Image analysis of prostate cancer patients with AI Prediction of recurrence with high accuracy 19:51 on December 18
A group such as RIKEN announced that AI = artificial intelligence was used to analyze images of prostate cancer patients and could predict the possibility of recurrence with high accuracy. AI also found cell features that seem to be associated with a previously unknown cancer recurrence.
This research was conducted by a group of team leaders Yoichiro Yamamoto of RIKEN and Associate Professor Tsuyoshi Kimura of Nippon Medical School, and published it in the science magazine Nature Communications.
The research group had AI read 100 prostate cancer images and trained with “deep learning” technology without giving answers such as where the cancer was.
As a result, AI naturally found out the characteristics of the cancer, and found that the density of cells around the cancer, which was not known before, may be related to the likelihood of recurrence. It is said.
Based on this, we analyzed approximately 15,000 prostate tissue images and compared it with information on whether the cancer had recurred, and it was possible to predict cancer recurrence with higher accuracy than before. about it.
By combining with the doctor's diagnosis, the recurrence can be predicted with high accuracy, Yamamoto said, “We were able to discover new cancer characteristics that humans could not reach using AI. Well, I want to proceed with research to see if it can be applied to rare diseases. "
AI to be introduced in the medical field
AI = artificial intelligence is being introduced to the medical field, and research and development are actively conducted especially in the field of diagnostic imaging to find out if there is a disease from a patient's image.
Diagnostic imaging is performed by doctors looking at patient images based on knowledge and experience. By using AI to assist doctors in diagnosis, mistakes such as oversight of illness can be prevented and more accurate. It is expected to be able to make a high diagnosis.
So far, medical institutions, universities, and companies in various regions have announced that they have developed an AI system that can quickly and accurately diagnose gastric cancer and liver cancer. In October, AI learns itself. A system that uses deep learning technology to detect diseases from brain images has received national approval as a medical device.
By utilizing AI, it is thought that the diagnosis can be performed efficiently at the medical site which is busy due to aging and the advancement of medical care, and the burden on medical staff can be reduced.
Furthermore, as new features of cells that lead to the recurrence of prostate cancer were discovered in this study, it was revealed by AI that things that were not known in previous studies may be revealed in the medical field. The movement to use is likely to strengthen.