Cancer detection is one of the big use cases AI cheerleaders have shouted about in recent years, so today let’s take a look at an early-stage startup in that particular space.

Read on to find out all about Spotlight Pathology.

– Martin

Spotlight Pathology is using AI to get the right blood cancer diagnosis, faster

Can AI really provide a helping hand in cancer detection? Plenty of people believe it can. and Spotlight Pathology is a startup that agrees.

The team’s software is initially being built for enabling faster diagnosis of lymphoma and leukaemia by analysing digitised images of tissue biopsies using proprietary computer vision models trained on UK hospital datasets.

The software’s algorithm then identifies if the sample in question is cancerous, and if so, what subtype of cancer is present, CEO Sam Perona says. But the tech is not designed to replace human specialists.

“People are still very much doing the diagnosis. And in the case of blood cancer, there are a lot of different pieces of evidence and data that they get to help make that definition. So this is one more piece of data to help get a correct diagnosis, but the way it works also helps to streamline and speed up the diagnosis,” Perona says.

Spotlight Pathology’s software can help generalist professionals figure out which samples to send to a specialist for a diagnosis. While the specialist’s workload can be sped up significantly.

“Blood cancer is particularly aggressive, so that time to diagnosis really does matter,” Perona says.

“It's allowing them to only need to sit and look at the sample once, as opposed to two or maybe multiple times. [Without Spotlight Pathology,] each time they look at them, they have to send off for other tests, so you've got an extra day or two days in between. So it's allowing them to sit down once with all the information and make the diagnosis, which shortens that time to make the diagnosis.

“We integrate with the scanners and the systems so that when the hematopathologist looks at their image on the same screen they would have looked at it before, they can also push a button and get the diagnosis that Spotlight offers as well, to support their diagnosis, or to make them look again at their diagnosis, or to help streamline their workflow.”

The story so far

Now based at Sci-Tech Daresbury in Cheshire, Spotlight Pathology began life as a spinout from the University of Manchester. It was co-founded by Dr Martin Fergie (previously CTO of DigitalBridge, a Manchester-based computer vision startup I wrote about a decade ago) alongside pathology specialist Professor Richard Byers.

Fergie and Byers were joined by Perona, who took up the CEO role last year following a successful exit in 2022 for her previous company, Perfectus Biomed Group.

The software was developed and trained on a combination of training data sets and additional hospital data sourced specifically to confirm the tech’s capabilities. Now the startup is in the midst of the necessary audits before it submits the product for regulatory approval in the UK, followed by the EU and USA.

Perona is optimistic that this process will be complete relatively quickly, allowing Spotlight Pathology to begin making sales in the UK and Europe next year. Its target customer base includes NHS trusts, private hospitals, and third-party labs.

One essential prerequisite for a potential Spotlight Pathology customer is that their scans are available in digital form.

“They need to have moved from using a microscope to a scanner to look at their images,” Perona says.

“A couple of years ago there was a bit more of a question of whether departments should digitise. But now what we're finding when we speak to people is that their sites are either already digitised or on the way to being digitised, or they see the benefit and they want to be digitised.”

Subscribe to Premium to read the rest.