Using AI for safer CT scans

---

Computerised tomography (CT or CAT) scans are widely used in all fields of medicine and surgery. Where treatment of a blood vessel is being considered, doctors need a detailed view inside the blood vessel. For example, blood clots are often found inside arteries or veins. The position and shape of the clot, in relation to the blood vessel wall, is important for the planning of treatment such as the placement of a stent.

On routine CT scans, the human eye cannot detect the presence of abnormalities inside the blood vessel. To make it clearer, a special dye, called a contrast agent, is injected into the patient. This makes it easier to visualise the blood flowing through the blood vessel and the blood clot. These are called ‘contrast enhanced’ CT scans. However, there are several problems with the use of contrast dye. Contrast enhanced CT scans take longer to complete and require more radiation exposure for the patient. Patients usually get some mild reactions to the contrast dye and may be allergic to it, in which case they cannot have contrast enhanced CT scans.

Importantly, the use of contrast dye requires the insertion of a needle into the patient’s arm. This is uncomfortable and can result in problems such as contrast dye leakage outside the injection site causing skin problems. Contrast dye can sometimes damage the kidneys, which is called ‘contrast induced nephropathy’. This can be a particular problem in elderly patients whose kidneys are not working at full capacity.

Although human eyes cannot tell the difference between blood flow and clot within the blood vessel on a routine CT scan, there are minute details in the scan that can be used to differentiate them. Professor Lee’s team has developed a technology to generate contrast enhanced CT scans without the need to inject contrast dye. This research uses artificial intelligence (Al) which is an advanced computing technology whereby a computer is programmed to mimic the learning and problem-solving abilities of the human mind. HRUK will support Professor Lee’s team to further develop their technology.

If successful, this research will allow diagnosis and treatment of blood vessel conditions with a reduced risk of complications. The reduced scanning time, radiation dose and lower cost would bring important benefits. Furthermore, the method can be extended for diagnosis and treatment planning for other medical conditions where contrast enhanced CT scanning is required.