Prediction of low cardiac output syndrome in patients following cardiac surgery using machine learning

A review of evaluation approaches for explainable AI with applications in cardiology

Explainable artificial intelligence (XAI) elucidates the decision-making process of complex AI models and is important in building trust in model predictions. XAI explanations themselves require evaluation as to accuracy and reasonableness and in the context of use of the underlying AI model. This review details the evaluation of XAI in cardiac AI applications and has found that, of the studies examined, 37% evaluated XAI quality using literature results, 11% used clinicians as domain-experts, 11% used proxies or statistical analysis, with the remaining 43% not assessing the XAI used at all. We aim to inspire additional studies within healthcare, urging researchers not only to apply XAI methods but to systematically assess the resulting explanations, as a step towards developing trustworthy and safe models.

Supplementary Information

The online version contains supplementary material available at 10.1007/s10462-024-10852-w.

Protected cardiac surgery: strategic mechanical circulatory support to improve postcardiotomy mortality

PURPOSE OF REVIEW

To examine the evolving landscape of cardiac surgery, focusing on the increasing complexity of patients and the role of mechanical circulatory support (MCS) in managing perioperative low cardiac output syndrome (P-LCOS).

RECENT FINDINGS

P-LCOS is a significant predictor of mortality in cardiac surgery patients. Preoperative risk factors, such as cardiogenic shock and elevated lactate levels, can help identify those at higher risk. Proactive use of MCS, rather than reactive implementation after P-LCOS develops, may lead to improved outcomes by preventing severe organ hypoperfusion. The emerging concept of "protected cardiac surgery" emphasizes early identification of these high-risk patients and planned MCS utilization. Additionally, specific MCS strategies are being developed and refined for various cardiac conditions, including AMI-CS, valvular surgeries, and pulmonary thromboendarterectomy.

SUMMARY

This paper explores the shifting demographics and complexities in cardiac surgery patients. It emphasizes the importance of proactive, multidisciplinary approaches to identify high-risk patients and implement early MCS to prevent P-LCOS and improve outcomes. The concept of protected cardiac surgery, involving planned MCS use and shared decision-making, is highlighted. The paper also discusses MCS strategies tailored to specific cardiac procedures and the ethical considerations surrounding MCS implementation.

A novel interpretative tool for early prediction of low cardiac output syndrome after valve surgery: online machine learning models

OBJECTIVE

Low cardiac output syndrome (LCOS) is a severe complication after valve surgery, with no uniform standard for early identification. We developed interpretative machine learning (ML) models for predicting LCOS risk preoperatively and 0.5 h postoperatively for intervention in advance.

METHODS

A total of 2218 patients undergoing valve surgery from June 2019 to Dec 2021 were finally enrolled to construct preoperative and postoperative models. Logistic regression, support vector machine (SVM), random forest classifier, extreme gradient boosting, and deep neural network were executed for model construction, and the performance of models was evaluated by area under the curve (AUC) of the receiver operating characteristic and calibration curves. Our models were interpreted through SHapley Additive exPlanations, and presented as an online tool to improve clinical operability.

RESULTS

The SVM algorithm was chosen for modeling due to better AUC and calibration capability. The AUCs of the preoperative and postoperative models were 0.786 (95% CI 0.729-0.843) and 0.863 (95% CI 0.824-0.902), and the Brier scores were 0.123 and 0.107. Our models have higher timeliness and interpretability, and wider coverage than the vasoactive-inotropic score, and the AUC of the postoperative model was significantly higher. Our preoperative and postoperative models are available online at http://njfh-yxb.com.cn:2022/lcos.

CONCLUSIONS

The first interpretable ML tool with two prediction periods for online early prediction of LCOS risk after valve surgery was successfully built in this study, in which the SVM model has the best performance, reserving enough time for early precise intervention in critical care.

Analysis of P(v-a)CO2/C(a-v)O2 Ratio and Other Perfusion Markers in a Population of 98 Pediatric Patients Undergoing Cardiac Surgery

BACKGROUND

The so-called Low Cardiac Output Syndrome (LCOS) is one of the most common complications in pediatric patients with congenital heart disease undergoing corrective surgery. LCOS requires high concentrations of inotropes to support cardiac contractility and improve cardiac output, allowing for better systemic perfusion. To date, serum lactate concentrations and central venous oxygen saturation (ScVO2) are the most commonly used perfusion markers, but they are not completely reliable in identifying a state of global tissue hypoxia. The study aims to evaluate whether the venoarterial carbon dioxide difference/arterial-venous oxygen difference ratio [P(v-a)CO2/C(a-v)O2] can be a good index to predict the development of LCOS in the aforementioned patients, so as to treat it promptly.

METHODS

This study followed a population of 98 children undergoing corrective cardiac surgery from June 2018 to October 2020 at the Department of Cardiac Surgery of University Hospital Integrated Trust and their subsequent admission at the Postoperative Cardiothoracic Surgery Intensive Care Unit. During the study, central arterial and venous blood gas analyses were carried out before and after cardiopulmonary bypass (CPB) (pre-CPB and post-CPB), at admission to the intensive care unit, before and after extubation, and at any time of instability or modification of the patient's clinical and therapeutic conditions.

RESULTS

The data analysis shows that 46.9% of the children developed LCOS (in line with the current literature) but that there is no statistically significant association between the P(v-a)CO2/C(a-v)O2 ratio and LCOS onset. Despite the limits of statistical significance, however, a 31% increase in the ratio emerged from the pre-CPB phase to the post-CPB phase when LCOS is present.

CONCLUSIONS

This study confirms a statistically significant association between the most used markers in adult patients (serum lactate concentration, ScVO2, and oxygen extraction ratio-ERO2) measured in the pre-CPB phase and the incidence of LCOS onset, especially in patients with hemodynamic instability before surgery.