Automated diagnosis of flatfoot using cascaded convolutional neural network for angle measurements in weight-bearing lateral radiographs

被引:4
|
作者
Ryu, Seung Min [1 ,2 ]
Shin, Keewon [1 ]
Shin, Soo Wung [3 ,4 ]
Lee, Sun Ho [5 ]
Seo, Su Min [6 ]
Cheon, Seung-Uk [6 ]
Ryu, Seung-Ah [6 ]
Kim, Min-Ju [7 ]
Kim, Hyunjung [1 ]
Doh, Chang Hyun [2 ]
Choi, Young Rak [2 ]
Kim, Namkug [3 ,8 ]
机构
[1] Univ Ulsan, Asan Med Inst Convergence Sci & Technol, Asan Med Ctr, Dept Biomed Engn,Coll Med, 26, Olymp Ro,43 Gil, Seoul 05506, South Korea
[2] Univ Ulsan, Asan Med Ctr, Dept Orthoped Surg, Coll Med, 88,Olymp Ro,43 Gil, Seoul 05505, South Korea
[3] Univ Ulsan, Res Inst Radiol, Asan Med Ctr, Dept Radiol,Coll Med, 88,Olymp Ro,43 Gil, Seoul 05506, South Korea
[4] Seoul Natl Univ, Dept Comp Sci & Engn, 1,Gwanak Ro, Seoul 08826, South Korea
[5] Chonnam Natl Univ Hosp, Dept Orthoped Surg, 42,Jebong Ro, Gwangju Gwangyeogsi 61469, South Korea
[6] Seoul Med Ctr, Dept Anesthesiol & Pain Med, 156,Sinnae Ro, Seoul 02053, South Korea
[7] Univ Ulsan, Asan Med Ctr, Dept Clin Epidemiol & Biostat, Coll Med, 88,Olymp Ro 43 Gil, Seoul 05505, South Korea
[8] Univ Ulsan, Asan Med Inst Convergence Sci & Technol, Asan Med Ctr, Dept Convergence Med,Coll Med, 26,Olymp Ro 43 Gil, Seoul 05506, South Korea
基金
新加坡国家研究基金会;
关键词
Deep Learning; Flatfoot; X-rays; Observer variation; Computer-assisted diagnosis; FOOT; CHILDREN; RELIABILITY; ANATOMY;
D O I
10.1007/s00330-023-09442-1
中图分类号
R8 [特种医学]; R445 [影像诊断学];
学科分类号
1002 ; 100207 ; 1009 ;
摘要
ObjectivesDiagnosis of flatfoot using a radiograph is subject to intra- and inter-observer variabilities. Here, we developed a cascade convolutional neural network (CNN)-based deep learning model (DLM) for an automated angle measurement for flatfoot diagnosis using landmark detection.MethodsWe used 1200 weight-bearing lateral foot radiographs from young adult Korean males for the model development. An experienced orthopedic surgeon identified 22 radiographic landmarks and measured three angles for flatfoot diagnosis that served as the ground truth (GT). Another orthopedic surgeon (OS) and a general physician (GP) independently identified the landmarks of the test dataset and measured the angles using the same method. External validation was performed using 100 and 17 radiographs acquired from a tertiary referral center and a public database, respectively.ResultsThe DLM showed smaller absolute average errors from the GT for the three angle measurements for flatfoot diagnosis compared with both human observers. Under the guidance of the DLM, the average errors of observers OS and GP decreased from 2.35 degrees +/- 3.01 degrees to 1.55 degrees +/- 2.09 degrees and from 1.99 degrees +/- 2.76 degrees to 1.56 degrees +/- 2.19 degrees, respectively (both p < 0.001). The total measurement time decreased from 195 to 135 min in observer OS and from 205 to 155 min in observer GP. The absolute average errors of the DLM in the external validation sets were similar or superior to those of human observers in the original test dataset.ConclusionsOur CNN model had significantly better accuracy and reliability than human observers in diagnosing flatfoot, and notably improved the accuracy and reliability of human observers.
引用
收藏
页码:4822 / 4832
页数:11
相关论文
共 50 条
  • [31] Rolling element bearing fault diagnosis using convolutional neural network and vibration image
    Hoang, Duy-Tang
    Kang, Hee-Jun
    COGNITIVE SYSTEMS RESEARCH, 2019, 53 : 42 - 50
  • [32] Intelligent Fault Diagnosis of Bearing Using Multiwavelet Perception Kernel Convolutional Neural Network
    Zhou, Yuanyuan
    Wang, Hang
    Liu, Yongbin
    Liu, Xianzeng
    Cao, Zheng
    IEEE SENSORS JOURNAL, 2024, 24 (08) : 12728 - 12739
  • [33] Cascaded Deep Learning Neural Network for Automated Liver Steatosis Diagnosis Using Ultrasound Images
    Rhyou, Se-Yeol
    Yoo, Jae-Chern
    SENSORS, 2021, 21 (16)
  • [34] Automatic diagnosis of retention pseudocyst in the maxillary sinus on panoramic radiographs using a convolutional neural network algorithm
    Ha, Eun-Gyu
    Jeon, Kug Jin
    Choi, Hanseung
    Lee, Chena
    Choi, Yoon Joo
    Han, Sang-Sun
    SCIENTIFIC REPORTS, 2023, 13 (01)
  • [35] Inter-rater reliability and validity of angle measurements using smartphone applications for weight-bearing ankle dorsiflexion range of motion measurements
    Awatani, Takenori
    Enoki, Taisuke
    Morikita, Ikuhiro
    PHYSICAL THERAPY IN SPORT, 2018, 34 : 113 - 120
  • [36] Deep Convolutional Neural Network Model for Automated Diagnosis of Schizophrenia Using EEG Signals
    Oh, Shu Lih
    Vicnesh, Jahmunah
    Ciaccio, Edward J.
    Yuvaraj, Rajamanickam
    Acharya, U. Rajendra
    APPLIED SCIENCES-BASEL, 2019, 9 (14):
  • [37] On the use of a cascaded convolutional neural network for three-dimensional flow measurements using astigmatic PTV
    Koenig, Joerg
    Chen, Minqian
    Roesing, Wiebke
    Boho, David
    Maeder, Patrick
    Cierpka, Christian
    MEASUREMENT SCIENCE AND TECHNOLOGY, 2020, 31 (07)
  • [38] Deep convolutional neural network for the automated detection and diagnosis of seizure using EEG signals
    Acharya, U. Rajendra
    Oh, Shu Lih
    Hagiwara, Yuki
    Tan, Jen Hong
    Adeli, Hojjat
    COMPUTERS IN BIOLOGY AND MEDICINE, 2018, 100 : 270 - 278
  • [39] Automatic diagnosis of retention pseudocyst in the maxillary sinus on panoramic radiographs using a convolutional neural network algorithm
    Eun-Gyu Ha
    Kug Jin Jeon
    Hanseung Choi
    Chena Lee
    Yoon Joo Choi
    Sang-Sun Han
    Scientific Reports, 13
  • [40] Automated Detection of Periodontal Bone Loss Using Deep Learning and Panoramic Radiographs: A Convolutional Neural Network Approach
    Ryu, Jihye
    Lee, Dong-Min
    Jung, Yun-Hoa
    Kwon, OhJin
    Park, SunYoung
    Hwang, JaeJoon
    Lee, Jae-Yeol
    APPLIED SCIENCES-BASEL, 2023, 13 (09):