Two-Stage Classification of Future Knee Osteoarthritis Severity After 8 Years Using MRI: Data from the Osteoarthritis Initiative.

IF 3 2区 医学 Q3 ENGINEERING, BIOMEDICAL Annals of Biomedical Engineering Pub Date : 2024-07-09 DOI:10.1007/s10439-024-03578-x
Teemu A T Nurmirinta, Mikael J Turunen, Rami K Korhonen, Jussi Tohka, Mimmi K Liukkonen, Mika E Mononen
{"title":"Two-Stage Classification of Future Knee Osteoarthritis Severity After 8 Years Using MRI: Data from the Osteoarthritis Initiative.","authors":"Teemu A T Nurmirinta, Mikael J Turunen, Rami K Korhonen, Jussi Tohka, Mimmi K Liukkonen, Mika E Mononen","doi":"10.1007/s10439-024-03578-x","DOIUrl":null,"url":null,"abstract":"<p><p>Currently, there are no methods or tools available in clinical practice for classifying future knee osteoarthritis (KOA). In this study, we aimed to fill this gap by classifying future KOA into three severity grades: KL01 (healthy), KL2 (moderate), and KL34 (severe) based on the Kellgren-Lawrance scale. Due to the complex nature of multiclass classification, we used a two-stage method, which separates the classification task into two binary classifications (KL01 vs. KL234 in the first stage and KL2 vs. KL34 in the second stage). Our machine learning (ML) model used two Balanced Random Forest algorithms and was trained with gender, age, height, weight, and quantitative knee morphology obtained from magnetic resonance imaging. Our training dataset comprised longitudinal 8-year follow-up data of 1213 knees from the Osteoarthritis Initiative. Through extensive experimentation with various feature combinations, we identified KL baseline and weight as the most essential features, while gender surprisingly proved to be one of the least influential feature. Our best classification model generated a weighted F1 score of 79.0% and a balanced accuracy of 65.9%. The area under the receiver operating characteristic curve was 83.0% for healthy (KL01) versus moderate (KL2) or severe (KL34) KOA patients and 86.6% for moderate (KL2) versus severe (KL34) KOA patients. We found a statistically significant difference in performance between our two-stage classification model and the traditional single-stage classification model. These findings demonstrate the encouraging results of our two-stage classification model for multiclass KOA severity classification, suggesting its potential application in clinical settings in future.</p>","PeriodicalId":7986,"journal":{"name":"Annals of Biomedical Engineering","volume":null,"pages":null},"PeriodicalIF":3.0000,"publicationDate":"2024-07-09","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"Annals of Biomedical Engineering","FirstCategoryId":"5","ListUrlMain":"https://doi.org/10.1007/s10439-024-03578-x","RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"Q3","JCRName":"ENGINEERING, BIOMEDICAL","Score":null,"Total":0}
引用次数: 0

Abstract

Currently, there are no methods or tools available in clinical practice for classifying future knee osteoarthritis (KOA). In this study, we aimed to fill this gap by classifying future KOA into three severity grades: KL01 (healthy), KL2 (moderate), and KL34 (severe) based on the Kellgren-Lawrance scale. Due to the complex nature of multiclass classification, we used a two-stage method, which separates the classification task into two binary classifications (KL01 vs. KL234 in the first stage and KL2 vs. KL34 in the second stage). Our machine learning (ML) model used two Balanced Random Forest algorithms and was trained with gender, age, height, weight, and quantitative knee morphology obtained from magnetic resonance imaging. Our training dataset comprised longitudinal 8-year follow-up data of 1213 knees from the Osteoarthritis Initiative. Through extensive experimentation with various feature combinations, we identified KL baseline and weight as the most essential features, while gender surprisingly proved to be one of the least influential feature. Our best classification model generated a weighted F1 score of 79.0% and a balanced accuracy of 65.9%. The area under the receiver operating characteristic curve was 83.0% for healthy (KL01) versus moderate (KL2) or severe (KL34) KOA patients and 86.6% for moderate (KL2) versus severe (KL34) KOA patients. We found a statistically significant difference in performance between our two-stage classification model and the traditional single-stage classification model. These findings demonstrate the encouraging results of our two-stage classification model for multiclass KOA severity classification, suggesting its potential application in clinical settings in future.

