CAD improves polyp detection
CT colonography exams are difficult and time consuming to read and interpret, a potential barrier to more widespread colon cancer screening. Computer-aided detection has a potential role to play in that effort, according to a study presented at the
CT colonography exams are difficult and time consuming to read and interpret, a potential barrier to more widespread colon cancer screening. Computer-aided detection has a potential role to play in that effort, according to a study presented at the American Roentgen Ray Society meeting in May.
Dr. Abraham H. Dachman and colleagues at the University of Chicago retrospectively studied 20 CT colonography data sets: 10 sets consisted of 11 polyps, 5 to 12 mm in size, and 10 were normal. Four readers (two attending radiologists, a radiology resident, and a gastroenterologist) reviewed the data sets with and without computer-aided detection.
Readers missed 42% of the polyps without CAD; they later identified 75% of the missed polyps with the help of the CAD system. CAD detected all the polyps missed by the readers, while readers correctly dismissed 77% of the computer's false positives.
The average area under the ROC curve (Az values) for radiologists without and with CAD were 0.70 and 0.85, respectively. The difference was statistically significant. The increase in the Az value was the largest for the gastroenterologist (0.21) among the four observers.
Considering Breast- and Lesion-Level Assessments with Mammography AI: What New Research Reveals
June 27th 2025While there was a decline of AUC for mammography AI software from breast-level assessments to lesion-level evaluation, the authors of a new study, involving 1,200 women, found that AI offered over a seven percent higher AUC for lesion-level interpretation in comparison to unassisted expert readers.
