Study Highlights Role Of AI In Improving Scan Diagnostics

The importance of medical image sharing will be well known to many involved in patient care, with the ease with which this information can be transmitted between different settings enabling greater access for all those delivering treatment or seeking to carry out a diagnosis.

Since early and accurate diagnosis is so often the key to successful treatment outcomes, the usefulness of medical images produced by scans can be increased further by enhancements in diagnostics. This is an area in which the use of artificial intelligence has produced some exciting developments, often spotting problems the human eye would usually miss.

These have been highlighted in a new study conducted by India-based healthcare technology researcher Sriram Sitaraman, Tech Bullion reports.

His research, published in the International Journal of Scientific Research, highlighted just how much advancement AI is bringing to the area of medical diagnostics.

Perhaps the most telling finding was that suspicious lesions are now detected 2.3 years earlier on average when AI is used than when it is not.

It is not just the all-important early diagnosis rate that has been improved, but workflows for busy staff, with the typical time taken to interpret an image falling from 6.5 minutes to 3.2 minutes, while maintaining a diagnostic accuracy of 96.2 per cent. Anatomical segmentation rates are also high, at 89.2 per cent accuracy.

The report highlighted that while upfront costs for such AI systems are high, they will pay for themselves in two or three years. Such a concern may apply most to the private medical sector, but within a state-run body like the NHS it may be considered that it brings the wider social and economic value of a healthier population.

Indeed, as Public Technology recently reported, the NHS is planning to set up a £180 million framework to enable health bodies to invest in AI for diagnostics. It is set to start taking bids this summer.