基于深度学习的多足畸形分类地标检测模型:双中心研究。

IF 2.8 4区 医学 Q1 MEDICINE, GENERAL & INTERNAL
Su Ji Lee, Hangyul Yoon, Seongsu Bae, Inyoung Paik, Jong Hak Moon, Seongeun Park, Chan Woong Jang, Jung Hyun Park, Edward Choi, Eunho Yang, Ji Cheol Shin
{"title":"基于深度学习的多足畸形分类地标检测模型:双中心研究。","authors":"Su Ji Lee, Hangyul Yoon, Seongsu Bae, Inyoung Paik, Jong Hak Moon, Seongeun Park, Chan Woong Jang, Jung Hyun Park, Edward Choi, Eunho Yang, Ji Cheol Shin","doi":"10.3349/ymj.2024.0246","DOIUrl":null,"url":null,"abstract":"<p><strong>Purpose: </strong>To introduce heatmap-in-heatmap (HIH)-based model for automated diagnosis of foot deformities using weight-bearing foot radiographs, aiming to address the labor-intensive and variable nature of manual diagnosis.</p><p><strong>Materials and methods: </strong>From January 2004 to September 2022, a dual-center retrospective study was conducted. In the first center, 1561 anterior-posterior (AP) and 1536 lateral images from 806 patients were used for model training, while 374 AP and 373 lateral images from 196 patients were allocated to the validation set. For external validation at the second center, 527 AP and 529 lateral images from 270 patients were allocated. Five deformities were diagnosed using four and three angles between the predicted landmarks in the AP and lateral images, respectively. The results were compared with those of the baseline model (FlatNet).</p><p><strong>Results: </strong>The HIH model demonstrated robust performance in diagnosing multiple foot deformities. On the test set, it outperformed FlatNet with higher accuracy (FlatNet vs. HIH: 78.9% vs. 85.1%), sensitivity (78.9% vs. 84.1%), specificity (79.0% vs. 85.9%), positive predictive value (77.3% vs. 84.4%), and negative predictive value (80.5% vs. 85.7%). Additionally, HIH exhibited significantly lower absolute pixel and angle errors, lower normalized mean errors, higher successful detection rate, faster training and inference speeds, and fewer parameters.</p><p><strong>Conclusion: </strong>The HIH model showed robust performance in diagnosing multiple foot deformities with high efficacy in internal and external validation. Our approach is expected to be effective for various tasks using landmarks in medical imaging.</p>","PeriodicalId":23765,"journal":{"name":"Yonsei Medical Journal","volume":"66 8","pages":"491-501"},"PeriodicalIF":2.8000,"publicationDate":"2025-08-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC12303672/pdf/","citationCount":"0","resultStr":"{\"title\":\"Deep Learning-Based Landmark Detection Model for Multiple Foot Deformity Classification: A Dual-Center Study.\",\"authors\":\"Su Ji Lee, Hangyul Yoon, Seongsu Bae, Inyoung Paik, Jong Hak Moon, Seongeun Park, Chan Woong Jang, Jung Hyun Park, Edward Choi, Eunho Yang, Ji Cheol Shin\",\"doi\":\"10.3349/ymj.2024.0246\",\"DOIUrl\":null,\"url\":null,\"abstract\":\"<p><strong>Purpose: </strong>To introduce heatmap-in-heatmap (HIH)-based model for automated diagnosis of foot deformities using weight-bearing foot radiographs, aiming to address the labor-intensive and variable nature of manual diagnosis.</p><p><strong>Materials and methods: </strong>From January 2004 to September 2022, a dual-center retrospective study was conducted. In the first center, 1561 anterior-posterior (AP) and 1536 lateral images from 806 patients were used for model training, while 374 AP and 373 lateral images from 196 patients were allocated to the validation set. For external validation at the second center, 527 AP and 529 lateral images from 270 patients were allocated. Five deformities were diagnosed using four and three angles between the predicted landmarks in the AP and lateral images, respectively. The results were compared with those of the baseline model (FlatNet).</p><p><strong>Results: </strong>The HIH model demonstrated robust performance in diagnosing multiple foot deformities. On the test set, it outperformed FlatNet with higher accuracy (FlatNet vs. HIH: 78.9% vs. 85.1%), sensitivity (78.9% vs. 84.1%), specificity (79.0% vs. 85.9%), positive predictive value (77.3% vs. 84.4%), and negative predictive value (80.5% vs. 85.7%). Additionally, HIH exhibited significantly lower absolute pixel and angle errors, lower normalized mean errors, higher successful detection rate, faster training and inference speeds, and fewer parameters.</p><p><strong>Conclusion: </strong>The HIH model showed robust performance in diagnosing multiple foot deformities with high efficacy in internal and external validation. Our approach is expected to be effective for various tasks using landmarks in medical imaging.</p>\",\"PeriodicalId\":23765,\"journal\":{\"name\":\"Yonsei Medical Journal\",\"volume\":\"66 8\",\"pages\":\"491-501\"},\"PeriodicalIF\":2.8000,\"publicationDate\":\"2025-08-01\",\"publicationTypes\":\"Journal Article\",\"fieldsOfStudy\":null,\"isOpenAccess\":false,\"openAccessPdf\":\"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC12303672/pdf/\",\"citationCount\":\"0\",\"resultStr\":null,\"platform\":\"Semanticscholar\",\"paperid\":null,\"PeriodicalName\":\"Yonsei Medical Journal\",\"FirstCategoryId\":\"3\",\"ListUrlMain\":\"https://doi.org/10.3349/ymj.2024.0246\",\"RegionNum\":4,\"RegionCategory\":\"医学\",\"ArticlePicture\":[],\"TitleCN\":null,\"AbstractTextCN\":null,\"PMCID\":null,\"EPubDate\":\"\",\"PubModel\":\"\",\"JCR\":\"Q1\",\"JCRName\":\"MEDICINE, GENERAL & INTERNAL\",\"Score\":null,\"Total\":0}","platform":"Semanticscholar","paperid":null,"PeriodicalName":"Yonsei Medical Journal","FirstCategoryId":"3","ListUrlMain":"https://doi.org/10.3349/ymj.2024.0246","RegionNum":4,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"Q1","JCRName":"MEDICINE, GENERAL & INTERNAL","Score":null,"Total":0}
引用次数: 0

