Skip to main content
Tessa NewsLink
Paste a health news link, or browse
Latest checks

Source study unconfirmed

Story checked

AI could help identify people at higher risk of common cancers using routine health data (opens in a new tab)

medicalxpress.com · 2026-10-26

Short answerEvidenceSource

Short answer

Source study unconfirmed

Study not confirmed.

The search did not establish a source study for this article. Its claims remain unverified; this does not mean they are incorrect.

15 candidate papersExplore the search results

These papers were considered as possible sources, but none could be confirmed as the study behind this story.

Share this check

Follow the evidence trail
1
2

NewsLink checks it

No single source study could be confirmed.

The search identified related research but did not establish a single source study. Related papers are provided for context, not as verification of the article's claims.

View article claims
3

Search result

Source study unconfirmed.

The search identified related research but did not establish a single source study. Related papers are provided for context, not as verification of the article's claims.

The literature search

15 candidate papers

Studies identified during the search

These papers appeared in the search for the story’s source. None was confirmed as a direct match. Inclusion here does not mean a paper supports or contradicts the story, or that its full text was reviewed.

Show 11 more search results
Then inspect each claim

Evidence layer

Claim by claim

The article's claims are listed below. Without a confirmed source study, they have not been assessed as supported or unsupported.

Claim 1 of 8Not checkedMachine learning models using routinely collected health information could help identify people at higher risk of several common cancers, according to an international study reported at the ESMO Congress 2026.View evidenceHide evidence

Not checked

No source study was confirmed for this article. This claim remains unverified, rather than assessed as supported or unsupported.

Claim 2 of 8Not checkedResearchers developed approximately 500 machine learning models to predict risk across 20 common cancers using data from 463,620 UK Biobank participants followed for approximately 18 years; 57,055 participants developed cancer during that period.View evidenceHide evidence

As statedapproximately 500 models; 20 cancers; 463,620 participants; 18 years; 57,055 cancer cases

Not checked

No source study was confirmed for this article. This claim remains unverified, rather than assessed as supported or unsupported.

Claim 3 of 8Not checkedThe models were externally tested in Scotland and Wales, 42,036 people from the U.S. All of Us research program, and more than 3 million people from Japan's JMDC database, and cancers diagnosed within the first year were excluded from the relevant models to reduce the influence of undiagnosed disease.View evidenceHide evidence

As stated42,036 U.S. participants; more than 3 million in Japan

Not checked

No source study was confirmed for this article. This claim remains unverified, rather than assessed as supported or unsupported.

Claim 4 of 8Not checkedFor a combined cancer outcome covering cancers that could be predicted from routine clinical data, the models achieved an AUROC of 0.83 at 15 years and outperformed traditional statistical approaches, though performance varied by cancer type.View evidenceHide evidence

As statedAUROC of 0.83 at 15 years

Not checked

No source study was confirmed for this article. This claim remains unverified, rather than assessed as supported or unsupported.

Claim 5 of 8Not checkedPerformance was particularly strong for lung, esophageal and liver cancers, but more limited for ovarian, breast and cervical cancers.View evidenceHide evidence

Not checked

No source study was confirmed for this article. This claim remains unverified, rather than assessed as supported or unsupported.

Claim 6 of 8Not checkedMuch of the predictive information came from routine blood tests and clinical history; blood tests continued to add predictive value for cancers diagnosed more than five years later, while genetic and metabolomic data added only a small further improvement.View evidenceHide evidence

Not checked

No source study was confirmed for this article. This claim remains unverified, rather than assessed as supported or unsupported.

Claim 7 of 8Not checkedThe team also developed PRE-Screen, a toolbox that lets AI agents based on large language models run the risk models consistently, and the article says agents using it performed substantially better than LLMs making predictions unaided.View evidenceHide evidence

Not checked

No source study was confirmed for this article. This claim remains unverified, rather than assessed as supported or unsupported.

Claim 8 of 8Not checkedLead author Dr. Paul-Henry Koop said that for selected cancers the models could serve as a low-cost pre-screening step using data already available in routine care, but the article notes that many people flagged as higher risk would not develop cancer and that further research is needed to show earlier diagnosis or better outcomes.View evidenceHide evidence

Not checked

No source study was confirmed for this article. This claim remains unverified, rather than assessed as supported or unsupported.

Finally, the search trail

Method layer

Study not confirmed.

The search results are available for you to inspect. Without a confirmed source study, they cannot establish whether the story’s claims are supported or contradicted.

Search results

Candidate papers, not confirmed sources.

PubMed, Crossref, Europe PMC · 15 candidate papers

CandidateOpen access

From Nutritional Epidemiology to Clinical Prediction: Interpretable Machine Learning Reveals Subtle Associations Between Dietary Antioxidant Profiles and Non-Melanoma Skin Cancer in Older Adults.

Food Science & Nutrition · 2026 · PubMed, Europe PMC

Tessa (opens in a new tab)Source (opens in a new tab)

Explore all search results