Our paper has been published again by ”Scientific Reports”!
Deep Learning has been used in various medical fields recently especially in radiology and pathology where medical image recognition has been used frequently.
We have announced in January 2020 the establishment of an AI model to detect epithelial tumors in the stomach and colon, and this has been published by Scientific Reports.
(Scientific Reports 2020, volume 10, Article (number: 1504)
We are so pleased to announce that on June 9th, 2020, our paper co-authored by the doctors from the National Hospital Organization Kyushu Medical Center and members of Medmain Inc. was again published in Scientific Reports!
The topic of this paper is “Weakly-supervised learning for lung carcinoma classification using deep learning “.
Lung cancer is one of the leading causes of cancer death in many countries around the world.
In diagnosing lung cancer, clinicians take tissue out from the patient’s lung, then pathologists identify the cancer under a microscope. This process is called “histopathological examination” and the diagnosis based on this examination will be “definitive diagnosis”. Thus, this is very important in optimal treatment planning.
We have established an AI model that detects lung cancer on pathological tissue with extremely high accuracy, and have successfully verified its accuracy on world’s largest numbers of test sets.
Using weakly-supervised learning, we let lesion imaging information extracted from over 3,000 lung histopathological specimens deep-learned.
With test cases from four domestic hospitals and international public databases, we validated the detection of lung cancer tissue and confirmed the high accuracy of the established AI model.
(ROC-AUC: 0.975, 0.974, 0.988, 0.981)
Medmain will work even harder to respond to the growing expectations for AI-based pathological image analysis.
“Weakly-supervised learning for lung carcinoma classification using deep learning” Link https://www.nature.com/articles/s41598-020-66333-x
Fahdi Kanavati, Gouji Toyokawa, Seiya Momosaki, Michael Rambeau, Yuka Kozuma, Fumihiro Shoji, Koji Yamazaki, Sadanori Takeo, Osamu Iizuka, and Masayuki Tsuneki
PubMed Link https://pubmed.ncbi.nlm.nih.gov/32518413
Nature Scientific Reports https://www.nature.com/srep/