{
  "abstract": "Barraclough and colleagues1 are to be congratulated on their paper published in Heart. The authors report on the work of a diverse team of clinicians, imaging scientists and bioinformaticians to discover and validate a machine learning algorithm to predict cardiovascular events from routine mammograms. Mammographic features, such as breast arterial calcification and tissue density, have previously been recognised as associated with the risk of cardiovascular disease. However, Barraclough et al 1 have taken a novel approach to progress this further. They have applied machine learning approaches to mammograms from the 49 196 women in the Lifepool cohort with linked hospitalisation and death outcome data. This modelling benefited from a median follow-up of 8.8 years, and a substantial number of first major cardiovascular events. Using a deep learning model, and only mammography features and participant age, the team were able to predict cardiovascular events with a similar performance to complex risk algorithms, such as the America Heart Association PREVENT (Predicting Risk of Cardiovascular Disease Events) equation.",
  "authors": [
    {
      "affiliations": [
        "Kolling Institute, University of Sydney, Sydney, New South Wales, Australia",
        "Cardiology, Royal North Shore Hospital, St Leonards, New South Wales, Australia"
      ],
      "name": "Gemma A Figtree"
    },
    {
      "affiliations": [
        "Charles Perkins Centre, University of Sydney, Sydney, New South Wales, Australia"
      ],
      "name": "Stuart M Grieve"
    }
  ],
  "title": "Mammography—an opportunity to optimise women’s heart health?",
  "uid": "57ff6cab-82b7-5148-9ba5-1a28b6817cb8"
}