AI Insight
This large retrospective cohort study of over 575,000 adults with newly diagnosed type 2 diabetes found that cancer risk varies significantly by diabetes subtype. Severe insulin-deficient and mixed subtypes showed the highest risks for colorectal, pancreatic, liver, endometrial, and ovarian cancers, while the mild obesity-related subtype had lower risks for prostate and breast cancer. Cancer screening rates were lower across all diabetes subtypes compared to adults without diabetes.
Why it matters
These findings suggest that type 2 diabetes should not be treated as a uniform disease when assessing cancer risk. Classifying patients into clinical subtypes using routine health data could enable more personalized cancer surveillance strategies, potentially improving early detection in high-risk groups while avoiding overscreening in lower-risk populations.
Understand the Science
⚠️ Preprint – Noch nicht peer-reviewed
Dieser Artikel wurde noch nicht von unabhängigen Experten begutachtet. Die Ergebnisse sind vorläufig und sollten mit Vorsicht interpretiert werden.
BackgroundType 2 diabetes (T2D) is associated with elevated rates of several cancers and is increasingly recognized as a heterogeneous disease, but whether its clinically distinct subtypes carry different cancer risks is unknown.
MethodsIn this matched retrospective cohort study using electronic health record data from the Epic Cosmos Research Platform (2012-2025), adults with newly diagnosed T2D were classified into severe insulin-deficient (SIDD, 21.6%), mild obesity-related (MOD, 23.5%), mild age-related (MARD, 40.7%), or mixed (14.1%) subtypes using validated algorithms and matched to adults without diabetes on age, sex, and body mass index. Cause-specific Cox models estimated adjusted hazard ratios (HRs) for seven site-specific cancers, accounting for competing risks. Cancer screening uptake was assessed as a secondary outcome.
ResultsAmong 575,139 adults with T2D and 689,719 without diabetes (median follow-up, 3.8 years), MARD had the highest cancer incidence (17.3 per 1,000 person-years). Relative to adults without diabetes, rates of colorectal, pancreatic, liver, endometrial, and ovarian cancer were elevated across subtypes, with the highest hazards in SIDD (HR=3.87, 95% CI=3.51 to 4.27) and mixed phenotypes. Prostate cancer rates were lower in all subtypes, most markedly in MOD (HR=0.60, 95% CI=0.55 to 0.64). Rates of breast cancer were higher among mixed (HR=1.12, 95% CI=1.05 to 1.19) and lower among MOD (HR=0.85, 95% CI=0.80 to 0.90). Mammography and prostate-specific antigen screening were lower across subtypes.
ConclusionsSite-specific cancer incidence and screening uptake differed across clinically defined subtypes of T2D. Subtype classification from routine clinical data may inform targeted cancer surveillance, though further study is needed before clinical use.