 |
Vol.60 No.2 contents | Japanese/English |
 | Full Text of PDF (2457K) Article in Japanese |
- Invited Review Article -
Update Concerning Digital Pathology and Artificial Intelligence Application
Wataru Uegami1,2, Taro Sakamoto1, Kishio Kuroda1,2, Junya Fukuoka1,21Department of Pathology, Nagasaki University Graduate School of Biomedical Sciences, Japan, 2Department of Anatomic Pathology, Kameda Medical Center, Japan
The emergence of whole slide image technology has enabled us to make a pathological diagnosis via computer. The use of digital pathology is expanding into daily diagnoses, beyond merely functioning as support for remote institutes suffering from a shortage of pathologists. A large archival database of digital pathology images has now been developed, capable of functioning as the basis for image analyses using artificial intelligence (AI). At present, several AI applications have been reported, including ones for the detection of metastatic foci, counting tumor cells, and prediction of gene mutations. While rapid advancement in this field is expected to continue, most institutes are not prepared to introduce AI into their daily practice due to the inability to scan pathological specimens and obtain digital images. This limitation has slowed the spread of digitization. Furthermore, difficulty in developing training data and a lack of explainability have also hampered the development of AI in this field. Despite these challenges, AI is expected to play an integral role in making pathological diagnoses in the near future, requiring pathologists to invest in AI literacy to ensure effective management.
key words: Deep learning, Machine learning, Lung, Remote diagnosis, Consultation
JJLC 60 (2): 81-89, 2020
