Literature DB >> 35508684

Differential Performance of Machine Learning Models in Prediction of Procedure-Specific Outcomes.

Kevin A Chen1, Matthew E Berginski2, Chirag S Desai1, Jose G Guillem1, Jonathan Stem1, Shawn M Gomez2,3, Muneera R Kapadia4.   

Abstract

BACKGROUND: Procedure-specific complications can have devastating consequences. Machine learning-based tools have the potential to outperform traditional statistical modeling in predicting their risk and guiding decision-making. We sought to develop and compare deep neural network (NN) models, a type of machine learning, to logistic regression (LR) for predicting anastomotic leak after colectomy, bile leak after hepatectomy, and pancreatic fistula after pancreaticoduodenectomy (PD).
METHODS: The colectomy, hepatectomy, and PD National Surgical Quality Improvement Program (NSQIP) databases were analyzed. Each dataset was split into training, validation, and testing sets in a 60/20/20 ratio, with fivefold cross-validation. Models were created using NN and LR for each outcome. Models were evaluated primarily with area under the receiver operating characteristic curve (AUROC).
RESULTS: A total of 197,488 patients were included for colectomy, 25,403 for hepatectomy, and 23,333 for PD. For anastomotic leak, AUROC for NN was 0.676 (95% 0.666-0.687), compared with 0.633 (95% CI 0.620-0.647) for LR. For bile leak, AUROC for NN was 0.750 (95% CI 0.739-0.761), compared with 0.722 (95% CI 0.698-0.746) for LR. For pancreatic fistula, AUROC for NN was 0.746 (95% CI 0.733-0.760), compared with 0.713 (95% CI 0.703-0.723) for LR. Variables related to intra-operative information, such as surgical approach, biliary reconstruction, and pancreatic gland texture were highly important for model predictions. DISCUSSION: Machine learning showed a marginal advantage over traditional statistical techniques in predicting procedure-specific outcomes. However, models that included intra-operative information performed better than those that did not, suggesting that NSQIP procedure-targeted datasets may be strengthened by including relevant intra-operative information.
© 2022. The Society for Surgery of the Alimentary Tract.

Entities:  

Keywords:  Anastomotic leak; Artificial intelligence; Hepatectomy; Machine learning; Pancreatic fistula

Mesh:

Year:  2022        PMID: 35508684      PMCID: PMC9444966          DOI: 10.1007/s11605-022-05332-x

Source DB:  PubMed          Journal:  J Gastrointest Surg        ISSN: 1091-255X            Impact factor:   3.267


  28 in total

1.  Using the NSQIP Pancreatic Demonstration Project to Derive a Modified Fistula Risk Score for Preoperative Risk Stratification in Patients Undergoing Pancreaticoduodenectomy.

Authors:  Olga Kantor; Mark S Talamonti; Henry A Pitt; Charles M Vollmer; Taylor S Riall; Bruce L Hall; Chi-Hsiung Wang; Marshall S Baker
Journal:  J Am Coll Surg       Date:  2017-04-10       Impact factor: 6.113

2.  Risk factors and consequences of anastomotic leak after colectomy: a national analysis.

Authors:  Emily F Midura; Dennis Hanseman; Bradley R Davis; Sarah J Atkinson; Daniel E Abbott; Shimul A Shah; Ian M Paquette
Journal:  Dis Colon Rectum       Date:  2015-03       Impact factor: 4.585

3.  Comparing the areas under two or more correlated receiver operating characteristic curves: a nonparametric approach.

Authors:  E R DeLong; D M DeLong; D L Clarke-Pearson
Journal:  Biometrics       Date:  1988-09       Impact factor: 2.571

4.  Development and Validation of a New Nomogram for Predicting Clinically Relevant Postoperative Pancreatic Fistula After Pancreatoduodenectomy.

Authors:  Xi-Tai Huang; Chen-Song Huang; Chen Liu; Wei Chen; Jian-Peng Cai; He Cheng; Xing-Xing Jiang; Li-Jian Liang; Xian-Jun Yu; Xiao-Yu Yin
Journal:  World J Surg       Date:  2020-09-08       Impact factor: 3.352

5.  Surgical Risk Is Not Linear: Derivation and Validation of a Novel, User-friendly, and Machine-learning-based Predictive OpTimal Trees in Emergency Surgery Risk (POTTER) Calculator.

Authors:  Dimitris Bertsimas; Jack Dunn; George C Velmahos; Haytham M A Kaafarani
Journal:  Ann Surg       Date:  2018-10       Impact factor: 12.969

Review 6.  Prediction of anastomotic leak in colorectal cancer surgery based on a new prognostic index PROCOLE (prognostic colorectal leakage) developed from the meta-analysis of observational studies of risk factors.

Authors:  S A Rojas-Machado; M Romero-Simó; A Arroyo; A Rojas-Machado; J López; R Calpena
Journal:  Int J Colorectal Dis       Date:  2015-10-27       Impact factor: 2.571

7.  Assessing the utility of deep neural networks in predicting postoperative surgical complications: a retrospective study.

Authors:  Alexander Bonde; Kartik M Varadarajan; Nicholas Bonde; Anders Troelsen; Orhun K Muratoglu; Henrik Malchau; Anthony D Yang; Hasan Alam; Martin Sillesen
Journal:  Lancet Digit Health       Date:  2021-06-29

8.  Video Ratings of Surgical Skill and Late Outcomes of Bariatric Surgery.

Authors:  Christopher P Scally; Oliver A Varban; Arthur M Carlin; John D Birkmeyer; Justin B Dimick
Journal:  JAMA Surg       Date:  2016-06-15       Impact factor: 14.766

9.  Development of a Risk Score to Predict Anastomotic Leak After Left-Sided Colectomy: Which Patients Warrant Diversion?

Authors:  Nicholas P McKenna; Katherine A Bews; Robert R Cima; Cynthia S Crowson; Elizabeth B Habermann
Journal:  J Gastrointest Surg       Date:  2019-06-26       Impact factor: 3.452

10.  Models predicting the risks of six life-threatening morbidities and bile leakage in 14,970 hepatectomy patients registered in the National Clinical Database of Japan.

Authors:  Hideki Yokoo; Hiroaki Miyata; Hiroyuki Konno; Akinobu Taketomi; Tatsuhiko Kakisaka; Norimichi Hirahara; Go Wakabayashi; Mitsukazu Gotoh; Masaki Mori
Journal:  Medicine (Baltimore)       Date:  2016-12       Impact factor: 1.817

View more

北京卡尤迪生物科技股份有限公司 © 2022-2023.