Clinical image quality evaluation of mammography for breast cancer screening

Abstract

Objective: To evaluate mammographic image quality in a real clinical practice scenario for breast cancer screening.
Methods: This was an observational, prospective study that analyzed images from digital mammograms from diagnostic services in the state of Goiás in 2019. A specific protocol was created based on the evaluation criteria of the Brazilian
College of Radiology, the European Guidelines, and the American College of Radiology. For each variable score, one was
attributed to conformity, and zero to nonconformity. Logistic regression model was utilized and the following independent
variables were considered: location (city vs. country); the Brazilian Unified Health System (SUS) (public vs. private); number
of monthly exams (≤300 vs. >300); device manufacturing year (≤2011 vs. >2011); and breast density (≤75% vs. >75% of the
parenchyma). Results: Of 163 fully functioning mammograms, 151 (92.6%) were eligible, with 53 (32.5%) equipment participating in the research, which produced a total of 1,024 images. In the clinical image analysis, regarding the positioning
of the patient, it was observed the higher conformity for symmetry parameters, in both projections (>90%). The conformity
rate amongst the other parameters varied from 18.6% to 100%. In the multivariable analysis, it was observed that only the
variable monthly exams (odds ratio [OR] 3.44; 95% confidence interval [CI] 1.67–7.09; p=0.0008) and mammogram device
manufacturing year (OR 2.46; 1.02–5.95; p=0.04) were associated with a higher conformity rate. After the percentage consolidation conformity rate per diagnostic service, regarding the final clinical mammography quality, no diagnostic service presented desirable conformity (>90%), 28 obtained acceptable conformity (between 70 and 89%), and 25 presented
conformity below 70%. Conclusion: Conformity rate of mammographic exams is extremely low and varies accordingly to
the multiple parameters analyzed. Mammographies performed at centers with less productivity (≤300/month) and with
newer devices (>2011) presented higher chances of conformity at the clinical imaging evaluation.