A growing number of clinical risk scores have been proposed to predict allograft failure after liver transplantation. However, validation of currently available scores in the Eurotransplant region is still lacking. We aimed to analyze all clinically relevant donor and recipient risk scores on a large German liver transplantation data set and performed a retrospective cohort analysis of liver transplantations performed at the Charité—Universitätsmedizin Berlin from January 2007 until December 2021 with organs from donation after brain death. We analyzed 9 previously published scores in 906 liver transplantations [Eurotransplant donor risk index (ET-DRI/DRI), donor age and model for end-stage liver disease (D-MELD), balance of risk (BAR), early allograft dysfunction (EAD), model for early allograft function (MEAF), liver graft assessment following transplantation (L-GrAFT₇), early allograft failure simplified estimation (EASE), and a score by Rhu and colleagues). The EASE score had the best predictive value for 3-month, 6-month, and 12-month graft survival with a c-statistic of 0.8, 0.77, and 0.78, respectively. In subgroup analyses, the EASE score was suited best for male recipients with a high-MELD (>25) and an EAD organ. Scores only based on pretransplant data performed worse compared to scores including postoperative data (eg, ET-DRI vs. EAD, p<0.001 at 3-month graft survival). Out of these, the BAR score performed best with a c-statistic of 0.6. This a comprehensive comparison of the clinical utility of risk scores after liver transplantation. The EASE score sufficiently predicted 12-month graft and patient survival. Despite a relatively complex calculation, the EASE score provides significant prognostic value for patients and health care professionals in the Eurotransplant region.

In recent decades, robotic-assisted minimally invasive esophagectomy (RAMIE) has been increasingly adopted for patients with esophageal cancer (EC) or cancer of the gastroesophageal junction (GEJ). However, concerns regarding its costs compared to conventional minimally invasive esophagectomy (MIE) have emerged. This study examined outcomes and costs of RAMIE versus total MIE in 128 patients who underwent Ivor Lewis esophagectomy for EC/GEJ at our department between 2017 and 2021. Surgical costs were higher for RAMIE (EUR 12,370 vs. EUR 10,059, p < 0.001). Yet, median daily (EUR 2023 vs. EUR 1818, p = 0.246) and total costs (EUR 30,510 vs. EUR 29,180, p = 0.460) were comparable. RAMIE showed a lower incidence of postoperative pneumonia (8% vs. 25%, p = 0.029) and a trend towards shorter hospital stays (15 vs. 17 days, p = 0.205), which may have equalized total costs. Factors independently associated with higher costs included readmission to the intensive care unit (hazard ratio [HR] = 7.0), length of stay (HR = 13.5), anastomotic leak (HR = 17.0), and postoperative pneumonia (HR = 5.4). In conclusion, RAMIE does not impose an additional financial burden. This suggests that RAMIE may be considered as a valid alternative approach for esophagectomy. Attention to typical cost factors can enhance postoperative care across surgical methods.

Non-adherence to immunosuppressant therapy reduces long-term graft and patient survival after solid organ transplantation. The objective of this 24-month prospective study was to determine adherence, efficacy and safety after conversion of stable liver transplant (LT) recipients from a standard twice-daily immediate release Tacrolimus (IR-Tac) to a novel once-daily life cycle pharma Tacrolimus (LCP-Tac) formulation. We converted a total of 161 LT patients at baseline, collecting Tacrolimus trough levels, laboratories, physical examination data and the BAASIS^© questionnaire for self-reported adherence to immunosuppression at regular intervals. With 134 participants completing the study period (17% dropouts), the overall adherence to the BAASIS^© increased by 57% until month 24 compared to baseline (51% vs. 80%). Patients who required only a morning dose of their concomitant medications reported the largest improvement in adherence after conversion. The intra-patient variability (IPV) of consecutive Tacrolimus trough levels after conversion did not change significantly compared to pre-conversion levels. Despite reducing the daily dose by 30% at baseline as recommended by the manufacturer, Tac-trough levels remained stable, reflected by an increase in the concentration-dose (C/D) ratio. No episodes of graft rejection or loss occurred. Our data suggest that the use of LCP-Tac in liver transplant patients is safe and can increase adherence to immunosuppression compared to conventional IR-Tac.

