New Zealand throws away thousands of bowel screening samples every year once they have been checked for traces of blood. Auckland researchers now think those leftover tubes could hold the key to catching bowel cancer far earlier, and they want to use artificial intelligence to do it.
The idea comes from Dr Theo Portlock, a research fellow at the University of Auckland’s Liggins Institute. He wants to take the residual material left over in the home test kits used by the national bowel screening programme and train an AI to read the mix of gut bacteria inside, looking for the signature of cancer long before a tumour would normally be found.
The test at the centre of the plan is the faecal immunochemical test, known as the FIT. It is the simple home kit posted out to people in the screening programme, and it works by detecting tiny amounts of blood in a stool sample. Blood can be an early warning sign of bowel cancer, but it can also turn up for plenty of harmless reasons, which is part of why so many people who return a positive test go on to have a clear colonoscopy.
That is the gap Dr Portlock wants to close. Rather than throwing the tubes out after the blood check, his team would analyse the gut microbiome inside them, the vast community of bacteria that lives in everyone’s digestive system. Certain shifts in that community appear to track with bowel cancer, and the relationships between the different species are far too tangled for a person to untangle by eye.
“AI is very good at doing that understanding, something that we wouldn’t be able to do,” Dr Portlock told RNZ. He explained that the patterns are rarely as simple as one bug going up or down. “Sometimes you might have an increase in one or a decrease in another,” he said, describing the kind of subtle, multi-species relationship that a trained model can learn but a standard lab analysis would miss.
The team is aiming for accuracy of up to 90 percent in sorting people who genuinely have bowel cancer from those who returned a positive blood test for some other reason. If it works, Dr Portlock believes it could sharpen the whole screening pipeline. The approach would “help reduce false positives, it will help reduce false negatives, and hopefully reduce waiting times,” he said.
Each of those gains matters in its own way. A false positive sends a worried person to a colonoscopy they did not need, using up scarce specialist time. A false negative is worse, because it can mean a cancer is missed entirely until it has grown and spread. Trimming both at once, while easing the queues for follow-up procedures, would be a rare win for a screening system that is under constant pressure.
The reason the work feels urgent is grim. New Zealand has the fastest rate of early-onset bowel cancer in the world, meaning the disease is striking younger people here more quickly than anywhere else, and researchers still do not know why. Bowel cancer is also one of the country’s biggest cancer killers. More than 3,300 New Zealanders are diagnosed with it every year and more than 1,200 die from it, according to Health New Zealand.
Catching the disease early changes the odds dramatically. Bowel cancer found at an early stage is often very treatable, while cancer found late is much harder to beat. That is the entire logic of screening, and it is why the programme exists in the first place.
At the moment the national bowel screening programme offers free testing to people aged 58 to 74 across the country. From 30 September 2026 the starting age will drop to 56, a change the Government has said will bring more people into screening earlier in life. The expansion will only add to the number of FIT kits flowing through the system, and with them, the pile of leftover samples that Dr Portlock’s project would put to use.
One of the quiet strengths of the plan is that it asks very little of the public. The samples are already being collected, already being posted in, and already being processed. People would not need to do a new test or hand over anything extra. The material that is currently binned would simply be analysed a second time before being discarded, turning a piece of medical waste into a research and diagnostic resource.
There is a catch, and it is the familiar one. The project still needs funding to get off the ground, and no firm timeline has been set for when it might feed into the real screening programme. Like a lot of promising New Zealand research, it sits in the waiting room between a good idea and the money to prove it works at scale.
Using AI to read medical samples is not unique to this study. Around the world, researchers are training models to spot polyps during colonoscopies, to read scans and to pre-screen biopsies, and the bowel is one of the most active areas of that work. What makes the Auckland approach stand out is how cheaply it could slot into an existing system, riding on tests that hundreds of thousands of New Zealanders already complete each year.
For now it remains a proposal rather than a tool in clinics. But the basic pitch is a tidy one. The country has a serious bowel cancer problem, a screening programme that already reaches deep into the population, and a stream of samples that currently goes in the bin. The hope is that a well-trained model can read more out of those samples than anyone has managed before, and give doctors a head start on a disease where every week counts.
What do you make of using AI and leftover screening samples to catch bowel cancer earlier? Have you been through the bowel screening programme yourself? Share your thoughts in the comments below.
This article was written by AI, briefed to report the facts, hopefully without some of the bias people bring to the job 🙂
Leave a Reply