Using National Health Data to Predict Cancer Risks

Scientists from the #German Cancer Research Center (DKFZ) and the #European Bioinformatics Institute EMBL-EBI have developed a model to predict individual risks for 20 different types of cancer using data from Danish health registers. This model, trained on data from 6.7 million adults between 1995 and 2014, includes over 1,000 prior diagnoses, family #cancer history, age, and lifestyle factors like smoking and obesity.

Validated with data from 4.7 million Danes from 2015 to 2018, the model showed high predictive accuracy, particularly for cancers of the #digestive system, #thyroid, #kidney, and #uterus. It achieved an overall accuracy of 81% over a lifetime and 59% when considering age and gender. The model's performance was confirmed using data from the #UK Biobank, indicating its potential applicability to other #healthcare systems.

This model can help identify individuals at high risk for cancer, allowing for targeted #early detection programs, which could #improve cure rates and reduce the need for #intensive treatments. Researchers suggest that collecting comprehensive health data, including body measurements and known risk factors, could enhance the model’s accuracy. They advocate for the development of national digital health infrastructures to support such #predictive models in other countries.

Journal Reference: Alexander W Jung, Peter C Holm, Kumar Gaurav, Jessica Xin Hjaltelin, Davide Placido, Laust Hvas Mortensen, Ewan Birney, S⊘ren Brunak, Moritz Gerstung. Multi-cancer risk stratification based on national health data: a retrospective modelling and validation study. The Lancet Digital Health, 2024; 6 (6): e396 DOI: 10.1016/S2589-7500(24)00062-1