Artificial Intelligence (AI) Stroke Imaging System

The Health Board are leading on a Procurement for Artificial Intelligence (AI) Stroke Imaging for NHS Wales.

In the UK, most stroke patients present to Primary Stroke Centres (PSC) where intravenous thrombolysis can be delivered, provided patients are identified and the physicians are able to administer therapy. However, there can be a lack of specialist neuroradiology or clinical expertise to aid physicians in the interpretation of acute stroke scans and the time-critical management decisions of these patients. There is a need to support physicians to make decisions about both intravenous thrombolysis and mechanical thrombectomy. Effective determination of stroke diagnosis can be challenging clinically, but positive identification of ischemia on imaging is associated with improved treatment decisions and patient flow.

At PSCs imaging is frequently limited to non-contrast CT scanning alone, or at most with the addition of CT angiography (CTA), and without ready access to MRI or CTP, further challenging appropriate patient selection. A key challenge for the NHS healthcare system is to identify patients who will benefit from thrombectomy without transferring excessive numbers of unsuitable patients to the CSC, consuming unnecessary and costly clinical resources, and overwhelming the capacity of these specialist services.

Once patients at PSCs are screened for transfer to a CSC, both imaging and clinical information then needs to be transferred to the CSC at the time that a referral takes place to make the final decision. While clinical data can usually be communicated by telephone, imaging data necessarily requires a computer-based solution.

Therefore, the Health Board is seeking a purchase of an AI stroke imaging system with support services over a three to five year period.