1. Machine learning to predict early recurrence after oesophageal cancer surgery
- Author
-
S A Rahman, R C Walker, M A Lloyd, B L Grace, G I van Boxel, B F Kingma, J P Ruurda, R van Hillegersberg, S Harris, S Parsons, S Mercer, E A Griffiths, J R O'Neill, R Turkington, R C Fitzgerald, T J Underwood, Ayesha Noorani, Rachael Fels Elliott, Paul A W Edwards, Nicola Grehan, Barbara Nutzinger, Jason Crawte, Hamza Chettouh, Gianmarco Contino, Xiaodun Li, Eleanor Gregson, Sebastian Zeki, Rachel de la Rue, Shalini Malhotra, Simon Tavaré, Andy G Lynch, Mike L Smith, Jim Davies, Charles Crichton, Nick Carroll, Peter Safranek, Andrew Hindmarsh, Vijayendran Sujendran, Stephen J Hayes, Yeng Ang, Shaun R Preston, Sarah Oakes, Izhar Bagwan, Vicki Save, Richard J E Skipworth, Ted R Hupp, J Robert O'Neill, Olga Tucker, Andrew Beggs, Philippe Taniere, Sonia Puig, Timothy J Underwood, Fergus Noble, James P Byrne, Jamie J Kelly, Jack Owsley, Hugh Barr, Neil Shepherd, Oliver Old, Jesper Lagergren, James Gossage, Andrew Davies Fuju Chang, Janine Zylstra, Vicky Goh, Francesca D Ciccarelli, Grant Sanders, Richard Berrisford, Catherine Harden, David Bunting, Mike Lewis, Ed Cheong, Bhaskar Kumar, Simon L Parsons, Irshad Soomro, Philip Kaye, John Saunders, Laurence Lovat, Rehan Haidry, Victor Eneh, Laszlo Igali, Michael Scott, Shamila Sothi, Sari Suortamo, Suzy Lishman, George B Hanna, Christopher J Peters, and Anna Grabowska
- Subjects
Male ,Esophageal Neoplasms ,Lymphovascular invasion ,medicine.medical_treatment ,030230 surgery ,computer.software_genre ,Machine Learning ,0302 clinical medicine ,Upper GI ,Neoadjuvant therapy ,Aged, 80 and over ,Incidence (epidemiology) ,General Medicine ,Middle Aged ,Esophageal cancer ,Regression ,Random forest ,Oncology ,Esophagectomy ,Area Under Curve ,030220 oncology & carcinogenesis ,Cohort ,Adenocarcinoma ,Female ,Original Article ,Elastic net regularization ,Adult ,medicine.medical_specialty ,Early Recurrence ,Machine learning ,Risk Assessment ,03 medical and health sciences ,SDG 3 - Good Health and Well-being ,Clinical Decision Rules ,medicine ,Humans ,Aged ,Ensemble forecasting ,Receiver operating characteristic ,business.industry ,General surgery ,Original Articles ,medicine.disease ,ROC Curve ,Surgery ,Artificial intelligence ,Neoplasm Recurrence, Local ,business ,computer ,Cancer surgery - Abstract
Background Early cancer recurrence after oesophagectomy is a common problem, with an incidence of 20–30 per cent despite the widespread use of neoadjuvant treatment. Quantification of this risk is difficult and existing models perform poorly. This study aimed to develop a predictive model for early recurrence after surgery for oesophageal adenocarcinoma using a large multinational cohort and machine learning approaches. Methods Consecutive patients who underwent oesophagectomy for adenocarcinoma and had neoadjuvant treatment in one Dutch and six UK oesophagogastric units were analysed. Using clinical characteristics and postoperative histopathology, models were generated using elastic net regression (ELR) and the machine learning methods random forest (RF) and extreme gradient boosting (XGB). Finally, a combined (ensemble) model of these was generated. The relative importance of factors to outcome was calculated as a percentage contribution to the model. Results A total of 812 patients were included. The recurrence rate at less than 1 year was 29·1 per cent. All of the models demonstrated good discrimination. Internally validated areas under the receiver operating characteristic (ROC) curve (AUCs) were similar, with the ensemble model performing best (AUC 0·791 for ELR, 0·801 for RF, 0·804 for XGB, 0·805 for ensemble). Performance was similar when internal–external validation was used (validation across sites, AUC 0·804 for ensemble). In the final model, the most important variables were number of positive lymph nodes (25·7 per cent) and lymphovascular invasion (16·9 per cent). Conclusion The model derived using machine learning approaches and an international data set provided excellent performance in quantifying the risk of early recurrence after surgery, and will be useful in prognostication for clinicians and patients., Early recurrence after surgery for adenocarcinoma of the oesophagus is common. A risk prediction model was derived using modern machine learning methods that accurately predicts risk of early recurrence using postoperative pathology. Machine learning may help
- Published
- 2019
- Full Text
- View/download PDF