Melbourne-based Peter MacCallum Cancer Centre has trialled the use of IBM Watson and found it helped reduce the time it takes for clinicians to match lung cancer patients to relevant clinical trials.
According to IBM, using past records of 102 lung cancer patients, IBM Watson for Clinical Trial Matching was able to match each patient to 10 potential trials, achieving a 92% accuracy when compared to manual clinician matching that would typically take "hours".
Peter MacCallum Cancer Centre oncologist and co-chief medical information officer Dr Dishan Herath said the six-month trial showed the potential of using artificial intelligence to eliminate "increasingly complex" trial criteria.
"One example is that one of the trials we included in the IBM project had multiple different arms -- protocols that patients could be allocated to -- and cohorts -- patients at different stages of their treatment. Each of these have their own eligibility criteria that Watson needed to analyse," he said.
See also: IBM Watson: A cheat sheet (TechRepublic)
The trial with Peter MacCallum Cancer Centre marked the first trial of IBM Watson for Clinical Trial Matching in Australia.
"The strengths and limitations have been fed back to IBM to assist with ongoing improvements to Clinical Trial Matching. Peter MacCallum will work with IBM on the improvements over this time," according to Herath.
Herath added the centre will also wait for the implementation of the Parkville electronic medical record (EMR) before deciding next steps.
Under, the EMR project, Peter MacCallum Cancer Centre, Melbourne Health, and Royal Women's Hospital will link patient records together so records can be updated in real-time and reduce duplication.
The project is part an AU$124 million commitment by the Victorian government announced last May.
The trial by Peter MacCallum Cancer Centre follows on from similar a trial by US-based Highland Oncology Group. It used Watson for Clinical Trial Matching and found it was able cut the time required to screen patients for clinical trial eligibility by 78% -- from 1 hour and 50 minutes of manual curation to 24 minutes, IBM claimed.
Leveraging Google Cloud to identify associations between microorganisms and genes to predict different disease states.
If we understand more about cancer on the molecular level, we can learn to treat it more effectively.
With the help of Harrison-AI, the IVF clinic is preparing to commence trials to validate the use of AI for IVF treatments.
IBM takes on Alzheimer's disease with machine learning
IBM hopes ML can provide the framework for a way to diagnose the illness without the need for spinal fluid extraction.
Fujitsu Australia, GE Healthcare, Macquarie University, and Macquarie Medical Imaging are all involved.