Carolyn D Seib1, James P Roose2, Alan E Hubbard2, Insoo Suh3. 1. Stanford-Surgery Policy Improvement Research and Education Center (S-SPIRE), Department of Surgery, Stanford University School of Medicine, Stanford, CA, United States; Division of General Surgery, Palo Alto Veterans Affairs Health Care System, United States. Electronic address: cseib@stanford.edu. 2. University of California, Berkeley, Division of Biostatistics, Berkeley, United States. 3. University of California, San Francisco, Section of Endocrine Surgery, San Francisco, United States.
Abstract
BACKGROUND: Accurate prediction of thyroidectomy complications is necessary to inform treatment decisions. Ensemble machine learning provides one approach to improve prediction. METHODS: We applied the Super Learner (SL) algorithm to the 2016-2018 thyroidectomy-specific NSQIP database to predict complications following thyroidectomy. Cross-validation was used to assess model discrimination and precision. RESULTS: For the 17,987 patients undergoing thyroidectomy, rates of recurrent laryngeal nerve injury, post-operative hypocalcemia prior to discharge or within 30 days, and neck hematoma were 6.1%, 6.4%, 9.0%, and 1.8%, respectively. SL improved prediction of thyroidectomy-specific outcomes when compared with benchmark logistic regression approaches. For postoperative hypocalcemia prior to discharge, SL improved the cross-validated AUROC to 0.72 (95%CI 0.70-0.74) compared to 0.70 (95%CI 0.68-0.72; p < 0.001) when using a manually curated logistic regression algorithm. CONCLUSION: Ensemble machine learning modestly improves prediction for thyroidectomy-specific outcomes. SL holds promise to provide more accurate patient-level risk prediction to inform treatment decisions. Published by Elsevier Inc.
BACKGROUND: Accurate prediction of thyroidectomy complications is necessary to inform treatment decisions. Ensemble machine learning provides one approach to improve prediction. METHODS: We applied the Super Learner (SL) algorithm to the 2016-2018 thyroidectomy-specific NSQIP database to predict complications following thyroidectomy. Cross-validation was used to assess model discrimination and precision. RESULTS: For the 17,987 patients undergoing thyroidectomy, rates of recurrent laryngeal nerve injury, post-operative hypocalcemia prior to discharge or within 30 days, and neck hematoma were 6.1%, 6.4%, 9.0%, and 1.8%, respectively. SL improved prediction of thyroidectomy-specific outcomes when compared with benchmark logistic regression approaches. For postoperative hypocalcemia prior to discharge, SL improved the cross-validated AUROC to 0.72 (95%CI 0.70-0.74) compared to 0.70 (95%CI 0.68-0.72; p < 0.001) when using a manually curated logistic regression algorithm. CONCLUSION: Ensemble machine learning modestly improves prediction for thyroidectomy-specific outcomes. SL holds promise to provide more accurate patient-level risk prediction to inform treatment decisions. Published by Elsevier Inc.
Authors: S F Khuri; J Daley; W Henderson; K Hur; J Demakis; J B Aust; V Chong; P J Fabri; J O Gibbs; F Grover; K Hammermeister; G Irvin; G McDonald; E Passaro; L Phillips; F Scamman; J Spencer; J F Stremple Journal: Ann Surg Date: 1998-10 Impact factor: 12.969
Authors: James C Iannuzzi; Fergal J Fleming; Kristin N Kelly; Daniel T Ruan; John R Monson; Jacob Moalem Journal: Surgery Date: 2014-11-11 Impact factor: 3.982
Authors: Karl Y Bilimoria; Yaoming Liu; Jennifer L Paruch; Lynn Zhou; Thomas E Kmiecik; Clifford Y Ko; Mark E Cohen Journal: J Am Coll Surg Date: 2013-09-18 Impact factor: 6.113
Authors: Jennifer N Cooper; Lai Wei; Soledad A Fernandez; Peter C Minneci; Katherine J Deans Journal: Comput Biol Med Date: 2014-12-08 Impact factor: 4.589
Authors: Christa R Abraham; Ashar Ata; Carrie B Carsello; Tiffany L Chan; Steven C Stain; Todd D Beyer Journal: J Am Coll Surg Date: 2014-03-02 Impact factor: 6.113
Authors: Joeky T Senders; Patrick C Staples; Aditya V Karhade; Mark M Zaki; William B Gormley; Marike L D Broekman; Timothy R Smith; Omar Arnaout Journal: World Neurosurg Date: 2017-10-03 Impact factor: 2.104