Prediction Model Reduces Inaccuracies in MRI Breast Cancer Screenings

According to a study from the Radiological Society of North America, prediction models based on clinical characteristics and imaging findings could reduce the number of false positives in MRI breast cancer screenings.

Researchers have found that women with dense breast tissue have a higher risk of breast cancer than those with average breast density. Additionally, higher breast density significantly reduces the sensitivity of mammography, allowing cancer to hide within dense breast tissue.

For women with dense breast tissue, breast MRIs are considered a better alternative to mammography screening. MRIs are the most sensitive imaging method for diagnosing breast cancer and can differentiate accurately between lesions and abnormalities of the breast.

However, the high sensitivity of the imaging tool also means MRI will often detect benign lesions that would otherwise go unnoticed. Women who get contacted for additional workups due to these findings potentially face repeat MRI scans, targeted ultrasound, and biopsy. The need for the additional examination can cause patients anxiety, increase healthcare costs, and lead to biopsy-related complications.

“The reduction of the false-positive recall rate is an important issue when considering the use of breast MRI as a screening tool,” said study lead author Bianca M. den Dekker, MD, in a press release.

In the study, researchers developed prediction models to distinguish true-positive MRI screenings from false positives. In creating the models, the team combined MRI findings with clinical characteristics, including body mass index, family history of breast cancer, and the use of hormone replacement therapy.

Researchers used data from the Dense Tissue and Early Breast Neoplasm Screening (DENSE) trial, which evaluated the effectiveness of screening with mammography and MRI versus mammography alone in participants aged 50 to 75.

Of the 454 women with positive MRI results in a first supplemental round of screenings, 79 were diagnosed with breast cancer, indicating that 375 women had false-positive results. According to the researchers, the full prediction model could have prevented 45.5 percent false-positive recalls and 21.3 percent benign biopsies.

The model based entirely on readily accessible MRI findings and age had a similar performance, potentially preventing 35.5 percent of false-positive MRI screenings and 13 percent of benign biopsies.

“Our prediction models may identify a substantial number of false positives after first-round supplemental MRI screenings, reducing false-positive recalls and benign biopsies without missing any cancers,” den Dekker said. “This brings supplemental screening MRI for women with dense breasts one step closer to implementation.”

The researchers plan to perform validation studies using data from different populations and to study the performance of prediction models in future screening rounds. Den Dekker added that with the model, the false-positive rate in the study group fell from 78.9 per 1000 screenings in the first round to 26.3 per 1000 in the second.  

“This can be partly explained by the availability of prior MRI examinations, which allows comparison for interval change,” she said. “As incident screening rounds have a much lower false-positive rate, separate models may have to be created.”