Photo Credit: Yacobchuk

For analyzing negative screening digital breast tomosynthesis (DBT) examinations, patient characteristics influence the case and risk scores of an artificial intelligence (AI) algorithm, according to a study published online May 21 in Radiology.

Derek L. Nguyen, M.D., from the Duke University School of Medicine in Durham, North Carolina, and colleagues examined the impact of patient characteristics on performance of an AI algorithm interpreting negative screening DBT examinations. The retrospective cohort study identified negative screening DBT examinations, all of which had two years of follow-up without a diagnosis of atypia or breast malignancy. The DBT studies were interpreted by a U.S. Food and Drug Administration-approved AI algorithm, which generated case scores and risk scores for each mammogram.

Data were included for 4,855 patients. The researchers found that false-positive case scores were significantly more likely in Black than White patients (odds ratio [OR], 1.5) and were significantly less likely in Asian patients (OR, 0.7). Compared with patients aged 51 to 60 years, false-positive case scores were more likely in patients aged 71 to 80 years (OR, 1.9) and less likely in those aged 41 to 50 years (OR, 0.6). False-positive risk scores were more likely in Black versus White patients, in those aged 61 to 70 versus 51 to 60 years, and in those with extremely dense versus fatty-density breasts.

“The Food and Drug Administration should provide clear guidance on the demographic characteristics of samples used to develop algorithms, and vendors should be transparent about how their algorithms were developed,” the authors write. “Continued efforts to train future AI algorithms on diverse data sets are needed to ensure standard performance across all patient populations.”

Several authors disclosed ties to the medical device and technology industries.

Abstract/Full Text

Copyright © 2024 HealthDay. All rights reserved.