Evaluating Cellularity Estimation Methods: Comparing AI Counting with Pathologists’ Visual Estimates
Diagnostics · 28.05.2024 · 14 (11) · 1115
DOI: 10.3390/diagnostics14111115 ↗Kurz gefasst
Fachartikel in der Zeitschrift Diagnostics, online erschienen am 28.05.2024. Verfasst von Tomoharu Kiyuna und 17 weiteren Autor:innen. Beteiligte Einrichtung: Healthcare Life Science Division, NEC Corporation, Tokyo 108-8556, Japan. Die Arbeit wurde bisher 8-mal zitiert (Crossref) und nennt 32 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0).
Abstract (Originalsprache)
The development of next-generation sequencing (NGS) has enabled the discovery of cancer-specific driver gene alternations, making precision medicine possible. However, accurate genetic testing requires a sufficient amount of tumor cells in the specimen. The evaluation of tumor content ratio (TCR) from hematoxylin and eosin (H&E)-stained images has been found to vary between pathologists, making it an important challenge to obtain an accurate TCR. In this study, three pathologists exhaustively labeled all cells in 41 regions from 41 lung cancer cases as either tumor, non-tumor or indistinguishable, thus establishing a “gold standard” TCR. We then compared the accuracy of the TCR estimated by 13 pathologists based on visual assessment and the TCR calculated by an AI model that we have developed. It is a compact and fast model that follows a fully convolutional neural network architecture and produces cell detection maps which can be efficiently post-processed to obtain tumor and non-tumor cell counts from which TCR is calculated. Its raw cell detection accuracy is 92% while its classification accuracy is 84%. The results show that the error between the gold standard TCR and the AI calculation was significantly smaller than that between the gold standard TCR and the pathologist’s visual assessment (p
Schlagwörter
Stichwörter der Autor:innen
Autor:innen und Institutionen
- Healthcare Life Science Division, NEC Corporation, Tokyo 108-8556, Japan.
- Department of Machine Learning, NEC Laboratories America, Princeton, NJ 08540, USA.
- Center for Development of Advanced Diagnostics (C-DAD), Hokkaido University Hospital, Sapporo 060-8648, Japan.
- Department of Pathology, Kanagawa Cancer Center, Yokohama 241-8515, Japan.
- Department of Pathology, Kansai Medical University, Osaka 573-1010, Japan.
- Department of Pathology, Saitama Cancer Center, Saitama 362-0806, Japan.
ORCID-Profile: Tomoharu Kiyuna, Kanako C Hatanaka, Tomoyuki Yokose, Noriko Motoi, Keishi Makita, Ai Shimizu, Emi Takakuwa, Yasunari Takakuwa, Takahiro Tsuji, Mitsuhiro Tsujiwaki, Mitsuru Yanai, Sayaka Yuzawa, Maki Ogura, Yutaka Hatanaka
Zitieren
APA
Kiyuna, T., Cosatto, E., Hatanaka, K. C., Yokose, T., Tsuta, K., Motoi, N., Makita, K., Shimizu, A., Shinohara, T., Suzuki, A., Takakuwa, E., Takakuwa, Y., Tsuji, T., Tsujiwaki, M., Yanai, M., Yuzawa, S., Ogura, M., & Hatanaka, Y. (2024). Evaluating Cellularity Estimation Methods: Comparing AI Counting with Pathologists’ Visual Estimates. Diagnostics, 14(11), 1115. https://doi.org/10.3390/diagnostics14111115
Vancouver
Kiyuna T, Cosatto E, Hatanaka KC, Yokose T, Tsuta K, Motoi N, et al. Evaluating Cellularity Estimation Methods: Comparing AI Counting with Pathologists’ Visual Estimates. Diagnostics. 2024;14(11):1115. doi:10.3390/diagnostics14111115
BibTeX
@article{Kiyuna2024,
author = {Kiyuna, Tomoharu and Cosatto, Eric and Hatanaka, Kanako C and Yokose, Tomoyuki and Tsuta, Koji and Motoi, Noriko and Makita, Keishi and Shimizu, Ai and Shinohara, Toshiya and Suzuki, Akira and Takakuwa, Emi and Takakuwa, Yasunari and Tsuji, Takahiro and Tsujiwaki, Mitsuhiro and Yanai, Mitsuru and Yuzawa, Sayaka and Ogura, Maki and Hatanaka, Yutaka},
title = {Evaluating Cellularity Estimation Methods: Comparing AI Counting with Pathologists’ Visual Estimates},
journal = {Diagnostics},
year = {2024},
volume = {14},
number = {11},
pages = {1115},
doi = {10.3390/diagnostics14111115}
}
EUVIMED