AI machine learning tools offer increasing ability to extract secondary analysis from imaging studies, which translates to meaningful disease insight that was previously overlooked.  Incidental findings are a term used to represent abnormalities that were noticed during the interpretation of an imaging exam that are unrelated to the primary health concern.  One such finding that has been labeled “incidental” by traditional breast radiologists is breast arterial calcification.  Breast arterial calcium (BAC) is infrequently reported and, to our knowledge, no established quantification or scoring guidelines exist or have been conventionally accepted.

Breast arterial calcification (BAC), or calcium within the wall of the breast arteries, is a relatively unambiguous imaging finding though has been traditionally overlooked and underreported on mammography.  Despite the relative ease of identifying this pattern of calcification and benignity in terms of association with breast cancer, multiple studies have demonstrated an association of simply the presence of BAC with “shorter overall survival” mostly due to “death due to cardiovascular events” [1].  In fact, BAC presence occurred in 26.5% of patients in a cohort of over 5000 women aged 60-79 years and was associated with a significantly increased hazard of atherosclerotic CVD (ASCVD) events (HR 1.51) over 6.5 year follow-up [1].  

Although coronary artery calcification(CAC) can be thought of as a natural analogue for BAC, the pathophysiology of BAC materially differs from CAC in that BAC represents medial calcification rather than intimal atherosclerotic deposits [2].  This medial artery calcification compromises the natural elasticity of the artery and leads to increased vessel stiffness, which translates to secondary strain on cardiac afterload.  While the presence of BAC is known to track with severe cardiovascular events and worse mortality, BAC may represent an adequate proxy for medial artery calcification in other medium and large vessels within the body.

In our study, a quantitative BAC model that has been previously shown to be accurate was compared to CT imaging within the head, neck, and body to determine if an association exists between the BAC score and the amount of calcification within medium and large vessels.