Next Article in Journal
Biochemical Characterization of a Novel α/β-Hydrolase/FSH from the White Shrimp Litopenaeus vannamei
Previous Article in Journal
Characterization of the Ohmyungsamycin Biosynthetic Pathway and Generation of Derivatives with Improved Antituberculosis Activity
Open AccessArticle

Illuminating Clues of Cancer Buried in Prostate MR Image: Deep Learning and Expert Approaches

1
Pathology Informatics Team, RIKEN Center for Advanced Intelligence Project, Tokyo 103-0027, Japan
2
Department of Urology, Nippon Medical School Hospital, Tokyo 113-8603, Japan
3
Department of Radiology, Nippon Medical School Hospital, Tokyo 113-8603, Japan
4
Statistical Genetics Team, RIKEN Center for Advanced Intelligence Project, Tokyo 103-0027, Japan
5
Tohoku Medical Megabank Organization, Tohoku University, Miyagi 980-8575, Japan
6
Department of Pathology, Kitasato University Kitasato Institute Hospital, Tokyo 108-8642, Japan
7
Department of Analytic Human Pathology, Nippon Medical School, Tokyo 113-8602, Japan
8
Department of Surgical Pathology, Aichi Medical University Hospital, Aichi 480-1195, Japan
*
Author to whom correspondence should be addressed.
Biomolecules 2019, 9(11), 673; https://doi.org/10.3390/biom9110673
Received: 25 September 2019 / Revised: 24 October 2019 / Accepted: 28 October 2019 / Published: 30 October 2019
(This article belongs to the Special Issue Application of Artificial Intelligence for Medical Research)
Deep learning algorithms have achieved great success in cancer image classification. However, it is imperative to understand the differences between the deep learning and human approaches. Using an explainable model, we aimed to compare the deep learning-focused regions of magnetic resonance (MR) images with cancerous locations identified by radiologists and pathologists. First, 307 prostate MR images were classified using a well-established deep neural network without locational information of cancers. Subsequently, we assessed whether the deep learning-focused regions overlapped the radiologist-identified targets. Furthermore, pathologists provided histopathological diagnoses on 896 pathological images, and we compared the deep learning-focused regions with the genuine cancer locations through 3D reconstruction of pathological images. The area under the curve (AUC) for MR images classification was sufficiently high (AUC = 0.90, 95% confidence interval 0.87–0.94). Deep learning-focused regions overlapped radiologist-identified targets by 70.5% and pathologist-identified cancer locations by 72.1%. Lymphocyte aggregation and dilated prostatic ducts were observed in non-cancerous regions focused by deep learning. Deep learning algorithms can achieve highly accurate image classification without necessarily identifying radiological targets or cancer locations. Deep learning may find clues that can help a clinical diagnosis even if the cancer is not visible. View Full-Text
Keywords: deep learning; black box; prostate cancer; MRI; pathology deep learning; black box; prostate cancer; MRI; pathology
Show Figures

Figure 1

MDPI and ACS Style

Akatsuka, J.; Yamamoto, Y.; Sekine, T.; Numata, Y.; Morikawa, H.; Tsutsumi, K.; Yanagi, M.; Endo, Y.; Takeda, H.; Hayashi, T.; Ueki, M.; Tamiya, G.; Maeda, I.; Fukumoto, M.; Shimizu, A.; Tsuzuki, T.; Kimura, G.; Kondo, Y. Illuminating Clues of Cancer Buried in Prostate MR Image: Deep Learning and Expert Approaches. Biomolecules 2019, 9, 673.

Show more citation formats Show less citations formats
Note that from the first issue of 2016, MDPI journals use article numbers instead of page numbers. See further details here.

Article Access Map by Country/Region

1
Back to TopTop