New research suggests that access to prior mammography exam findings may enhance specificity rates and reduce false positives for current mammography assessments without a significant impact on sensitivity.
For the study, recently published in Clinical Radiology, researchers compared mammography assessment with the availability of two to four years of prior mammograms in an initial reading session to mammography interpretation without prior mammography access in a second reading session six months later. Eight readers (including six radiologists and a senior radiology trainee) participated in the reading sessions of 72 screening mammograms, according to the study.
Overall, the researchers found that access to prior mammograms bolstered the specificity rate to 91.1 percent in comparison to 76.3 percent without access to a patient’s prior mammography results. The study authors also noted that access to prior mammograms resulted in a significantly lower mean false positive rate (3.6 percent vs. 9.5 percent) a significantly higher mean percentage of true negative assessments (36.4 percent vs. 30.5 percent) in contrast to no access to prior exams.
“The findings demonstrate that access to mammograms from the previous screening round improves radiologists’ ability to correctly identify women who have no signs of cancer on their images without affecting the detection and classification (of) breast cancer lesions in women who have cancer,” wrote lead study author Judith D. Akwo, M.D., who is affiliated with the Medical Image Optimization and Perception Group with the Sydney School of Health Sciences at the University of Sydney in Camperdown, Australia, and colleagues.
Even after considering potential confounding factors such as breast density as well as clinician experience in reading mammograms, the researchers found that specificity rates consistently increased with access to prior mammography exams.
For example, access to prior mammography exams for patients with dense breasts led to a 15.6 percent increase in specificity (92.5 percent vs. 76.9 percent). For patients with non-dense breasts, access to prior mammography exams led to a 16.9 percent increase with specificity rates (92.5 percent vs. 75.6 percent). The researchers also noted that exposure to prior mammography led to a 15.7 percent increase in specificity for radiologists who spend four hours or less a week interpreting mammograms and a 13.8 percent increase in specificity for those who spend more than four hours a week reading mammograms.
“These findings highlight the importance of prior mammograms to all radiologists regardless of their years of experience and workload-related characteristics, and women of all breast densities, and we consider this the first time such an effect has been shown,” explained Akwo and colleagues.
Three Key Takeaways
1. Access to prior mammograms enhances specificity. Having access to prior mammography exams significantly improves the specificity of current mammography assessments, increasing it from 76.3 percent to 91.1 percent. This helps in more accurately identifying women without cancer.
2. Reduction in false positives. Access to prior exams leads to fewer false positives (3.6 percent vs. 9.5 percent) and increases the rate of true negative assessments (36.4 percent vs. 30.5 percent), improving diagnostic confidence.
3. Impact across varying breast density and radiologist experience levels. The benefit of access to prior mammograms extends across varying breast densities and radiologists with different experience levels, improving specificity consistently for both dense and non-dense breasts and among radiologists with varied workload characteristics.
The study authors also noted no significant overall difference with respect to sensitivity rates, pointing out 67.5 percent sensitivity with access to priors versus 71.3 percent without priors. However, the researchers pointed out the receiver operating characteristic (ROC) curve and jackknife alternative free-response ROC (JAFROC) revealed slightly higher reader capabilities for lesion detection and ascertaining malignancy level with exposure to prior mammography exams.
“… Unlike perturbations from normal breast parenchyma or benign lesions, the perturbations due to breast cancer gradually change over time as the cancer grows or invades surrounding tissues. Therefore, comparison with prior mammograms should help radiologists detect changes in mammographic features that are most discriminative of breast cancer,” emphasized Akwo and colleagues.
(Editor’s note: For related content, see “Comparing Digital Breast Tomosynthesis to Digital Mammography: What a Long-Term Study Reveals,” “Mammography Study: Can Stand-Alone AI Enhance Detection of Interval Breast Cancer?” and “Predicting DCIS Upgrade to Invasive Breast Cancer: Can Contrast-Enhanced Ultrasound Have an Impact?”)
In regard to study limitations, the authors acknowledged that the small sample size and relying on eight readers for the analysis may limit extrapolation of the study findings for a broader population.
