AI could play vital role in catching cancer early

The use of artificial intelligence for the early detection of cancer is still in its infancy, but it offers huge promise for the future. Getting in on the ground floor is Durotimi AI Technologies, a new cancer-detection platform founded by Doyin Bademosi, which helps clinicians spot the disease before a patient starts showing obvious symptoms.

Bademosi is a physicist by background who became aware of the difficulties of diagnosing cancer without symptoms when a family member was found to have the disease out of the blue. This got Bademosi, who also has an MBA and a postgraduate diploma in innovation and entrepreneurship from Trinity College Dublin, thinking about the need for a diagnostic tool that could dig into a patient’s history and flag the possibility of cancer before the symptoms became apparent.

Durotimi is not Bademosi’s first foray into AI-led diagnostics. He is already the founder of Radmol AI Systems, which is focused on minimising the risk of errors in radiology reporting. During its commercial pilot he had seen radiologists spot numerous cases of cancer in patients who were actually undergoing scanning for something else.

“As things stood there was no way of knowing if someone was at high risk until the symptoms of cancer were visibly apparent by which time it could be too late. Even visiting a GP was no reassurance that someone will be referred for further tests particularly if they are not exhibiting signs of cancer that would make their physician suspicious,” says Bademosi.

A chill on Apple’s fortunes may be abating

---

Student accommodation: A market plagued by bed shortages, rising rents and 51-week leases

---

An attempt to buy the next election with a giveaway budget would be a gross miscalculation

---

Durotimi joins the dots in a patient’s medical history and looks for correlations between symptoms, biomarkers and anything else that might suggest something is wrong

Having become aware that delayed or late diagnosis was a regular occurrence, Bademosi started looking at the possibility of training a machine to analyse patient symptoms and medical histories and to red flag the need for further investigation if something untoward was found.

“The big advantage of machine learning is its ability to analyse a significantly large amount of data to provide credible prediction every time,” says Bademosi, who adds that the expertise of Prof Willie Hamilton of Exeter University was invaluable in modelling Durotimi for clinical use.

In simple terms, Durotimi joins the dots in a patient’s medical history and looks for correlations between symptoms, biomarkers and anything else that might suggest something is wrong. The tool is integrated into existing records systems and whirrs away in the background as the GP goes through a normal consultation or check-up.

[ AI-supported mammography screening is safe and almost halves radiologist workload, Lancet study finds ]

“Our platform enables non-intrusive cancer screening with every doctor’s visit. It cross-references multiple diagnostic pathways and gives GPs the ability to concurrently check combinations of signs, symptoms and risk factors in an intuitive format during the consultation when the opportunity to affect decision-making is at its greatest,” says Bademosi.

Durotimi was established earlier this year following an 18-month research period, and the company, which now employs six, is based at UCD’s Nexus centre for industry partnerships, where it is also collaborating with researchers from CeADAR, Ireland’s National Centre for Applied AI.

Investment in the business to date is about €500,000 between support from EIT Digital (the European Institute of Innovation and Technology) and backing from Radmol. The company’s SaaS solution is aimed at a global audience of GPs, hospitals, and other clinicians and is due for launch in Q1 next year.

The next big step for Durotimi is a fundraising round of €5 million, which is in preparation at the moment. “Radmol is on a mission to minimise the risk of delay and errors in medical diagnosis, while Durotimi is on a mission to minimise the risk of delay and errors in cancer diagnosis. Our ability to integrate both solutions sets us apart in the emerging world of AI diagnosis,” says Bademosi. “At the moment we are working on lung and colorectal cancer but with our funding in place we will be expanding our pilots, deploying our system with more clinicians and extending it into others areas such as breast and cervical cancers.”