Our paper, "90-Day Mortality Prediction in Elective Visceral Surgery Using Machine Learning: A Retrospective Multicenter Development, Validation, and Comparison Study" has been published online ahead of print in the International Journal of Surgery.
Authors are C. Riepe, R. van de Water, A. Winter, B. Pfitzner, L. Faraj, R. Ahlborn, M. Schulze, D. Zuluaga, C. Schineis, K. Beyer, J. Pratschke, B. Arnrich, I.M. Sauer, and M.M. Maurer

Machine Learning (ML) is increasingly being adopted in biomedical research, however, its potential for outcome prediction in visceral surgery remains uncertain. This study compares the potential of ML methods for preoperative 90-day mortality (90DM) prediction of an aggregated multi-organ approach to conventional scoring systems and individual organ models.

This retrospective cohort study enrolled patients undergoing major elective visceral surgery between 2014 and 2022 across two tertiary centers. Multiple ML models for preoperative 90DM prediction were trained, externally validated and benchmarked against the American Society of Anesthesiologists (ASA) score and revised Charlson Comorbidity Index (rCCI). Areas under the receiver operating characteristic (AUROC) and precision recall curves (AUPRC) including standard deviations were calculated. Additionally, individual models for esophageal, gastric, intestinal, liver, and pancreatic surgery were developed and compared to an aggregated approach. A total of 7,711 cases encompassing 78 features were included. Overall 90DM was 4% (n = 309). An XBoost classifier demonstrated the best performance and high robustness following external validation (AUROC: 0.86 [0.01]; AUPRC: 0.2 [0.04]). All models outperformed the ASA score (AUROC: 0.72; AUPRC: 0.08) and rCCI (AUROC: 0.81; AUPRC: 0.11). rCCI, patient age and C-reactive protein emerged as most decisive model weights. Models for gastric (AUROC: 0.88 [0.13]; AUPRC: 0.24 [0.26]) and intestinal surgery (AUROC: 0.87 [0.05]; AUPRC: 0.17 [0.09]) revealed the highest organ-specific performances, while pancreatic surgery yielded the lowest results (AUROC: 0.66 [0.08]; AUPRC: 0.22 [0.12]). A combined multi-organ approach (AUROC: 0.84 [0.04]; AUPRC: 0.21 [0.06]) demonstrated superiority over the weighted average across all organ-specific models (AUROC: 0.82 [0.07]; AUPRC: 0.2 [0.13]).

ML offers robust preoperative risk stratification for 90DM in elective visceral surgery. Leveraging training across multi-organ cohorts may improve accuracy and robustness compared to organ-specific models. Prospective studies are needed to confirm the potential of ML in surgical outcome prediction.

We are delighted to welcome Dr. Zeynep Akbal as a new member of the team!
Before she started as a post-doctoral researcher at the Digital Surgery Lab she studied communication sciences, media sciences, philosophy, and worked on developing her interdisciplinary method around virtual reality (VR) technology. She did her doctorate in philosophy at Universität Potsdam. Her dissertation is recently published as a monograph, titled "Lived-Body Experiences in Virtual Reality. A Phenomenology of the Virtual Body.”
Her research focuses on the intersection of philosophy of perception, cognitive sciences and VR. In her recent research project “Tactile Stimulation in VR” at Max Planck Institute for Human Cognitive and Brain Sciences, she focused on the behavioral consequences of haptic feedback in a VR task.

Wellcome to the team!

The Federal Ministry of Education and Research (BMBF) funds the project "Tangible reality - skilful interaction of user hands and fingers with real tools in mixed reality worlds (GreifbAR)" – a cooperation of the Augmented Vision group of the DFKI (Prof. Dr. Didier Stricker), the Department of Psychology and Human-Machine Interaction of the University of Passau (Prof. Dr. Susanne Mayr), the company NMY Mixed Reality Communication (Christoph Lenk), and the Experimental Surgery of Charité – Universitätsmedizin Berlin (Prof. Dr. Igor M. Sauer).

The goal of the GreifbAR project is to make extended reality (XR) worlds, including virtual (VR) and mixed reality (MR), tangible and graspable by allowing users to interact with real and virtual objects with their bare hands. Hand accuracy and dexterity is paramount for performing precise tasks in many fields, but the capture of hand-object interaction in current XR systems is woefully inadequate. Current systems rely on hand-held controllers or capture devices that are limited to hand gestures without contact with real objects. GreifbAR solves this limitation by proposing a sensing system that detects both the full hand grip including hand surface and object pose when users interact with real objects or tools. This sensing system will be integrated into a mixed reality training simulator.

Competent handling of instruments and suture material is the basis of every surgical activity. The main instruments used in surgery are in the hands of the surgical staff. Their work is characterised by the targeted use of a large number of instruments that have to be operated and controlled in different ways. Until now, surgical knotting techniques have been learned by means of personal instruction by experienced surgeons, blackboard images and video-based tutorials. A training and teaching concept based on the acquisition of finger movement does not yet exist in surgical education and training. Learning surgical account techniques through participant observation and direct instruction by experienced surgeons is cost-intensive and hardly scalable. This type of training is increasingly reaching its limits in daily clinical practice, which can be attributed in particular to the changed economic, social and regulatory conditions in surgical practice. Students and trainees as well as specialist staff in further training are therefore faced with the problem of applying and practising acquired theoretical knowledge in a practice-oriented manner. Text- and image-based media allow scalable theoretical knowledge acquisition independent of time and place. However, gestures and work steps can only be passively observed and subsequently imitated. Moreover, the learning success cannot be quantitatively measured and verified.

The aim of the Charité's sub-project is therefore to develop a surgical application scenario for Mixed/Augmented Reality (MR/AR) for the spatial guidance and verifying recording of complex fine motor finger movements for the creation of surgical knots, the practical implementation and technical testing of the developed concept within the framework of a demonstrator, and the evaluation of the usability of the system for use in a clinical context.