摘要

目的:介绍基于热图中的热图(HIH)模型的负重足部x线片足部畸形自动诊断,以解决人工诊断的劳动密集型和可变性。材料与方法:2004年1月至2022年9月进行双中心回顾性研究。在第一个中心,来自806名患者的1561张前后(AP)和1536张侧位图像被用于模型训练,而来自196名患者的374张前后(AP)和373张侧位图像被分配到验证集。为了在第二个中心进行外部验证,分配了来自270名患者的527张AP图像和529张侧位图像。五个畸形被诊断分别使用四个和三个角度的预测标志之间的AP和侧位图像。结果与基线模型(FlatNet)的结果进行比较。结果:HIH模型在诊断多发性足部畸形方面表现出稳健的性能。在测试集上,它以更高的准确率(FlatNet vs. HIH: 78.9% vs. 85.1%)、灵敏度(78.9% vs. 84.1%)、特异性(79.0% vs. 85.9%)、阳性预测值(77.3% vs. 84.4%)和阴性预测值(80.5% vs. 85.7%)优于FlatNet。此外,HIH具有更低的绝对像素和角度误差,更低的归一化平均误差,更高的成功检测率,更快的训练和推理速度以及更少的参数。结论:HIH模型对多发性足部畸形的诊断具有较强的稳健性,具有较高的内、外验证效果。我们的方法有望对医学成像中使用地标的各种任务有效。
本文章由计算机程序翻译,如有差异,请以英文原文为准。
Deep Learning-Based Landmark Detection Model for Multiple Foot Deformity Classification: A Dual-Center Study.

Purpose: To introduce heatmap-in-heatmap (HIH)-based model for automated diagnosis of foot deformities using weight-bearing foot radiographs, aiming to address the labor-intensive and variable nature of manual diagnosis.

Materials and methods: From January 2004 to September 2022, a dual-center retrospective study was conducted. In the first center, 1561 anterior-posterior (AP) and 1536 lateral images from 806 patients were used for model training, while 374 AP and 373 lateral images from 196 patients were allocated to the validation set. For external validation at the second center, 527 AP and 529 lateral images from 270 patients were allocated. Five deformities were diagnosed using four and three angles between the predicted landmarks in the AP and lateral images, respectively. The results were compared with those of the baseline model (FlatNet).

Results: The HIH model demonstrated robust performance in diagnosing multiple foot deformities. On the test set, it outperformed FlatNet with higher accuracy (FlatNet vs. HIH: 78.9% vs. 85.1%), sensitivity (78.9% vs. 84.1%), specificity (79.0% vs. 85.9%), positive predictive value (77.3% vs. 84.4%), and negative predictive value (80.5% vs. 85.7%). Additionally, HIH exhibited significantly lower absolute pixel and angle errors, lower normalized mean errors, higher successful detection rate, faster training and inference speeds, and fewer parameters.

Conclusion: The HIH model showed robust performance in diagnosing multiple foot deformities with high efficacy in internal and external validation. Our approach is expected to be effective for various tasks using landmarks in medical imaging.

求助全文
通过发布文献求助,成功后即可免费获取论文全文。 去求助
来源期刊
Yonsei Medical Journal
Yonsei Medical Journal 医学-医学:内科
CiteScore
4.50
自引率
0.00%
发文量
167
审稿时长
3 months
期刊介绍: The goal of the Yonsei Medical Journal (YMJ) is to publish high quality manuscripts dedicated to clinical or basic research. Any authors affiliated with an accredited biomedical institution may submit manuscripts of original articles, review articles, case reports, brief communications, and letters to the Editor.
×
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
确定
请完成安全验证×
copy
已复制链接
快去分享给好友吧!
我知道了
右上角分享
点击右上角分享
0
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 布克学术 All rights reserved.
京ICP备2023020795号-1
ghs 京公网安备 11010802042870号
Book学术文献互助
Book学术文献互助群
群 号:604180095
Book学术官方微信