Blinded review of retrospectively visible unreported breast cancers: An eye-position analysis

PURPOSE: To determine whether unreported retrospectively identified cancers on mammograms receive prolonged visual attention and can be reliably detec ted in a blinded review. MATERIALS AND METHODS: Four experienced mammographers performed a blinded r eview of a test set of 20 retrospective cases where the cancer was not dete cted until the next mammographic evaluation, 10 prospective cases where the cancer was initially detected, and 10 cancer-free cases. Two views were di gitized and displayed on a workstation. The experiment consisted of an init ial impression, during which eye position was monitored, and a final impres sion, during which viewers zoomed on regions of interest and localized susp icious lesions. Eye-position data were analyzed to determine whether retros pectively visible cancers attracted attention to the same degree as prospec tively visible cancers. The initial impression used 1,000 msec as the eye-f ixation dwell criterion for detecting a lesion. RESULTS: Initially, 70% of retrospective cancers and 50% of prospective can cers did not attract prolonged visual attention. In prospective cases, deta iled examination significantly improved the mean receiver operating charact eristic area, from .73 to .88 (P <.01) but in retrospective cases the mean receiver operating characteristic area barely increased, from .60 to .68, d ue to a high true-positive-to-false-positive ratio. CONCLUSION: At blinded review, detection of retrospectively visible cancers was significantly inferior to that of prospective cancers. It cannot be as sumed that retrospectively identified cancers are intrinsically detectable, because they do not draw prolonged visual attention during visual search f or breast cancers.