Division of Gastrointestinal Surgery, Department of Surgery, University of North Carolina, 101 Manning Drive, Burnett Womack Building, Suite 4038, Chapel Hill, NC, 27599, USA.
Department of Pharmacology, University of North Carolina, 120 Mason Farm Rd, Genetic Medicine Building, Chapel Hill, NC, 27599, USA.
J Gastrointest Surg. 2022 Aug;26(8):1732-1742. doi: 10.1007/s11605-022-05332-x. Epub 2022 May 4.
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).
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).
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.
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.
特定手术的并发症可能会带来毁灭性的后果。基于机器学习的工具在预测其风险和指导决策方面有可能优于传统的统计建模。我们试图开发和比较深度学习神经网络 (NN) 模型，这是一种机器学习，用于预测结肠切除术吻合口漏、肝切除术胆漏和胰十二指肠切除术 (PD) 后的胰腺瘘。
对结肠切除术、肝切除术和 PD 国家手术质量改进计划 (NSQIP) 数据库进行了分析。每个数据集都以 60/20/20 的比例分割为训练、验证和测试集，并进行五重交叉验证。为每个结果使用 NN 和 LR 创建模型。主要使用接受者操作特征曲线下的面积 (AUROC) 来评估模型。
共纳入 197488 例结肠切除术、25403 例肝切除术和 23333 例 PD 患者。对于吻合口漏，NN 的 AUROC 为 0.676（95%置信区间 0.666-0.687），而 LR 的 AUROC 为 0.633（95%置信区间 0.620-0.647）。对于胆漏，NN 的 AUROC 为 0.750（95%置信区间 0.739-0.761），而 LR 的 AUROC 为 0.722（95%置信区间 0.698-0.746）。对于胰腺瘘，NN 的 AUROC 为 0.746（95%置信区间 0.733-0.760），而 LR 的 AUROC 为 0.713（95%置信区间 0.703-0.723）。与术中信息相关的变量，如手术方法、胆道重建和胰腺质地，对模型预测非常重要。
机器学习在预测特定手术的结果方面显示出优于传统统计技术的微小优势。然而，包含术中信息的模型比不包含术中信息的模型表现更好，这表明包括相关术中信息可能会增强 NSQIP 针对特定手术的数据集。
