Search

Machine learning (ML) is increasingly applied to predict adverse postoperative outcomes in cardiac surgery. Commonly used ML models fail to translate to clinical practice due to absent model explainability, limited uncertainty quantification, and no flexibility to missing data. We aimed to develop and benchmark a novel ML approach, the uncertainty-aware attention network (UAN), to overcome these common limitations. Two Bayesian uncertainty quantification methods were tested, generalized variational inference (GVI) or a posterior network (PN). The UAN models were compared with an ensemble of XGBoost models and a Bayesian logistic regression model (LR) with imputation. The derivation datasets consisted of 153,932 surgery events from the Australian and New Zealand Society of Cardiac and Thoracic Surgeons (ANZSCTS) Cardiac Surgery Database. An external validation consisted of 7343 surgery events which were extracted from the Medical Information Mart for Intensive Care (MIMIC) III critical care dataset. The highest performing model on the external validation dataset was a UAN-GVI with an area under the receiver operating characteristic curve (AUC) of 0.78 (0.01). Model performance improved on high confidence samples with an AUC of 0.81 (0.01). Confidence calibration for aleatoric uncertainty was excellent for all models. Calibration for epistemic uncertainty was more variable, with an ensemble of XGBoost models performing the best with an AUC of 0.84 (0.08). Epistemic uncertainty was improved using the PN approach, compared to GVI. UAN is able to use an interpretable and flexible deep learning approach to provide estimates of model uncertainty alongside state-of-the-art predictions. The model has been made freely available as an easy-to-use web application demonstrating that by designing uncertainty-aware models with innately explainable predictions deep learning may become more suitable for routine clinical use., Competing Interests: The authors have declared that no competing interests exist., (Copyright: © 2023 Penny-Dimri et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.)

Purpose: Trends towards surgical sub-specialisation to improve patient-outcomes are well-documented and largely supported by evidence. However few studies have examined whether this benefit exists within adult-cardiac surgery. To answer whether sub-specialisation within adult-cardiac surgery improves patient-outcomes, this study assessed the relationship between procedure-specific and total-cardiac surgeon-volume and mortality and morbidity in cardiac-valve and coronary artery bypass grafting (CABG) surgery., Methods: Data came from the Australian and New Zealand Society of Cardiac and Thoracic Surgeons (ANZSCTS) registry from 2001 to 2010 and included 23 hospitals, 109 surgeons, 20,619 patients with isolated-CABG-surgery and 11,536 patients with a valve-procedure. Hierarchical logistic regression using generalised estimating equations was used to analyse outcomes. Measures included operative-mortality and occurrence of a complication (deep sternal wound infection, new stroke, acute kidney injury)., Results: Crude operative mortality (and complication rates) were 1.7% (4.9%) and 4% (11%) in the isolated-CABG and valve-surgical populations respectively. A greater procedure-specific surgeon volume was associated with reduced mortality and complication rates in valve-surgery but not isolated-CABG. There was a 33% decrease in odds of dying for every additional 50 valve procedures performed [OR 0.67, p=0.003]. Conversely, greater total-cardiac surgical volume for individual surgeons did not result in improved outcomes, for both isolated-CABG and valve populations., Conclusions: Our finding of an association between increased valve-specific surgeon volumes with improved valve-surgery outcomes, and absence of an association between these outcomes and annual total-cardiac surgical experience supports the case for sub-specialisation specifically within the field of valve surgery., (Copyright © 2014 Australian and New Zealand Society of Cardiac and Thoracic Surgeons (ANZSCTS) and the Cardiac Society of Australia and New Zealand (CSANZ). Published by Elsevier B.V. All rights reserved.)

Objective: To report the first "domino-donor" operation performed in Australia; review the results of the first five combined procedures at the Alfred Hospital; and discuss the advantages of the procedure and its place in transplantation., Patients and Outcome: The first domino-donor was a 25-year-old man with cystic fibrosis and pulmonary hypertension, receiving a heart-lung transplant, whose heart was transplanted into a 20-year-old woman with end-stage cardiomyopathy. The cardiac recipient's initial course was satisfactory, but at eight months she developed severe rejection, with failure of the donor heart, and required retransplantation. Subsequently four further combined procedures were performed. All 10 patients are alive and well. Four patients have had cytomegalovirus infection but without major complications. Improved function of the donor right ventricle has been observed when the ventricle is working against a lower resistance., Conclusions: Four advantages of the domino procedure are evident: an increased pool of donor allografts; the use of a heart with an hypertrophied ("prepared") right ventricle; absence of the adverse effects of brain-death in the live donor; and improved ability to match the donor and recipient before the operation.