EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

Deep learning-based diagnosis of feline hypertrophic cardiomyopathy

Jinhyung Rho iD, Sung-Min Shin, Kyoungsun Jhang, Gwanghee Lee, Keun-Ho Song, Hyunguk Shin, Kiwon Na, Hyo-Jung Kwon, Hwa-Young Son

PLOS ONE · 02.02.2023 · 18 (2) · e0280438

DOI: 10.1371/journal.pone.0280438 ↗

Kurz gefasst

Fachartikel in der Zeitschrift PLOS ONE, online erschienen am 02.02.2023. Verfasst von Jinhyung Rho und 8 weiteren Autor:innen. Beteiligte Einrichtung: Jeonbuk Pathology Research Group, Korea Institute of Toxicology, Jeonbuk, Republic of Korea. Hauptthemen laut MEDLINE: Cardiomyopathy, Hypertrophic, Deep Learning, Cardiomyopathies, Heart Diseases. Die Arbeit wurde bisher 16-mal zitiert (Crossref) und nennt 26 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: Chungnam National University.

AbstractSchlagwörterBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Feline hypertrophic cardiomyopathy (HCM) is a common heart disease affecting 10–15% of all cats. Cats with HCM exhibit breathing difficulties, lethargy, and heart murmur; furthermore, feline HCM can also result in sudden death. Among various methods and indices, radiography and ultrasound are the gold standards in the diagnosis of feline HCM. However, only 75% accuracy has been achieved using radiography alone. Therefore, we trained five residual architectures (ResNet50V2, ResNet152, InceptionResNetV2, MobileNetV2, and Xception) using 231 ventrodorsal radiographic images of cats (143 HCM and 88 normal) and investigated the optimal architecture for diagnosing feline HCM through radiography. To ensure the generalizability of the data, the x-ray images were obtained from 5 independent institutions. In addition, 42 images were used in the test. The test data were divided into two; 22 radiographic images were used in prediction analysis and 20 radiographic images of cats were used in the evaluation of the peeking phenomenon and the voting strategy. As a result, all models showed > 90% accuracy; Resnet50V2: 95.45%; Resnet152: 95.45; InceptionResNetV2: 95.45%; MobileNetV2: 95.45% and Xception: 95.45. In addition, two voting strategies were applied to the five CNN models; softmax and majority voting. As a result, the softmax voting strategy achieved 95% accuracy in combined test data. Our findings demonstrate that an automated deep-learning system using a residual architecture can assist veterinary radiologists in screening HCM.

Schlagwörter

MeSH-Schlagwörter (MEDLINE)

Cardiomyopathy, HypertrophicDeep LearningCardiomyopathiesHeart DiseasesCat DiseasesCatsAnimalsUltrasonography

Autor:innen und Institutionen

  • Jeonbuk Pathology Research Group, Korea Institute of Toxicology, Jeonbuk, Republic of Korea.
  • Center for Companion Animal New Drug Development, Korea Institute of Toxicology, Jeonbuk, Republic of Korea.
  • Do Animal Hospital, Incheon, Republic of Korea.
  • Department of Computer Engineering, Chungnam National University, Daejeon, Republic of Korea.
  • College of Veterinary Medicine, Chungnam National University, Daejeon, Republic of Korea.
  • 24 Africa Animal Medical Center, Daejeon, Republic of Korea.

ORCID-Profile: Jinhyung Rho

Förderung

  • Chungnam National University

Zitieren

APA

Rho, J., Shin, S. M., Jhang, K., Lee, G., Song, K. H., Shin, H., Na, K., Kwon, H. J., & Son, H. Y. (2023). Deep learning-based diagnosis of feline hypertrophic cardiomyopathy. PLOS ONE, 18(2), e0280438. https://doi.org/10.1371/journal.pone.0280438

Vancouver

Rho J, Shin SM, Jhang K, Lee G, Song KH, Shin H, et al. Deep learning-based diagnosis of feline hypertrophic cardiomyopathy. PLOS ONE. 2023;18(2):e0280438. doi:10.1371/journal.pone.0280438

BibTeX
@article{Rho2023,
  author = {Rho, Jinhyung and Shin, Sung-Min and Jhang, Kyoungsun and Lee, Gwanghee and Song, Keun-Ho and Shin, Hyunguk and Na, Kiwon and Kwon, Hyo-Jung and Son, Hwa-Young},
  title = {Deep learning-based diagnosis of feline hypertrophic cardiomyopathy},
  journal = {PLOS ONE},
  year = {2023},
  volume = {18},
  number = {2},
  pages = {e0280438},
  doi = {10.1371/journal.pone.0280438}
}
Kostenlos starten →