More young people are developing colorectal cancer

“The moment when the study results were presented is something I will carry with me for the rest of my life. At first, there was complete silence for 30 seconds. Then what we were seeing on the screen sank in, and the room erupted in jubilations,” says Johan Lindberg, Associate Professor at the Department of Medical Epidemiology and Biostatistics at Karolinska Institutet. 

The ALASCCA study showed that a low dose of acetylsalicylic acid halves the risk of recurrence in a common subtype of colorectal cancer. The effect is greater than that achieved by most other adjunctive therapies, says Anna Martling, Professor of Surgery at the Department of Molecular Medicine and Surgery at Karolinska Institutet. 

“The drug is also very inexpensive and has few side effects compared with chemotherapy or immunotherapy, or indeed compared with almost any other adjunctive treatment,” says Anna Martling, who was the principal investigator responsible for the study. 

When the study was launched in 2016, researchers had, for a long time, observed a recurring pattern: acetylsalicylic acid, ASA, appeared to influence cancer progression, but only in certain cases. Particularly in cardiovascular care, where it has long been used under the brand name Trombyl, and from that there were additional findings that had to be looked into. 

Attention turned to tumours with abnormalities in the PIK3 genes, which regulate cell growth and cell division. Disrupted signalling in these genes can lead to uncontrolled cell division, in other words cancer. 

Designed a large study

Using these indications, the researchers designed a large study. A total of 3,500 patients from 33 hospitals across the Nordic countries had their bowel tumours genetically analysed. Those with alterations in the PIK3 signalling pathway were randomly assigned to one of two groups, with one receiving low-dose ASA and the other a placebo. 

Johan Lindberg was given the task of building a laboratory capable of carrying out all these genetic analyses. It was a logistical challenge with some very practical difficulties, one being recruitment. Bioinformaticians specialising in genetic data analysis were in short supply when the study began. Furthermore, the laboratory, which was part of a research environment, had to deliver clinical results at the pace required by healthcare services, since the patients in the stidy were undergoing assessment and awaiting treatment. 

“Some weeks we had to deliver 30 analyses, other weeks none at all. It was a marathon to keep everything running, weekdays, holidays, all year round,” says Johan Lindberg. 

The study has now been completed. It shows that one in three colorectal cancer patients has genetic abnormalities somewhere along the PIK3 signalling pathway. 

Almost 15 per cent of patients treated according to standard care, meaning they received a placebo, experienced replase within three years. Among those who received ASA in addition, the figure was half that. 

In Sweden alone, ASA treatment is expected to save the lives of 200 to 300 colorectal cancer patients every year, by conservative estimates. Globally, this could translate into around 50,000 lives saved annually. 

Revised at record speed

Clinical guidelines for colorectal cancer treatment were revised at record speed. Patients with the relevant genetic alterations are now recommended to receive ASA as an adjunctive treatment. 

This places demands on healthcare services throughout the country. Every patient assessment must include the molecular genetic analyses required, known in professional terminology as a “broad panel”, covering approximately 400 genes. 

In Stockholm, clinical services were being prepared while the study was still ongoing. Johan Lindberg now also works part-time at Karolinska University Hospital, which has taken over both the technical software and the routines developed through the project. 

Each genetic analysis costs around SEK 20,000. The cost is justified because PIK3 mutations are common and ASA is both inexpensive and effective. However, these analyses also provide additional information that may be important when making treatment decisions. Sometimes they uncover other genetic alterations, some of which are very rare. In some cases, there are drugs that are an excellent match. 

“You could say that this particular precision diagnostics came as an added bonus to what we were doing,” says Johan Lindberg. 

Several AI models are being trained on this image material to answer different questions. One model learns to identify alterations in the PIK3CA gene, the gene most commonly mutated within the PIK3 signalling pathway. An AI system like that could potentially identify likely mutation carriers, meaning that a broad genetic panel would need to be ordered only for those patients. 

“Molecular diagnostics is still expensive, which limits its use. That is particularly evident in other parts of the world, and it is there that we see the greatest potential for the use of this type of AI tool,” says Mattias Rantalainen. 