Abstract Image

查看原文
分享 分享
微信好友 朋友圈 QQ好友 复制链接
本刊更多论文
使用磁共振成像对 8 年后未来膝关节骨性关节炎严重程度进行两阶段分类:骨关节炎倡议 "的数据。
目前,临床实践中还没有可用来对未来膝关节骨性关节炎(KOA)进行分级的方法或工具。在这项研究中,我们将未来 KOA 的严重程度分为三个等级,旨在填补这一空白:KL01(健康)、KL2(中度)和 KL34(重度)。由于多类分类的复杂性,我们采用了两阶段方法,将分类任务分为两个二元分类(第一阶段为 KL01 与 KL234,第二阶段为 KL2 与 KL34)。我们的机器学习(ML)模型使用了两种平衡随机森林算法,并通过性别、年龄、身高、体重和磁共振成像获得的膝关节形态定量数据进行训练。我们的训练数据集包括骨关节炎倡议(Osteoarthritis Initiative)1213 个膝关节的 8 年纵向随访数据。通过对各种特征组合的广泛试验,我们发现 KL 基线和体重是最基本的特征,而性别则出人意料地被证明是影响最小的特征之一。我们的最佳分类模型的加权 F1 得分为 79.0%,平衡准确率为 65.9%。健康(KL01)与中度(KL2)或重度(KL34)KOA 患者的接收者操作特征曲线下面积为 83.0%,中度(KL2)与重度(KL34)KOA 患者的接收者操作特征曲线下面积为 86.6%。我们发现,我们的两阶段分类模型与传统的单阶段分类模型在性能上有显著的统计学差异。这些研究结果表明,我们的两阶段分类模型在多级 KOA 严重程度分类中取得了令人鼓舞的结果,这表明它将来有可能应用于临床。
本文章由计算机程序翻译,如有差异,请以英文原文为准。
求助全文
约1分钟内获得全文 去求助
来源期刊
Annals of Biomedical Engineering
Annals of Biomedical Engineering 工程技术-工程:生物医学
CiteScore
7.50
自引率
15.80%
发文量
212
审稿时长
3 months
期刊介绍: Annals of Biomedical Engineering is an official journal of the Biomedical Engineering Society, publishing original articles in the major fields of bioengineering and biomedical engineering. The Annals is an interdisciplinary and international journal with the aim to highlight integrated approaches to the solutions of biological and biomedical problems.
期刊最新文献
The Potential of Sound Analysis to Reveal Hemodynamic Conditions of Arteriovenous Fistulae for Hemodialysis. Active Machine Learning for Pre-procedural Prediction of Time-Varying Boundary Condition After Fontan Procedure Using Generative Adversarial Networks. Simulating Pelvis Kinematics from Belt and Seat Loading in Frontal Car Crash Scenarios: Important Boundary Conditions that Influence the Outcome. Lumbar Spine Orientation Affects Compressive Fracture Outcome. An In vivo Pilot Study to Estimate the Swelling of the Aneurysm Wall Rabbit Model Generated with Pulsed Fluid Against the Aneurysm Wall.
×
引用
GB/T 7714-2015
复制
MLA
复制
APA
复制
导出至
BibTeX EndNote RefMan NoteFirst NoteExpress
×
×
提示
您的信息不完整,为了账户安全,请先补充。
现在去补充
×
提示
您因"违规操作"
具体请查看互助需知
我知道了
×
提示
现在去查看 取消
×
提示
确定
0
微信
客服QQ
Book学术公众号 扫码关注我们
反馈
×
意见反馈
请填写您的意见或建议
请填写您的手机或邮箱
已复制链接
已复制链接
快去分享给好友吧!
我知道了
×
扫码分享
扫码分享
Book学术官方微信
Book学术文献互助
Book学术文献互助群
群 号:481959085
Book学术
文献互助 智能选刊 最新文献 互助须知 联系我们:info@booksci.cn
Book学术提供免费学术资源搜索服务,方便国内外学者检索中英文文献。致力于提供最便捷和优质的服务体验。
Copyright © 2023 Book学术 All rights reserved.
ghs 京公网安备 11010802042870号 京ICP备2023020795号-1