0.875-0.984), and sensitivity of 0.72 (95% CI 0.537-0.852) on the test set. CONCLUSION: Preoperative MR radiomics can accurately predict SSIGN score of ccRCC, suggesting its promise as a prognostic tool that can be used in conjunction with diagnostic markers.
Entities:
Keywords:
Imaging analysis; Medical imaging; Neoplasm progression; Renal cancer
Authors: Ianto Lin Xi; Yijun Zhao; Robin Wang; Marcello Chang; Subhanik Purkayastha; Ken Chang; Raymond Y Huang; Alvin C Silva; Martin Vallières; Peiman Habibollahi; Yong Fan; Beiji Zou; Terence P Gade; Paul J Zhang; Michael C Soulen; Zishu Zhang; Harrison X Bai; S William Stavropoulos Journal: Clin Cancer Res Date: 2020-01-14 Impact factor: 12.531