Experts vs. machine - comparison of machine learning to expert-informed prediction of outcome after major liver surgery.

BACKGROUND: Machine learning (ML) has been successfully implemented for classification tasks (e.g., cancer diagnosis). ML performance for more challenging predictions is largely unexplored. This study's objective was to compare machine learning vs. expert-informed predictions for surgical outcome in patients undergoing major liver surgery. METHODS: Single tertiary center data on preoperative parameters and postoperative complications for elective hepatic surgery patients were included (2008-2021). Expert-informed prediction models were established on 14 parameters identified by two expert liver surgeons to impact on postoperative outcome. ML models used all available preoperative patient variables (n = 62). Model performance was compared for predicting 3-month postoperative overall morbidity. Temporal validation and additional analysis in major liver resection patients were conducted. RESULTS: 889 patients included. Expert-informed models showed low average bias (2-5 CCI points) with high over/underprediction. ML models performed similarly: average prediction 5-10 points higher than observed CCI values with high variability (95% CI -30 to 50). No performance improvement for major liver surgery patients. CONCLUSION: No clinical relevance in the application of ML for predicting postoperative overall morbidity was found. Despite being a novel hype, ML has the potential for application in clinical practice. However, at this stage it does not replace established approaches of prediction modelling.

Effect of donor pancreas extraction time on pancreas transplantation-a Swiss tertiary center experience.

We aimed to assess the effect of donor pancreas extraction time (ET) on postoperative complications and graft function after pancreas transplantation (PT). We analyzed all consecutive donor pancreas procurements for the simultaneous pancreas and kidney transplantation (SPK) and the associated PT in a Swiss transplant center over a 20-year period. Pancreas ET was defined as the time from cold flush to static storage of the pancreas on ice. The primary endpoint was the effect of extraction time on surgical complications. Secondary endpoints comprised the effect of ET on graft function (insulin-free survival) and graft pancreatitis. Of 115 procured pancreas grafts the median donor pancreas ET was 65 min (IQR: 48-78 min). In multivariable analysis, ET did not negatively affect major complications (OR 1.41 [95% CI: .59-3.36]; p = .438) and insulin-free survival (HR 1.42 [95% CI: .55-3.63]; p = .459). The median CIT was 522 (441-608) min. CIT was associated with major complications (OR 2.51 [95% CI: 1.11-5.68]; p = .027), but without impact on insulin-free survival (HR 1.94 [95% CI: .84-4.48]; p = .119). Patients with and without graft pancreatitis had no statistically significant differences in ET and CIT (p = .164 and p = .47, respectively). In multivariable analysis, Amylase levels > 270 U/L on postoperative day 1 were significantly associated with major complications (OR 3.61 [95% CI: 1.06-12.32]; p = .040). Our results suggest that although no effect of ET on complications and graft function after PT was found, shorter CIT and less graft pancreatitis can have a positive impact on surgical complications. Results could possibly be influenced by the exceptional quality of the pancreas donors, with short travel distances and preservation times in Switzerland.