Retrieval-Based Disease Prediction for Myocardial Injury after Noncardiac Surgery: Leveraging Language Models as Diagnostic Tools

Namjun Park, Donggeun Ko, Dongjun Lee, San Kim, Jaekwang KIM

Track: Traditional track

Keywords: Myocardial Injury after Noncardiac Surgery, Language Model, Disease Prediction, Natural Language Processing

TL;DR: The study introduces a Retrieval Based Disease (RBD) Prediction framework. This method surpasses traditional machine learning, particularly in managing datasets with class imbalances, enhancing predictive diagnostics in healthcare.

Abstract: Predicting Myocardial Injury after Noncardiac Surgery (MINS) is crucial for enhancing patient outcomes, as these injuries significantly affect health and survival rates. This study presents a novel approach for MINS prediction by transforming and converting collected comprehensive pre-operative and intra-operative medical data into a textual description format compatible with Language Models (LM). We employ a Retrieval Based Disease Prediction (RBD) framework, leveraging advanced natural language processing (NLP) techniques to interpret complex patient information. Our results demonstrate that this LM-based approach outperforms traditional machine learning methods. Furthermore, our findings indicate that leveraging LMs with medical data improves predictive performance and potentially enhances patient care and postoperative outcomes. Moreover, the versatility of the RBD framework in adapting to various medical data types highlights its potential as a transformative tool and a stepping stone in healthcare analytics and predictive diagnostics.

Presentation And Attendance Policy: I have read and agree with the symposium's policy on behalf of myself and my co-authors.

Ethics Board Approval: Yes, we have/will include(d) information about IRB approval or its equivalent, in the manuscript.

Data And Code Availability: Yes, we will make data and code available upon acceptance.

Primary Area: Clinical foundation models

Student First Author: Yes, the primary author of the manuscript is a student.

Submission Number: 39