Artificial intelligence (AI) is showing “great promise in enhancing clinical decision-making,” according to a new report.
GlobalData’s The Future of Work in Healthcare report details that AI can be used to help improve both disease diagnosis and treatment plans, leading to better patient outcomes. It notes in particular the use of the technology for helping to interpret visual data like scans and to determine therapy approaches.
Smarter leaders trust GlobalData
By way of context, the report explains that clinical decision-making can be challenging due in particular to “the large volume of patients with complex diseases, the array of diagnostic tools available and the need to interpret multiple sets of results.” It adds that interpretation technologies like AI can help to simplify the process.
AI diagnostics
For diagnostics, it says there is promise in computer vision (CV), a field of AI that allows machines to acquire visual data, interpret images and make decisions based on them. CV can help radiologists find abnormalities in patient scan images, with the report giving the example of Bayer’s Calantic Digital Solution, which offers AI apps covering neuro, thoracic, breast, prostate and cardiac diseases.
The report adds: “Studies have indicated that the sensitivity of AI readings remains consistent across different breast densities, whereas human sensitivity tends to diminish in the presence of dense breast tissue. This consistency allows radiologists and other healthcare providers (HCPs) to place greater confidence in their diagnoses and partially automates the clinical decision-making process.”
Crucially, the report also notes that AI aids the detection of incidental findings in medical images.
“These findings are abnormalities found unintentionally, unrelated to the patient’s initial concern,” it explains. “HCPs may miss these irregularities, particularly when patients do not fit the typical demographic profile associated with the disease. AI-powered software that analyses medical images is more adept at identifying these anomalies.”
AI therapy selection
Subsequent to diagnosis, the report outlines that AI can aid not only in therapy selection for patients but also for the patients’ prognosis.
By way of example, the report says: “Vesta, developed by Valar Labs, uses AI to analyse pathology slides and predict responses to bacillus Calmette-Guérin, the standard of care in high-grade non-muscle invasive bladder cancer. Vesta has been trained by machine learning models using thousands of digital pathology images. The model analyses the details like the shape and location of tumour cells, identifying patterns in the slides linked to poorer outcomes, allowing Vesta to understand the likelihood of recurrence and progression of the tumour.”
