Using Machine Learning to Predict Postoperative Liver Dysfunction After Aortic Arch Surgery

Risk assessment of hyperbilirubinemia using a three-factor model after cardiac surgery

BACKGROUND

Hyperbilirubinemia (HB) is a common occurrence after cardiopulmonary bypass and often results in increased rates of complications.

METHODS

We conducted a study on 411 patients who underwent cardiac surgery with extracorporeal circulation and divided them into a training set and a validation set. Least absolute shrinkage and selection operator (LASSO) regression was employed to screen out variables. Multivariate logistic regression was subsequently applied to establish prediction models, which were then evaluated using receiver operating characteristic (ROC) curves and calibration plots. Finally, restricted cubic splines (RCSs) curve pairs were used to calculate the adjusted ORs for each independent risk factor.

RESULTS

A clinical prediction model based on preoperative bilirubin elevation, aortic clamp time and intraoperative red blood cell (RBC) transfusion was successfully developed. The model presented an ROC of 0.855 (0.808, 0.901) in the training set. The validation set demonstrated a promising ROC of 0.839 (0.761, 0.917), and the calibration curve closely approximated the true incidence. The risk of postoperative HB was increased when the aortic clamp time was greater than 133 min, when the red blood transfusion volume was greater than 275 ml or when the preoperative total bilirubin concentration was greater than 16.65 mol/L.

CONCLUSION

In this study, the aortic clamp time was identified as a crucial factor for postoperative HB during cardiac surgery. A three-factor prediction model that includes the aortic clamp time for the prediction of HB after cardiac surgery was successfully established. Furthermore, patients with hyperbilirubinemia post-surgery had a worse short-term prognosis.

Prediction of Complications and Prognostication in Perioperative Medicine: A Systematic Review and PROBAST Assessment of Machine Learning Tools

BACKGROUND

The utilization of artificial intelligence and machine learning as diagnostic and predictive tools in perioperative medicine holds great promise. Indeed, many studies have been performed in recent years to explore the potential. The purpose of this systematic review is to assess the current state of machine learning in perioperative medicine, its utility in prediction of complications and prognostication, and limitations related to bias and validation.

METHODS

A multidisciplinary team of clinicians and engineers conducted a systematic review using the Preferred Reporting Items for Systematic Review and Meta-Analysis (PRISMA) protocol. Multiple databases were searched, including Scopus, Cumulative Index to Nursing and Allied Health Literature (CINAHL), the Cochrane Library, PubMed, Medline, Embase, and Web of Science. The systematic review focused on study design, type of machine learning model used, validation techniques applied, and reported model performance on prediction of complications and prognostication. This review further classified outcomes and machine learning applications using an ad hoc classification system. The Prediction model Risk Of Bias Assessment Tool (PROBAST) was used to assess risk of bias and applicability of the studies.

RESULTS

A total of 103 studies were identified. The models reported in the literature were primarily based on single-center validations (75%), with only 13% being externally validated across multiple centers. Most of the mortality models demonstrated a limited ability to discriminate and classify effectively. The PROBAST assessment indicated a high risk of systematic errors in predicted outcomes and artificial intelligence or machine learning applications.

CONCLUSIONS

The findings indicate that the development of this field is still in its early stages. This systematic review indicates that application of machine learning in perioperative medicine is still at an early stage. While many studies suggest potential utility, several key challenges must be first overcome before their introduction into clinical practice.

EDITOR’S PERSPECTIVE

Machine-Learning Models for Predicting Surgical Site Infections using Patient Pre-Operative Risk and Surgical Procedure Factors

BACKGROUND

Surgical site infections (SSIs) are a significant healthcare problem as they can cause increased medical costs and increased morbidity and mortality. Assessing a patient's pre-operative risk factors can improve risk stratification and help guide the surgical decision-making process. Previous efforts to use pre-operative risk factors to predict the occurrence of SSIs have relied upon traditional statistical modeling approaches. The aim of this paper is to develop and validate, using state-of-the-art machine learning (ML) approaches, classification models for the occurrence of SSI to improve upon previous models.

METHODS

In this work, using the American College of Surgeons' National Surgical Quality Improvement Program (ACS NSQIP) database, the performances (e.g., prediction accuracy) of seven different ML approaches (Logistic Regression (LR), Naïve Bayesian (NB), Random Forest (RF), Decision Tree (DT), Support Vector Machine (SVM), Artificial Neural Network (ANN), and Deep Neural Network (DNN)) were compared. The performance of these models was evaluated using the area under the curve, accuracy, precision, sensitivity, and F1-score metrics.

RESULTS

Overall, 2,882,526 surgical procedures were identified in the study for the SSI predictive models' development. The results indicate that the DNN model offers the best predictive performance with 10-fold compared to the other six approaches considered (area under the curve = 0.8518, accuracy = 0.8518, precision = 0.8517, sensitivity = 0.8527, F1-score = 0.8518). Emergency case surgeries, American Society of Anesthesiologists (ASA) Index of 4 (ASA_4), BMI, Vascular surgeries, and general surgeries were most significant influencing features towards developing an SSI.

CONCLUSION

Equally important is that the commonly used LR approach for SSI prediction displayed mediocre performance. The results are encouraging as they suggest that the prediction performance for SSIs can be improved using modern ML approaches.

Artificial intelligence and machine learning in aortic disease

PURPOSE OF REVIEW

Discuss foundational concepts for artificial intelligence (AI) and review recent literature on its application to aortic disease.

RECENT FINDINGS

Machine learning (ML) techniques are rapidly evolving for the evaluation of aortic disease - broadly categorized as algorithms for aortic segmentation, detection of pathology, and risk stratification. Advances in deep learning, particularly U-Net architectures, have revolutionized segmentation of the aorta and show potential for monitoring the size of aortic aneurysm and characterizing aortic dissection. These algorithms also facilitate application of more complex technologies including analysis of flow dynamics with 4D Flow magnetic resonance imaging (MRI) and computational simulation of fluid dynamics for aortic coarctation. In addition, AI algorithms have been proposed to assist in 'opportunistic' screening from routine imaging exams, including automated aortic calcification score, which has emerged as a strong predictor of cardiovascular risk. Finally, several ML algorithms are being explored for risk stratification of patients with aortic aneurysm and dissection, in addition to prediction of postprocedural complications.

SUMMARY

Multiple ML techniques have potential for characterization and risk prediction of aortic aneurysm, dissection, coarctation, and atherosclerotic disease on computed tomography and MRI. This nascent field shows considerable promise with many applications in development and in early preclinical evaluation.