OBJECTIVES: Amid the growing shortage of radiologists and workflow pressures, artificial intelligence (AI) software has the potential to increase the accuracy and efficiency of mammographic breast cancer screening. We built a model to evaluate the potential economic impact of AI implementation on screening mammography from the facility and payer perspectives.

METHODS: An Excel-based model was constructed to evaluate the impact of AI implementation on mammography with two types of AI solutions: (1) augmentative; and (2) autonomous rule-out. Augmentative AI detects suspicious lesions and generates a case recommendation that has been shown to reduce read times. Autonomous rule-out AI offers similar benefits, with the added feature of triaging out “normal” cases so radiologists can focus on reviewing high-priority exams. Model inputs were estimated primarily from published literature and the Merative MarketScan Commercial Claims Database. Key outputs included facility read time savings and potential change in revenue, and payer change in spending. Tornado diagrams evaluated the relative impact of key inputs.

RESULTS: Using base case assumptions for a facility that screens 10,000 women/year, augmentative AI could result in time savings of 57 mammography exam reading hours/year, which could be leveraged to potentially increase revenue up to $71.1k/year. Autonomous rule-out AI, assuming "normal" cases read by AI would not be eligible for professional reading fees and no additional reimbursement is available, could save up to 229 hours/year, but could result in a revenue decrease of up to $46.3k/year. Again, assuming no additional reimbursement, autonomous rule-out AI could result in payer savings of $412.5k/year. Facility throughput, read time reductions, and percentage of “normal" exams had the greatest impact on results.

CONCLUSIONS: Augmentative AI offers substantial gains in flexibility and savings for facilities, with a neutral impact on payers. Looking ahead, autonomous rule-out AI may offer significant savings for payers; however, reimbursement will be critical for adoption by facilities.