Novel Contrast-Enhanced Mammography Method

University of Pittsburgh researchers have developed an automated method to quantify background parenchymal enhancement (BPE) in contrast-enhanced mammography (CEM). Designed using an AI-based model to segment images of breast tissue into fibroglandular or fatty tissue, this method can quantify BPE, an indicator of blood flow in the breast tissue. Quantitative analysis of BPE volume, differences in BPE across tissue types, and changes in BPE with time could inform the development of risk-prediction and clinical decision-making tools in breast cancer.

A novel method has been developed to automate the quantification of BPE in CEM. Using low energy with recombined images and AI segmentation of fibroglandular and fatty tissue, overlay of these segments onto corresponding RI images can standardize the measurement of BPE.

Description

Breast cancer is the second leading cause of cancer-related death in women. Digital mammography, often supplemented by ultrasound, is commonly used for screening, but with some limitations. Magnetic resonance imaging (MRI) is required for high- or intermediate-risk cases and in individuals with dense breast tissue, but MRI is costly and not suitable for every patient. CEM is an emerging technique using contrast agents to provide visualization similar to MRI while being less expensive and more accessible. BPE is associated with increased breast cancer risk and reported subjectively (i.e., mild, moderate, marked) following CEM investigation. This novel method is designed to automate quantitative BPE measurements, making reporting easier and more consistent.

Applications

• Breast cancer screening programs
• Breast cancer risk prediction research

Advantages

Currently, BPE is measured subjectively by radiologists, with research indicating a lack of consistency in intra- and inter-radiologist measurements. Additionally, only qualitative BPE assessments are made (i.e., mild, moderate, marked), meaning granular data that could be used for developing breast cancer risk-prediction tools are missed.

This novel CEM method will quantify BPE assessments using an AI-based assessment. The approach will overcome the challenges posed by the subjectivity of current methods, thereby eliminating disagreement among radiologists. This method would result in a standardized BPE assessment, which could eventually be developed into a breast cancer risk-prediction tool. Finally, this novel method will streamline radiologists' workflows, improving access to CEM for more patients.

Invention Readiness

Using CEM images from over 1000 patients, a radiologist-trained AI model segmented fibroglandular and fatty tissue from low-energy images. Using recombined images, AI-dBPE (the difference between fibroglandular and fatty tissue BPE) demonstrated true tissue contrast. Strong correlations were found between AI-dBPE and radiologist-reported BPE (r=0.69, 95% CI: 0.66-0.71), and in women <45 years (r=0.68, 95% CI: 0.62-0.73). AI-dBPE correlated inversely with age, suggesting AI-dBPE may be a suitable, quantifiable measure of BPE classification.

IP Status

Patent Pending

Related Publication(s)

Nishikawa, R. M., & Lu, A. H. (2023). AI in Screening Mammography: Use One Radiologist and Improve Double Reads. Radiology, 309(2). https://doi.org/10.1148/radiol.232964