Using computer-aided detection software to help find target areas in images may result in readers missing areas not marked for viewing.
Using computer-aided detection (CAD) software to help find target areas in images may actually result in readers missing areas not marked for viewing, according to research published in the October issue of Academic Radiology.
Researchers performed two experiments as part of an eye-tracking study to determine eye movements as images were being read. Forty-seven observers with an average age of 24.3 years examined images for the targets, which were the letter “T.” Throughout the image, there were also several marks of the letter “L.”
The first experiment provided high noise and low similarity between the targets, meant to cause distraction and making the targets more difficult to locate. The second experiment used less distraction and provided more targets, making the targets easier to find.
Researchers found that although in the first experiment, observers were able to find many of the targets, guided by the CAD, this came at a cost: If targets were not in an area marked by the CAD, they were likely not located. For the second experiment, there was no noted behavioral benefit from CAD, but neither was there a cost in missed targets as there was in the first experiment.
In evaluating eye movements, the researchers saw that the CAD observers looked at a lower total area than did the non-CAD observers.
“CAD signals do not combine with observers' unaided performance in a straightforward manner,” concluded the authors. “CAD can engender a sense of certainty that can lead to incomplete search and elevated chances of missing unmarked stimuli.”
MRI-Based AI Radiomics Model Offers 'Robust' Prediction of Perineural Invasion in Prostate Cancer
July 26th 2024A model that combines MRI-based deep learning radiomics and clinical factors demonstrated an 84.8 percent ROC AUC and a 92.6 percent precision-recall AUC for predicting perineural invasion in prostate cancer cases.
Breast MRI Study Examines Common Factors with False Negatives and False Positives
July 24th 2024The absence of ipsilateral breast hypervascularity is three times more likely to be associated with false-negative findings on breast MRI and non-mass enhancement lesions have a 4.5-fold likelihood of being linked to false-positive results, according to new research.
Can Polyenergetic Reconstruction Help Resolve Streak Artifacts in Photon Counting CT?
July 22nd 2024New research looking at photon-counting computed tomography (PCCT) demonstrated significantly reduced variation and tracheal air density attenuation with polyenergetic reconstruction in contrast to monoenergetic reconstruction on chest CT.