A recent article in Forbes highlights a study on diagnostic performance published in The Lancet Digital Health. The study represents the first review to compare the diagnostic accuracy of deep learning models against health-care professionals locating diseases by using medical imaging.
Deep learning, a type of artificial intelligence (AI), shows the potential to improve diagnostic speed and accuracy in medical imaging.
In fact, after analyzing fourteen studies conducted between 2012 and 2019, the researchers found comparable accuracy in both diagnostic performances.
Bayer Pharmaceutical’s Involvement
In addition to diagnostic advances, AI is moving forward in treating diseases. The information being fed to AI is garnered from many factors. Some of the factors include disease causes, medical images, symptom data, and doctors’ reports.
Bayer has joined tech companies working to create software that will diagnose rare, complex conditions as well as the drugs that will treat these conditions.
In addition, Bayer has partnered with various researchers and hospitals in an effort to determine what is required for machine learning to be able to diagnose a patient’s medical condition.
The term “machine learning” came about at the same time AI appeared on the scene. AI refers to a machine performing tasks. Machine learning, on the other hand, refers to a computer application that processes and learns from data. In other words, it can learn certain behaviors. An example would be autonomous cars.
Angela Moeller, head of Bayer’s AI projects, explained that the industry is not attempting to replace doctors or make final decisions with respect to patient care. They believe that patients should have control over decisions that are made in their treatment. AI should be used to support those decisions and treatment recommendations should be based on accurate diagnosis.
Moeller admitted that it is difficult to get this technology to the patient so it may take at least two years before it becomes mainstream in medical practice.
At this moment there are approximately 148 startups that are using AI in drug discovery. However, there are no drug treatments developed by AI on the market as yet.
Deep Genomics Biotech Company
It has been five years since Deep Genomics, a Canadian company, began experimenting with drug development and machine learning in their quest to find a cure for Wilson disease.
Wilson disease, a rare and sometimes fatal disorder, causes copper buildup in the brain, liver and other organs. The accumulation of copper is due to a mutation that leads to the loss of function of the copper-binding protein ATP7B.
One of the researchers said that the disease has been studied for over twenty years without success. The AI system was able to provide an analysis in three hours.
He went on to explain that the Wilson mutation is one of a new class of mutations. The protein ATP7B is absent in Wilson disease because the mutation causes a disruption in the genome and prevents its production.
Deep Genomic’s new therapy is currently being tested on the first study volunteer. All those involved are hopeful for its success in treating Wilson disease, which affects about one person in every thirty thousand.
The healthcare industry, AI, and machine learning algorithms (calculations) are continuously advancing new treatments. Judging from all the AI activity, a number of new therapies should be forthcoming.
What are your thoughts about AI and its potential in the medical field?