Background: Limited data availability impedes the advancements of artificial intelligence (AI) applications in surgery to date. Federated learning (FL), a novel privacy-focused machine learning technique, introduces a decentralized framework to facilitate multicenter modeling. This study constructs a FL network, including differential privacy (DP), to predict 90-day mortality following colorectal surgery. Methods: Patients undergoing elective colorectal surgery across three tertiary centers (C1–C3) between January 2015 and December 2021 were retrospectively enrolled. Neural networks (NN) for mortality prediction were trained and validated for all three centers individually and after data aggregation using centralized and distributed FL analysis. Local and central DP was then applied as additional data protection framework components. Areas under the receiver operating characteristic (AUROC) and precision-recallcurves (AUPRC) including 95%-confidence intervals were calculated. Results: A total of 2959 patients (mean [SD] age: 56.9 [16.8] years; n = 1677 [56.7%] male) were enrolled. The 90-day mortality rate was 3.1% (n = 92). Local NNs achieved AUROCs of 0.80 ([0.74–0.87]; C1), 0.81 ([0.75–0.87]; C2), and 0.84 ([0.75–0.92]; C3) and corresponding AUPRCs reaching 0.26 [0.11–0.41], 0.31 [0.21–0.41], and 0.21 [0.11–0.32], respectively. An aggregated centralized NN (NN _CZ ) achieved an AUROC of 0.81 [0.76–0.85] and AUPRC of 0.29 [0.22–0.36]. Distributed FL (NN _FL ) was comparable (AUROC: 0.78 [0.72–0.84], P = 0.67; AUPRC: 0.26 [0.17–0.35], P = 0.44) to the centralized model. Central DP reduced the performance of FL-based prediction by 5% in AUROC (0.74 [0.64–0.84]) and 35% in AUPRC (0.17 [0.10–0.23]). Local DP almost diminished the performance (AUROC: 0.52 [0.49–0.55]; AUPRC: 0.05 [0.03–0.06]). Feature importance analysis revealed age, blood status and the Charlson Comorbidity Index as highest weighted features for both NN _CZ and NN _FL . Conclusion: Federated learning demonstrates similar performance to centralized machine learning in preoperative mortality prediction, providing an encouraging framework to accelerate the development of future surgical AI applications. Improving data privacy through DP is associated with compromises in model performance.

Introduction: Cytoreductive surgery (CRS) coupled with hyperthermic intraperitoneal chemotherapy (HIPEC) is a potentially curative approach to peritoneal disease (PD) and corresponds to the needs of patients with dire survival rates. However, the oncological community remains cautious toward this procedure because of its significant morbidity and mortality rates. This evolving evidence around CRS and HIPEC and its impact suggests a need for more standardized procedures in existing centers. Because of its complexity and potential for high morbidity and mortality, critical observation of our center’s complication rates using complication management documentation tools were crucial to further develop our standard operating procedures (SOP) and maximize patient safety. Methods: Our prospectively maintained institutional database was queried to identify all patients who underwent CRS and HIPEC and had a filled-out quality management (QM) and complication management documentation tool at discharge at the surgical department of the Charité—Universitätsmedizin Berlin, Campus Virchow Klinikum, Berlin, Germany, between January 2018 and December 2023. Results: A total of 155 patients had a surgical and/or medical complication recorded. In total, 305 surgeries were surveyed. Some patients had more than one postoperative complication; hence, 344 events in 50 categories were recorded, of which 267 were graded 3a or higher in 92 patients. The most common medical complications were of pulmonary and renal origin. On the surgical side, surgical site infections (SSI) were most common. The incidence of anastomotic leakage (AL) was 5% (n = 8), with no events between 2021 and the present. Patients with longer surgery duration times were at higher risk for developing postoperative complications. Conclusions: Major abdominal surgeries like CRS and HIPEC are associated with significant patient morbidity despite achieving optimal oncological outcomes. Postoperative complications are managed through strict surveillance and transparency, particularly in our large reference centers, to minimize patient risk. Quality management programs in our department have successfully maintained high standards of care without compromising patient safety.

Background: 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. Methods: 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. Results: 7711 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]). Conclusion: 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.