Jethro C C Kwong1,2, Adree Khondker3, Christopher Tran3, Emily Evans3, Adrian I Cozma4, Ashkan Javidan3, Amna Ali5, Munir Jamal1, Thomas Short1, Frank Papanikolaou1, John R Srigley6, Benjamin Fine5,7,8, Andrew Feifer1,5. 1. Division of Urology, Department of Surgery, University of Toronto, Toronto, ON, Canada. 2. Temerty Centre for AI Research and Education in Medicine, University of Toronto, Toronto, ON, Canada. 3. Temerty Faculty of Medicine, University of Toronto, Toronto, ON, Canada. 4. Department of Radiation Oncology, University of Toronto, Toronto, ON, Canada. 5. Institute for Better Health, Trillium Health Partners, Mississauga, ON, Canada. 6. Department of Laboratory Medicine and Pathobiology, University of Toronto, Toronto, ON, Canada. 7. Operational Analytics Lab, Trillium Health Partners, Mississauga, ON, Canada. 8. Department of Medical Imaging, University of Toronto, Toronto, ON, Canada.
Abstract
INTRODUCTION: We aimed to develop an explainable machine learning (ML) model to predict side-specific extraprostatic extension (ssEPE) to identify patients who can safely undergo nerve-sparing radical prostatectomy using preoperative clinicopathological variables. METHODS: A retrospective sample of clinicopathological data from 900 prostatic lobes at our institution was used as the training cohort. Primary outcome was the presence of ssEPE. The baseline model for comparison had the highest performance out of current biopsy-derived predictive models for ssEPE. A separate logistic regression (LR) model was built using the same variables as the ML model. All models were externally validated using a testing cohort of 122 lobes from another institution. Models were assessed by area under receiver-operating-characteristic curve (AUROC), precision-recall curve (AUPRC), calibration, and decision curve analysis. Model predictions were explained using SHapley Additive exPlanations. This tool was deployed as a publicly available web application. RESULTS: Incidence of ssEPE in the training and testing cohorts were 30.7 and 41.8%, respectively. The ML model achieved AUROC 0.81 (LR 0.78, baseline 0.74) and AUPRC 0.69 (LR 0.64, baseline 0.59) on the training cohort. On the testing cohort, the ML model achieved AUROC 0.81 (LR 0.76, baseline 0.75) and AUPRC 0.78 (LR 0.75, baseline 0.70). The ML model was explainable, well-calibrated, and achieved the highest net benefit for clinically relevant cutoffs of 10-30%. CONCLUSIONS: We developed a user-friendly application that enables physicians without prior ML experience to assess ssEPE risk and understand factors driving these predictions to aid surgical planning and patient counselling (https://share.streamlit.io/jcckwong/ssepe/main/ssEPE_V2.py).
INTRODUCTION: We aimed to develop an explainable machine learning (ML) model to predict side-specific extraprostatic extension (ssEPE) to identify patients who can safely undergo nerve-sparing radical prostatectomy using preoperative clinicopathological variables. METHODS: A retrospective sample of clinicopathological data from 900 prostatic lobes at our institution was used as the training cohort. Primary outcome was the presence of ssEPE. The baseline model for comparison had the highest performance out of current biopsy-derived predictive models for ssEPE. A separate logistic regression (LR) model was built using the same variables as the ML model. All models were externally validated using a testing cohort of 122 lobes from another institution. Models were assessed by area under receiver-operating-characteristic curve (AUROC), precision-recall curve (AUPRC), calibration, and decision curve analysis. Model predictions were explained using SHapley Additive exPlanations. This tool was deployed as a publicly available web application. RESULTS: Incidence of ssEPE in the training and testing cohorts were 30.7 and 41.8%, respectively. The ML model achieved AUROC 0.81 (LR 0.78, baseline 0.74) and AUPRC 0.69 (LR 0.64, baseline 0.59) on the training cohort. On the testing cohort, the ML model achieved AUROC 0.81 (LR 0.76, baseline 0.75) and AUPRC 0.78 (LR 0.75, baseline 0.70). The ML model was explainable, well-calibrated, and achieved the highest net benefit for clinically relevant cutoffs of 10-30%. CONCLUSIONS: We developed a user-friendly application that enables physicians without prior ML experience to assess ssEPE risk and understand factors driving these predictions to aid surgical planning and patient counselling (https://share.streamlit.io/jcckwong/ssepe/main/ssEPE_V2.py).
Authors: John R Srigley; Peter A Humphrey; Mahul B Amin; Sam S Chang; Lars Egevad; Jonathan I Epstein; David J Grignon; James M McKiernan; Rodolfo Montironi; Andrew A Renshaw; Victor E Reuter; Thomas M Wheeler Journal: Arch Pathol Lab Med Date: 2009-10 Impact factor: 5.534
Authors: Alberto Martini; Akriti Gupta; Sara C Lewis; Shivaram Cumarasamy; Kenneth G Haines; Alberto Briganti; Francesco Montorsi; Ashutosh K Tewari Journal: BJU Int Date: 2018-05-14 Impact factor: 5.588
Authors: Timo F W Soeterik; Harm H E van Melick; Lea M Dijksman; Heidi Küsters-Vandevelde; Saskia Stomps; Ivo G Schoots; Douwe H Biesma; J A Witjes; Jean-Paul A van Basten Journal: Eur Urol Oncol Date: 2020-09-22
Authors: Maarten de Rooij; Esther H J Hamoen; J Alfred Witjes; Jelle O Barentsz; Maroeska M Rovers Journal: Eur Urol Date: 2015-07-26 Impact factor: 20.096
Authors: Laura N Nguyen; Linden Head; Kelsey Witiuk; Nahid Punjani; Ranjeeta Mallick; Sonya Cnossen; Dean A Fergusson; Ilias Cagiannos; Luke T Lavallée; Christopher Morash; Rodney H Breau Journal: J Urol Date: 2017-03-09 Impact factor: 7.450
Authors: Timo F W Soeterik; Harm H E van Melick; Lea M Dijksman; Heidi V N Küsters-Vandevelde; Douwe H Biesma; J A Witjes; Jean-Paul A van Basten Journal: Urol Oncol Date: 2020-02-19 Impact factor: 3.498