The University of Oklahoma and Mercy Hospital Oklahoma City are developing new short-term breast cancer risk prediction models that aim to help increase cancer detection of breast magnetic resonance imaging screening. With a $2.5 million grant from the National Institutes of Health/National Cancer Institute, OU and Mercy will develop and apply two new short-term breast cancer risk prediction approaches: a rule-in approach for identifying women who are excluded from current breast MRI screening guidelines but have a higher risk of developing or harboring mammography-occult (hidden) breast cancers, which can be detected by breast MRI; and a rule-out approach for women with an elevated lifetime cancer risk but no imminent risk.
Collaborators are Bin Zheng and Hong Liu, professors in the School of Electrical and Computer Engineering, Gallogly College of Engineering, and affiliates of the newly established School of Biomedical Engineering and the Stephenson Cancer Center on the OU Health Sciences Center campus, and Dr. Alan Hollingsworth, medical director of the Mercy Breast Center. They will analyze image features related to bilateral asymmetry of mammographic tissue density and/or MRI tissue enhancement signals of the left and right breasts of the same woman, which is highly correlated to the biological process of cancer development, to develop the rule-in and rule-out risk prediction models.
"The goal is to significantly increase cancer detection of breast MRI screenings by combining these two new risk assessment approaches," says Zheng.
Although breast MRI is the most sensitive imaging modality in detecting breast cancer at an early stage when used for asymptomatic screening, it is currently approved only for a small group of women with significantly higher risk of developing breast cancer in their lifetime as compared to the general population. Even in this high-risk population, the cancer detection yield of breast MRI screening is low, around 3 percent.
OU researchers will assemble a large and diverse digital mammogram and breast MRI image database with sequential screening examinations and verified diagnostic results provided by Hollingsworth at Mercy. Nearly 10,000 images will be studied over a five-year period, many of them requiring additional confirmation of normalcy, prompting review by dedicated breast radiologists at Mercy, Dr. Rebecca Stough and Dr. Melanie Pearce. The research team will develop and optimize new short-term models to predict the risk of a woman having mammography-occult early breast cancer after a negative mammography screening.
OU researchers will use the rule-in method to identify women with high-risk mammography-occult cancer from the general mammography screening population. A prospective clinical study will be performed at the Mercy Breast Center that tracks a minimum of 4,000 negative screening mammography cases to validate performance of the new risk model by identifying women with mammography-occult breast cancer that can be found on MRI. Using the rule-out process, researchers aim to reduce the repeated negative MRI screenings among the currently identified women with elevated risk.
The research team will compare the new image feature based risk models with other existing breast cancer risk prediction models in assessing the short-term breast cancer risk. The development of a new feature fusion model will improve risk prediction performance. Finally, OU researchers will analyze data from the entire case pool, including subgroups of cases based on different classification criteria.
"The synergy of our two programs is unique," said Hollingsworth, a nationally recognized expert in breast cancer risk assessment and the epidemiology of screening with breast MRI. "Great discoveries at the basic science level can languish without a clinical counterpart to move the innovation into clinical trials. Professors Zheng and Liu have developed an image-based approach that could revolutionize breast cancer screening if this multi-pronged study is successful."
"The leadership and support from the University of Oklahoma and the Gallogly College of Engineering at OU and their vision of building a critical mass in medical imaging, the seamless collaboration between OU's Norman campus and the Stephenson Cancer Center on the OU Health Sciences Center campus was essential to our progress," states Liu.