Development of Machine Learning Model for VO2max Estimation Using a Patch-Type Single-Lead ECG Monitoring Device in Lung Resection Candidates - study summary

Aug 5, 2026

Upload by Noah

A cardiopulmonary exercise test (CPET) is essential for assessing suitability for lung resection, yet it can be difficult to perform in routine practice. This prospective single-center study developed a machine learning model to estimate maximal oxygen consumption (VO2max) using data from a patch-type single-lead ECG monitoring device. In 42 lung resection candidates undergoing CPET, algorithm-derived VO2max was compared with CPET results and with the FRIEND equation. Bland–Altman analysis showed smaller bias for the machine learning model, with improved consistency across maximal stage levels and sexes, supporting its use when CPET is not feasible.

LanguageEnglish

CategoryResearch & Report

FormatPDF

Pages14

Reading Time35 min

AccessFree to download

Abstract

Introduction

Clinical need for CPET in lung resection

Standards and limitations of VO2max assessment

Methods

Study design and patient cohort

ECG patch data and machine-learning algorithm

Reference comparisons and statistics

healthcare Article Development of Machine Learning Model for VO 2max Estimation Using a Patch-Type Single-Lead ECG Monitoring Device in Lung Resection Candidates Hyun Ah Lee 1,†, Woosik Yu 2,†, Jong Doo Choi 3, Young-sin Lee 3, Ji Won Park 1, Yun Jung Jung 1, Seung Soo Sheen 1, Junho Jung 2, Seokjin Haam 2, Sang Hun Kim 4, * and Ji Eun Park 1, * Citation: Lee, H.A.; Yu, W.; Choi, J.D.; Lee, Y.-s.; Park, J.W.; Jung, Y.J.; Sheen, S.S.; Jung, J.; Haam, S.; Kim, S.H.; et al. Development of Machine Learning Model for VO2max Estimation Using a Patch-Type Single-Lead ECG Monitoring Device in Lung Resection Candidates. Healthcare 2023, 11, 2863 . [https://doi.org/10.3390/](https://doi.org/10.3390/)[ ](https://doi.org/10.3390/)healthcare11212863 Academic Editors: Abbas Edalat and Ramin Ramezani Received: 19 September 2023 Revised: 27 October 2023 Accepted: 29 October 2023 Published: 30 October 2023 Copyright: © 2023 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license ([https://](https://)[ ](https://)[creativecommons.org/licenses/by/](creativecommons.org/licenses/by/)[ ](creativecommons.org/licenses/by/)[4.0/](4.0/)) . 1 Department of Pulmonary and Critical Care Medicine, Ajou University School of Medicine, Suwon 16499, Republic of Korea 2 Department of Thoracic and Cardiovascular Surgery, Ajou University School of Medicine, Suwon 16499, Republic of Korea; [yws081011@aumc.ac.kr](yws081011@aumc.ac.kr) (W.Y.) 3 Seers Technology Co., Seongnam-si 13558, Republic of Korea 4 Department of Rehabilitation Medicine, Biomedical Research Institute, Pusan National University Hospital, Busan 49241, Republic of Korea * Correspondence: [kel5504@gmail.com](kel5504@gmail.com) (S.H.K.); [petitprince012@ajou.ac.kr](petitprince012@ajou.ac.kr) (J.E.P.); Tel.: +82-51-240-7485 (S.H.K.); +82-31-219-5096 (J.E.P.)† These authors contributed equally to this work. Abstract: A cardiopulmonary exercise test (CPET) is essential for lung resection. However, performing a CPET can be challenging. This study aimed to develop a machine learning model to estimate maximal oxygen consumption (VO 2max) using data collected through a patch-type single-lead electrocardiogram (ECG) monitoring device in candidates for lung resection. This prospective, single-center study included 42 patients who underwent a CPET at a tertiary teaching hospital from October 2021 to July 2022 . During the CPET, a single-lead ECG monitoring device was applied to all patients, and the results obtained from the machine-learning algorithm using the information extracted from the ECG patch were compared with the CPET results. According to the Bland–Altman plot of measured and estimated VO2max, the VO2max values obtained from the machine learning model and the FRIEND equation showed lower differences from the reference value (bias: 􀀀0 .33 mL􀀁kg􀀀1􀀁min􀀀1, bias: 0.30 mL􀀁kg􀀀1􀀁min􀀀1, respectively) . In subgroup analysis, the developed model demonstrated greater consistency when applied to different maximal stage levels and sexes. In conclusion, our model provides a closer estimation of VO 2max values measured using a CPET than existing equations. This model may be a promising tool for estimating VO 2max and assessing cardiopulmonary reserve in lung resection candidates when a CPET is not feasible. Keywords: maximal oxygen consumption (VO 2max); cardiopulmonary exercise test (CPET); machine learning model; estimation; lung resection candidates 1. Introduction Lung cancer is the second most commonly diagnosed cancer and the leading cause of cancer-related mortality [1] . Surgical resection is considered the best curative option for lung cancer [2] . However, lung cancer patients are often elderly, have weakened lung function due to smoking, or have underlying medical conditions [3,4] . Although surgery may be feasible based on the cancer

FAQs

What problem does the study address regarding VO2max estimation?

CPET is important for lung resection assessment but can be challenging to perform. The study targets VO2max estimation when CPET is not feasible.

How was VO2max estimated in this research?

A patch-type single-lead ECG monitoring device collected ECG data during CPET, which was fed into a machine-learning algorithm to generate estimated VO2max values.

How did the machine learning model perform compared with the FRIEND equation?

Bland–Altman analysis showed the machine learning model had smaller differences from CPET reference VO2max than the FRIEND equation, indicating closer estimation.

Aug 5, 2026

Upload by Noah