Remember meForgot password?
    Log in with Twitter

article imageArtificial intelligence can improve on human doctors

By Tim Sandle     Jul 17, 2017 in Science
The headline of this article may require re-reading: can artificial intelligence really improve upon human medics? If machines can learn to take medical decisions, is this a good thing? A number of health tech investors think so.
One reason why machines are needed in medicine is a result of clinicians being faced with an avalanche of digital data every day. The plethora of information is making it increasingly hard for medical staff to male sense of everything. Commenting in this, Professor Robert Darnell, from The Rockefeller University, told medical site QMed: " “Time is critical when you are a sick patient, and time is critical when you are a physician scientist."
Britain s Prince William (R)  president of the Royal Marsden NHS (National Health Service) Foundatio...
Britain's Prince William (R), president of the Royal Marsden NHS (National Health Service) Foundation Trust, watches as head surgeon Pardeep Kumar (L) performs surgery for the removal of a bladder tumour on a male patient during a visit to the Royal Marsden hospital in London November 7, 2013
With permission by Reuters / Lefteris Pitarakis
This means, the researcher emphasizes, we need to see artificial intelligence in medicine as a partner and not a threat to medical competence. With this he adds: “We need the machines to scale what we do, and we need some component of human thought to at least be able to teach the machines. I do think there’s a human-machine collaboration that is upon us right now, and probably will be for a while.”
READ MORE: Medical technology will be transformed by robotics
So what does this mean? We take stock of three examples of artificial intelligence in medicine.
The first example comes from Professor Robert Darnell himself. Working with the New York Genome Center and IBM, the researcher demonstrated the power of IBM Watson for Genomics to analyze complex genomic data from sequencing of whole genomes. This study helped to pinpoint personalized therapies for cancer patients. The major case study was on the aggressive brain cancer glioblastoma. With this Watson was able to report of potentially clinically actionable insights within 10 minutes.
The second example relates to radiotherapy. With this area of medicine, scientists from University College London Hospitals NHS Foundation Trust are examining ways that machine learning can reduce the amount of time required to plan radiotherapy treatments for head and neck cancers. Machine learning has reduced this from a typical four hours to under one hour. This process has used Google's DeepMind to assist with the analytics. A second artificial intelligence machine capable of similar actions, is SOPHiA. SOPHiA can help doctors diagnose, treat, and monitor cancer.
READ MORE: Surgical Robotics: The Next 25 Years
The third and final example is Elon Musk's venture called Neuralink. This aims to connect artificial intelligence with the human brain, via an implantable chip. The aim is to assist people merge with computer software for treatments like deep brain stimulation, which can treat the symptoms of epilepsy and perhaps other neurodegenerative diseases like Parkinson's disease.
Machine learning and AI are making inroads into medicine. Time will tell whether they can contribute in the kind of way that business interests are currently suggesting.
More about Artifical intelligence, Medicine, machine learning
More news from