Inconsistency between Human Observation and Deep Learning Models: Assessing Validity of Postmortem Computed Tomography Diagnosis of Drowning
Journal of Imaging Informatics in Medicine · 2024
DOI: 10.1007/s10278-024-00974-6 ↗Kurz gefasst
Fachartikel in der Zeitschrift Journal of Imaging Informatics in Medicine, 2024. Verfasst von Yuwen Zeng und 8 weiteren Autor:innen. Beteiligte Einrichtung: Department of Radiological Imaging and Informatics, Tohoku University Graduate School of Medicine, Sendai, Japan. yuwen@tohoku.ac.jp. Hauptthemen laut MEDLINE: Deep Learning, Drowning, Tomography, X-Ray Computed, Autopsy. Die Arbeit wurde bisher 5-mal zitiert (Crossref) und nennt 35 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: Japan Society for the Promotion of Science.
Abstract (Originalsprache)
Drowning diagnosis is a complicated process in the autopsy, even with the assistance of autopsy imaging and the on-site information from where the body was found. Previous studies have developed well-performed deep learning (DL) models for drowning diagnosis. However, the validity of the DL models was not assessed, raising doubts about whether the learned features accurately represented the medical findings observed by human experts. In this paper, we assessed the medical validity of DL models that had achieved high classification performance for drowning diagnosis. This retrospective study included autopsy cases aged 8–91 years who underwent postmortem computed tomography between 2012 and 2021 (153 drowning and 160 non-drowning cases). We first trained three deep learning models from a previous work and generated saliency maps that highlight important features in the input. To assess the validity of models, pixel-level annotations were created by four radiological technologists and further quantitatively compared with the saliency maps. All the three models demonstrated high classification performance with areas under the receiver operating characteristic curves of 0.94, 0.97, and 0.98, respectively. On the other hand, the assessment results revealed unexpected inconsistency between annotations and models’ saliency maps. In fact, each model had, respectively, around 30%, 40%, and 80% of irrelevant areas in the saliency maps, suggesting the predictions of the DL models might be unreliable. The result alerts us in the careful assessment of DL tools, even those with high classification performance.
Schlagwörter
MeSH-Schlagwörter (MEDLINE)
Stichwörter der Autor:innen
Autor:innen und Institutionen
- Department of Radiological Imaging and Informatics, Tohoku University Graduate School of Medicine, Sendai, Japan. yuwen@tohoku.ac.jp.
- National Institute of Technology, Sendai College, Sendai, Japan.
- Department of Intelligent Biomedical System Engineering, Graduate School of Biomedical Engineering, Tohoku University, Sendai, Japan.
- Department of Radiological Imaging and Informatics, Tohoku University Graduate School of Medicine, Sendai, Japan.
- Faculty of Science, University of South Bohemia in Ceske Budejovice, Ceske Budejovice, Czech Republic.
- Mechanical Engineering, Czech Technical University in Prague, Prague, Czech Republic.
ORCID-Profile: Yuwen Zeng
Förderung
- Japan Society for the Promotion of Science
Zitieren
APA
Zeng, Y., Zhang, X., Wang, J., Usui, A., Ichiji, K., Bukovsky, I., Chou, S., Funayama, M., & Homma, N. (2024). Inconsistency between Human Observation and Deep Learning Models: Assessing Validity of Postmortem Computed Tomography Diagnosis of Drowning. Journal of Imaging Informatics in Medicine. https://doi.org/10.1007/s10278-024-00974-6
Vancouver
Zeng Y, Zhang X, Wang J, Usui A, Ichiji K, Bukovsky I, et al. Inconsistency between Human Observation and Deep Learning Models: Assessing Validity of Postmortem Computed Tomography Diagnosis of Drowning. Journal of Imaging Informatics in Medicine. 2024. doi:10.1007/s10278-024-00974-6
BibTeX
@article{Zeng2024,
author = {Zeng, Yuwen and Zhang, Xiaoyong and Wang, Jiaoyang and Usui, Akihito and Ichiji, Kei and Bukovsky, Ivo and Chou, Shuoyan and Funayama, Masato and Homma, Noriyasu},
title = {Inconsistency between Human Observation and Deep Learning Models: Assessing Validity of Postmortem Computed Tomography Diagnosis of Drowning},
journal = {Journal of Imaging Informatics in Medicine},
year = {2024},
doi = {10.1007/s10278-024-00974-6}
}
EUVIMED