Another AI model is being trained to assess how aggressive an individual patient’s illness is. It is conceivable, indeed likely, that algorithms will detect patterns invisible to the human eye. This is already happening in other cancers. AI can, for example, identify aggressive breast tumours that pathologists cannot distinguish using microscopic images alone. 

“The strength of AI is that it can move beyond known markers and tissue patterns associated with disease. We want AI to learn freely from the images themselves, combined with clinical outcome data. Ultimately, we may be able to produce what could be called a digital biomarker,” says Mattias Rantalainen. 

Such a biomarker would need to be validated to ensure that it accurately predicts outcomes across different study populations before being introduced into clinical practice. Even then, it may remain unclear exactly what the AI model is identifying in the images. It may be correct without researchers or clinicians fully understanding what it sees. 

“That is not unusual at all,” says Mattias Rantalainen. 

“The same is true of some biomarkers. We sometimes measure things that we know are important without fully understanding the molecular mechanisms behind them.” 

Assessing disease severity

The aim is to develop a well-trained AI model capable of independently assessing disease severity. That information could complement the rest of the clinical assessment and help guide doctors in treatment decisions. 

“In general, it is wise to proceed cautiously when introducing new biomarkers, whether they are digital or not,” says Mattias Rantalainen. 

Treatment for colorectal cancer has advanced considerably in recent decades. As with many other cancers, immunotherapy and targeted antibody treatments are now available, offering both effectiveness and precision. Surgery, chemotherapy and radiotherapy have also become more sophisticated and more personalised. Some patients receive more aggressive surgery, while others need none at all. In certain situations, radiotherapy can replace surgery, preserving the bowel and allowing patients to avoid a stoma. 

“This places greater demands on assessment, both before and during treatment. We can fine-tune and adapt treatment according to the characteristics of the tumour and how the patient responds. At the same time, being a patient becomes more challenging because treatment is less predictable,” says Anna Martling. 

She explains that younger patients are often treated with more intesive therapies than older patients, even when disease severity is comparable. A reason for that is that younger people are generally better able to manage intensive treatment. Another is that younger patients are more likely to have advanced disease at diagnosis. Researchers do not know why. 

“It may be because younger patients develop biologically more aggressive tumours. But it could also be because delays in detection and investigation are more common among younger people,” says Anna Martling. 

Researchers have long advocated lowering the age threshold for the national colorectal cancer screening programme. One of them is Anna Forsberg, Associate Professor at the Department of Medicine, Solna, Karolinska Institutet. She led the SCREESCO study, which provided important evidence for Sweden’s screening programme. 

The study examined two screening approaches. Some participants were asked to hand in stool samples on two occasions, while others underwent a single colonoscopy. Both approaches were compared with a shared control group. 

More cancers and polyps detected

In both cases, more cancers and polyps, which can be precursors to cancer, were detected. Participation was higher in the stool-sample group, with more people declining colonoscopy. 

Under the current programme, everyone aged 60 to 74 is invited to submit a stool sample every second year. If blood is detected, even in microscopic quantities invisible to the naked eye, a colonoscopy is offered. This reduces mortality by 15 per cent, corresponding to around 300 lives saved each year once the programme is fully implemented. And from this year, colorectal cancer screening has been rolled out nationwide for the entire target age group, although the regions have had to work hard to achieve this. 

“I think the chosen method is the right one. But we should lower the lower age limit to 50,” says Anna Forsberg. 

She was a member of the expert group consulted by the National Board of Health and Welfare when the age limits were recently reviewed. The group unanimously recommended lowering the screening age to 50. However, the Board reached a different conclusion, recommending that the regions continue screening people aged 60 to 74 because of limited colonoscopy capacity. 

“We discussed access to colonoscopy in the expert group. In my view, we could consider alternative solutions to make resources go further. For example, people with absolutely no blood in their stool, meaning a result of zero, could perhaps be invited back after four years instead,” says Anna Forsberg. 

The age limits may yet be reconsidered. The Swedish Government has given the National Board of Health and Welfare the task of reviewing all national screening programmes and assessing, among other things, whether age thresholds should be revised. An initial report is due to be submitted in February 2027.