Optimal Breast Biopsy Decision-Making Based on Mammographic Features and Demographic Factors

Area of Science:

• Radiology
• Medical Decision Making
• Biostatistics

Background:

• Breast cancer is a leading cancer in US women, with millions of mammograms performed annually.
• Over 700,000 breast biopsies are conducted yearly in the US, but 55-85% reveal benign lesions, leading to overtreatment and costs.
• Current radiologist decision-making for biopsies lacks a standardized, data-driven approach considering patient demographics.

Purpose of the Study:

• To develop an optimal decision-making model for breast biopsy referrals based on mammographic findings and patient demographics.
• To address the problem of unnecessary biopsies and their associated patient anxiety and healthcare expenditures.
• To create a model that assists radiologists in making more accurate biopsy decisions.

Main Methods:

• Formulation of the breast biopsy decision problem as a finite-horizon discrete-time Markov decision process.
• Development of an optimal policy derived from the Markov decision process model.
• Application and validation of the model using clinical data to compare its performance against radiologists.

Main Results:

• The optimal policy indicates that patient age is a critical factor in biopsy decisions.
• The model recommends a higher risk threshold for biopsy in older patients compared to younger patients.
• The developed model demonstrated superior performance compared to human radiologists in clinical data analysis for biopsy decision-making.

Conclusions:

• A Markov decision process model provides an effective framework for optimizing breast biopsy decisions.
• Incorporating patient age into the decision-making process can significantly improve the accuracy of biopsy recommendations.
• The model offers a data-driven approach to reduce unnecessary biopsies, patient anxiety, and healthcare costs in breast cancer screening.