Now you see it: Tracking eye gaze during CTC interpretation

Using a desk-mounted tracker to follow the gaze of radiologists as they interpreted 3D CT colonography (CTC) videos revealed that other than time to first pursuit, visual search pattern metrics don’t vary significantly between experienced and inexperienced readers, according to a group of U.K. researchers.

The results of the study, published online July 15 in Radiology, suggested that differences in effectiveness between experienced and inexperienced radiologists is the result of something other than simply spotting a lesion.

“We found that readers almost always examined polyps by a series of multiple pursuits, which suggested that readers recognize a lesion as visually important in most observations, and the errors are because of decision,” wrote Susan Mallett, DPhil, of the University of Oxford, and colleagues.

A total of 65 radiologists were included in the study, 27 of whom were experienced and the rest inexperienced. While their gaze was being tracked, they interpreted 23 endoluminal 3D CTC videos that were each 30 seconds long, and indicated when they saw a polyp by clicking a computer mouse.

Time to first pursuit was significantly shorter in experienced readers, with a hazard ratio of 1.22, however Mallett and colleagues indicated that the absolute time difference was small and it wasn’t clear how this would affect polyp identification. Other metrics such as assessment pursuit time and identification period were similar between experienced and inexperienced readers.

Inexperienced readers found a lower percentage of lesions than the experts—67 percent vs. 75 percent—but since there was a similar percentage of lesions with eye pursuits, the authors concluded that errors during interpretation of 3D CTC occurred either during the discovery or recognition phases of the interpretations, and were not related to problems with visual scanning technique.

Other findings of note included the fact that the longer polyps were visible, the longer identification was delayed. Readers spent about 40 percent of the time pursuing polyps and 60 percent looking at other areas of the video.

“Knowledge of the deficiencies in interpretation by inexperienced readers could improve the focus and structure of training,” wrote Mallett and colleagues.