EUVIMEDEUVIMED entdecken
Multizentrische Studie

Ovarian Cancer Detection in Ascites Cytology with Weakly Supervised Model on Nationwide Data Set

Jiwon Lee, Seonggyeong Choi, Seoyeon Shin, Mohammad Rizwan Alam, Jamshid Abdul-Ghafar, Kyung Jin Seo, Gisu Hwang, Daeky Jeong, Gyungyub Gong, Nam Hoon Cho, Chong Woo Yoo, Hyung Kyung Kim und 2 weitere

The American Journal of Pathology · 2025

DOI: 10.1016/j.ajpath.2025.04.004 ↗

Kurz gefasst

Multizentrische Studie in der Zeitschrift The American Journal of Pathology, 2025. Verfasst von Jiwon Lee und 13 weiteren Autor:innen. Beteiligte Einrichtung: College of Medicine, The Catholic University of Korea, Seoul, Republic of Korea. Hauptthemen laut MEDLINE: Ovarian Neoplasms, Neural Networks, Computer, Cytodiagnosis, Ascites. Die Arbeit wurde bisher 3-mal zitiert (Crossref) und nennt 38 Literaturangaben. Gefördert von: Ministry of Science and ICT, South Korea, The Catholic University of Korea Uijeongbu St Mary's Hospital, National Research Foundation of Korea.

AbstractSchlagwörterBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Conventional ascitic fluid cytology for detecting ovarian cancer is limited by its low sensitivity. To address this issue, this multicenter study developed patch image (PI)-based fully supervised convolutional neural network (CNN) models and clustering-constrained attention multiple-instance learning (CLAM) algorithms for detecting ovarian cancer using ascitic fluid cytology. Whole-slide images (WSIs), 356 benign and 147 cancer, were collected, from which 14,699 benign and 8025 cancer PIs were extracted. Additionally, 131 WSIs (44 benign and 87 cancer) were used for external validation. Six CNN algorithms were developed for cancer detection using PIs. Subsequently, two CLAM algorithms, single branch (CLAM-SB) and multiple branch (CLAM-MB), were developed. ResNet50 demonstrated the best performance, achieving an accuracy of 0.973. The performance when interpreting internal WSIs was an area under the curve (AUC) of 0.982. CLAM-SB outperformed CLAM-MB with an AUC of 0.944 for internal WSIs. Notably, in the external test, CLAM-SB exhibited superior performance with an AUC of 0.866 compared with ResNet50's AUC of 0.804. Analysis of the heatmap revealed that cases frequently misinterpreted by AI were easily interpreted by humans, and vice versa. Because AI and humans were found to function complementarily, implementing computer-aided diagnosis is expected to significantly enhance diagnostic accuracy and reproducibility. Furthermore, the WSI-based learning in CLAM, eliminating the need for patch-by-patch annotation, offers an advantage over the CNN model.

Schlagwörter

MeSH-Schlagwörter (MEDLINE)

Ovarian NeoplasmsNeural Networks, ComputerCytodiagnosisAscitesAscitic FluidHumansFemaleAlgorithms

Autor:innen und Institutionen

  • College of Medicine, The Catholic University of Korea, Seoul, Republic of Korea.
  • Department of Hospital Pathology, College of Medicine, The Catholic University of Korea, Seoul, Republic of Korea.
  • AI Team, DeepNoid Inc., Seoul, Republic of Korea.
  • Department of Pathology, Asan Medical Center, Seoul, Republic of Korea.
  • Department of Pathology, Yonsei University College of Medicine, Seoul, Republic of Korea.
  • Department of Pathology, National Cancer Center, Ilsan, Republic of Korea.

Förderung

  • Ministry of Science and ICT, South Korea
  • The Catholic University of Korea Uijeongbu St Mary's Hospital
  • National Research Foundation of Korea
  • Incorporated Research Institutions for Seismology
  • Catholic University of Korea

Zitieren

APA

Lee, J., Choi, S., Shin, S., Alam, M. R., Abdul-Ghafar, J., Seo, K. J., Hwang, G., Jeong, D., Gong, G., Cho, N. H., Yoo, C. W., Kim, H. K., Chong, Y., & Yim, K. (2025). Ovarian Cancer Detection in Ascites Cytology with Weakly Supervised Model on Nationwide Data Set. The American Journal of Pathology. https://doi.org/10.1016/j.ajpath.2025.04.004

Vancouver

Lee J, Choi S, Shin S, Alam MR, Abdul-Ghafar J, Seo KJ, et al. Ovarian Cancer Detection in Ascites Cytology with Weakly Supervised Model on Nationwide Data Set. The American Journal of Pathology. 2025. doi:10.1016/j.ajpath.2025.04.004

BibTeX
@article{Lee2025,
  author = {Lee, Jiwon and Choi, Seonggyeong and Shin, Seoyeon and Alam, Mohammad Rizwan and Abdul-Ghafar, Jamshid and Seo, Kyung Jin and Hwang, Gisu and Jeong, Daeky and Gong, Gyungyub and Cho, Nam Hoon and Yoo, Chong Woo and Kim, Hyung Kyung and Chong, Yosep and Yim, Kwangil},
  title = {Ovarian Cancer Detection in Ascites Cytology with Weakly Supervised Model on Nationwide Data Set},
  journal = {The American Journal of Pathology},
  year = {2025},
  doi = {10.1016/j.ajpath.2025.04.004}
}
Kostenlos starten →