Assessing satisfaction of search in virtual mammograms for experienced and novice searchers

Abstract

Purpose: Satisfaction of search (SOS) is a phenomenon where searchers are more likely to miss a lesion/target after detecting a first lesion/target. Here, we investigated SOS for masses and calcifications in virtual mammograms with experienced and novice searchers to determine the extent to which: (1) SOS affects breast lesion detection, (2) similarity between lesions impacts detection, and (3) experience impacts SOS rates.

Approach: The open virtual clinical trials framework was used to simulate the breast anatomy of patients, and up to two simulated masses and/or single-calcifications were inserted into the breast models. Experienced searchers (residents, fellows, and radiologists with breast imaging experience) and novice searchers (undergraduates who had no breast imaging experience) were instructed to search for up to two lesions (masses and calcifications) per image.

Results: $2\times 2$ mixed factors analysis of variances (ANOVAs) were run with: (1) single versus second lesion hit rates, (2) similar versus dissimilar second-lesion hit rates, and (3) similar versus dissimilar second-lesion response times as within-subject factors and experience as the between subject's factor. The ANOVAs demonstrated that: (1) experienced and novice searchers made a significant amount of SOS errors, (2) similarity had little impact on experienced searchers, but novice searchers were more likely to miss a dissimilar second lesion compared to when it was similar to a detected first lesion, (3) experienced and novice searchers were faster at finding similar compared to dissimilar second lesions.

Conclusions: We demonstrated that SOS is a significant cause of lesion misses in virtual mammograms and that reader experience impacts detection rates for similar compared to dissimilar abnormalities. These results suggest that experience may impact strategy and/or recognition with theoretical implications for determining why SOS occurs.

Keywords: attentional template; breast imaging; satisfaction of search; subsequent search misses; visual search.

Fig. 1

(Previously published in Ref. 17): (a) attentional template theory tenets. An attentional template is known in cognitive science as an active representation of the search goal held within visual working memory. The attentional template theory predicts that the maintenance of a detected target as a searchers’ attentional template is why SOS errors occur. The tenets of the attentional template theory were motivated by the SOS error literature and the broader visual working memory, attention, and visual search literature. (b) Time course of the fluctuation of working memory and attentional resources in a multiple target search. The battery illustration depicts the capacity limitations of visual working memory and attention. The double-sided arrows between the batteries illustrate the flow of a shared cognitive resource that underlies both visual working memory and attention, and that when one cognitive process is prioritized, the remaining cognitive process is hindered. After first target detection, visual working memory resources are prioritized to maintain the first target as an attentional template. Consequently, fewer attentional resources are available, which will decrease the probability of additional target detection. After detecting a first target, the detection of similar targets will improve compared to dissimilar targets because a first target attentional template will prime attention toward similar targets. Furthermore, the detection of similar, second targets will be faster than dissimilar second targets. Over time the first target may lose its prioritization as the attentional template, which will free up visual working memory resulting in improved and unbiased detection for similar and dissimilar targets.

Fig. 2

(a) Sample virtual breast mammogram. (b) Graph of single and second lesion hit rate (i.e., to determine whether there was an SOS effect). (c) Graph of second lesion hit rates when they were similar and dissimilar to a detected first lesion (i.e., to determine whether there was a similarity effect).

Fig. 3

Predicted values of total time searching from the full model. The “red” color represents the 1st trial and the “blue” color represents the last trial (i.e., trial 52). The shaded color around the lines represents 95% confidence intervals. Notice the difference in the mean value for trial 1 compared to trial 52 for novice searchers compared with experienced searchers—novice searchers have a steeper learning curve, as would be predicted given that this was their first time searching breast images.