NHS Shared Business Services has issued a tender worth £150m, split into six lots, covering the use of artificial intelligence (AI) in medical imaging and analysis to speed up diagnosis.
AI systems can be deployed to analyse medical images in seconds, and help detect conditions such as strokes, asymptomatic conditions, future diseases, suspicious areas, small tumours or subtle abnormalities. They can provide real-time, evidence-based recommendations during patient consultations. NHS Shared Business Services said these systems can suggest potential diagnoses and treatment options based on the latest clinical guidelines and patient data.
In a post on the NHS Shared Business Services website, Kelly Bevington, senior category manager of digital and IT at NHS Shared Business Services, said: “With our current artificial intelligence offerings – Artificial Intelligence Software in Neuroscience for Stroke Decision Support and Artificial Intelligence, Imaging and Radiotherapy Equipment, Associated Products and Diagnostic Imaging – due to expire, we wanted to bring together the wealth of experience of these framework agreements into one comprehensive offering that, to quote Lord Darzi’s independent review, helps take the NHS above the foothills of digital transformation.”
She said the framework agreement will focus on the application of AI in different specialities by providing efficient ways to prevent, diagnose, treat illness and optimise clinical workflow.
The first lot covers the use of AI image analysis, including diagnosis of neurological disorders such as Alzheimer’s disease, Parkinson’s disease and multiple sclerosis through imaging and other diagnostic tests; and the use of AI medical imaging to detect lumps, ovarian cancer, endometriosis, fetal structures, anatomical development, and early detection of pregnancy complications.
In oncology, AI can be used to aid in the early detection and diagnosis of various cancers by analysing imaging data, genetic information and patient records.
Other AI uses include supporting musculoskeletal conditions; cardiology; gastroenterology; urology; ophthalmology; and the analysis of neurological data to identify risk of stroke or stroke occurrence.
The second lot focuses on the adoption of AI algorithms and deep learning models, which are trained on large datasets and can be used to identify patterns that cannot be detected in a normal clinical setting or laboratory.
For instance, AI can analyse digital slides from biopsies, tissue, cells, blood and bone marrow to help detect cancer and other diseases, optimising workflow in pathology labs.
Lot three covers predictive analytics for improving patient flow and experience based on better interoperability with health records. Lot 4 is about research and development to support cross institution collaboration and research such as drug discovery and clinical trials. The Lot 5 is focused on operational efficiency, and Lot 6 covers AI consultancy, implementation and training.
