CDIML pixel news
Just last month, a former director of a leading AI medical image analysis venture, El Pixel Co., Ltd., was arrested on business allegations of embezzlement. The amount of damage is said to be almost the same as the 3 billion yen raised by the company in 2018, which is quite shocking.
Of course, I was surprised by the news of the arrest, but it is shocking that it is the time when Japanese medical venture companies can raise billions of yen with one financing, and the company is 2019 It was surprising again that the AI technology development was solidly successful despite the fact that the medical device analysis software “EIRL aneurysm” received medical device approval in 2014, and the funds have been embezzled this much. There will be. After all, I think that I have a solid ability.
Current location of medical image analysis by AI
By the way, I would like to write about the medical image by AI, which L-Pixel is working on. It's a world of rapid progress, so I'm worried that the content is out of date as I'm writing...
Over the last few years, I have also been involved in some AI-themed stories. AI is also a technology of interest, and the combination of medical images and AI technology is considered to be compatible, so research and development is being actively conducted in Japan. Looking at the image, it seems that it is possible to make a fairly accurate judgment at the level that makes a specific judgment such as "whether there is a suspicion of cancer". To the extreme, some people say that whether or not they can make a decision is not a competitive axis. *
Then, will AI become the world where diagnosis is done automatically? The story is not that simple. At the moment, at the moment, the Ministry of Health, Labor and Welfare said, “AI is just a support tool that improves the efficiency of diagnostic hypothesis formation support and treatment strategy planning support in the medical treatment process, and presents information. (Healthcare field AI development acceleration consortium January 2019)”. In other words, even if the judgment is made using AI, the doctor must make a diagnosis by looking at the image with the doctor. Japanese doctors are excellent and the diagnosis speed is already too fast, so if the AI's judgment function remains “support”, it is a little question how useful it is in clinical practice. In addition, whether a new insurance score for medical equipment will be added to the function of image judgment, which is the same as what a doctor does, is also a subtle issue.
Future of AI x medical image analysis
What about AI technology development? Personally, I think there are two directions for development. One is to "specialize in the operation support of doctors". It is a direction to focus on functions for breaking down the work of an actual doctor and reducing the work load or increasing the amount that the doctor can handle. The other is to create functions that only AI can do. For example, we are aiming for "better medical care" by combining inputs other than images to derive "prognosis prediction" and "optimal treatment". I think both approaches are attractive, and I don't think that the competition is decided only by the amount of learning data, and I think that Japan's technological capabilities can be put to good use. What do you guys think?
AI technology remains one of the hot topics in life science consulting. In addition, I will write the continuation for the next and subsequent occasions. Updates are also important, so if you would like to exchange information, please contact us.
*This column does not discuss based on the definition of individual accuracy, sensitivity, specificity, accuracy, etc.
Responsibility: Ai Ito
Ai Ito (CDI Medical Consultant Co., Ltd.)
Completed a master's course at the Graduate School of Pharmaceutical Sciences, Osaka University (pharmacist). Completed a master's degree at Kyoto University Graduate School of Medicine.
After working as a trading company, an independent venture capital firm, a healthcare / bio venture company, a management consulting firm, etc. Conducts consulting including strategy support and execution support, including mid-term business strategy, new business strategy, overseas development, open innovation strategy, etc., mainly in the life sciences and healthcare field.