Pathologists welcome artificial intelligence
AI in automation is accounting for laboratory advancements in areas that include testing systems, home collection, home testing, PoC (Point of Care) testing, genome sequencing, and digital pathology, shared Dr Junaid Bajwa, Chief Medical Scientist, Microsoft, UK, in his keynote. AI has the potential to reduce costs, support clinicians and potentially generate new novel hypotheses.
Autonomous labs “are not far away,” he added, offering the example of a Liverpool “self-driving” laboratory with robots.
Dr Aaron Han, Consultant Pathologist, Vice President, Emirates Pathology Society, Adjunct Professor, MBRU School of Medicine, shared insights into the future of digital pathology, including the increased adoption of AI.
Current examples of AI’s use in the lab include its differentiation between Helicobacter Pylori gastritis and autoimmune gastritis, with results comparable to gastrointestinal pathologists.
AI will ensure that processes are made more efficient and save time, he emphasised.
He shared the example of AI introduced for advance requesting of immunohistochemistry in diagnostically uncertain prostate biopsies. AI will reduce the time taken in ordering basal keratin or IHC by two days (from 5-6) once the machine screens and decides on whether to go ahead.
After two days, the pathologist will be informed by their senior tech that the “immunos” are ready, allowing them to review the case and sign out with a click.
The introduction of AI and automation to the laboratory was welcomed by Dr Sayed, who was speaking during a panel talk on the future of laboratory leaders. “It’s improving the way we do our job. We need it to improve our service and reporting methods. Utilising innovations can really help us in reaching accurate results in an efficient manner.”
Her sentiment was echoed by Dr Al-Delfi, who described technology as “our friend. Automation is already here. We should not be afraid of it. Change is the only constant in life.”
Dr Tanios was also in agreement. “We need to incorporate any new advances such as full lab automation, and staff need to be involved more in the process. This is the future.”
Pathologists needn’t worry about AI, Dr Han concluded. “We are the ones that decide on what’s important to study. Human operators will still be required to program the machine, validate and verify the performance.”
“Once we can use AI, digital pathology will have a big advantage over the microscope,” said Prof Ruthy Shaco-Levy, Head, Israeli Pathologists Association, Soroka Hospital. Diagnostic algorithms will save the pathologist time and help unsure diagnostics are more accurate.
Michel added that “labs with an ability to digitise and put machine learning into their laboratory information system are going to be much ahead in providing proactive care, while at the same time supporting population health.”
AI isn’t entirely infallible, however, commented Prof Ulysses G. J. Balis, Professor of Pathology & Director, Division of Pathology Informatics, Director, Computational Pathology Laboratory Section, University of Michigan.
Speaking on AI applications in clinical chemistry, he highlighted that it is especially important to validate algorithms before they are introduced. Thereafter, there needs to be supervision of algorithms, he continued, and constant vigilance, so that the process becomes iterative.
Dr Aaron Han