In their article titled “Time Course of Perception and Decision Making during Mammographic Interpretation,” Nodine et al. seek to compare and contrast the time it takes to accurately detect breast lesions between radiology trainees and experienced mammographers. Previous research found that prolonged decision time led to less accurate interpretations of scans, especially within the population of trainees who had not yet had much experience in reading these scans (Nodine et al.
, 917). The authors of this article sought to find if this same conclusion could be applied to the reading of mammograms. To do this, they used eye tracking technology and computer mouse clicks to determine the speed at which the subjects noticed and accurately identified mammogram scans as “abnormal” or “normal.” In addition, is there a specific decision time at which performance begins to decline when looking at these scans? To answer this question, 40 cases, in which half contained malignant lesion(s) and the other half were healthy, normal breasts, were shown to six trainees and three experienced mammographers (Nodine et al.
, 918). Each subject participated in two phases: during the initial-decision phase, instructions to evaluate each case and determine if the scans were normal or abnormal were given to each subject. They were then asked to give their level of confidence that their decision was correct using a cursor on a computer screen. An “eye-head tracker” was used to follow eye position when looking at each scan during this phase and each case was timed using a computer clock (Nodine et al., 918-919).
In the final-decision phase, subjects were expected to identify the lesion or skip to the next case if no abnormality was observed. Performance was found to decline after a 40-sec decision time, after which the true-positive rate began to slow down for both groups. As expected, experienced mammographers were able to make true-positive decisions with higher certainty than trainees. Eye-position data showed that experienced mammographers directed their attention on more true lesions than trainees did and failed to do so on less true lesions than trainees. Nodine et al.
found that the experienced subjects could identify true abnormalities “four times faster” than trainees within the first 25-50 seconds of looking at a case; after this time, it is recommended that the search for abnormalities be discontinued, especially if the individual looking at the scan believes they cannot make a “high confidence decision” as continuing to search is less likely to result in a true-positive finding after this time (Nodine et al., 921-923). These findings could have implications for other scans as well. Magnetic Resonance Imaging (MRI) scans are used in the diagnosis of Multiple Sclerosis (MS) patients, in addition to spinal tap and evoked potential procedures (National Multiple Sclerosis Society). While MRIs will not always provide a definite answer, they can be helpful in putting together a well informed and accurate diagnosis. If an MRI technologist or trainee were to read a scan in search for lesions, it would be useful to know how long it would take them to find a lesion, if any, and when to end their search if a confident decision cannot be made. Reference:Diagnosing MS.
(n.d.). Retrieved January 29, 2018, from https://www.
nationalmssociety.org/Symptoms-Diagnosis/Diagnosing-MSNodine, C. F., Mello-Thoms, C., Kundel, H.
L., & Weinstein, S. P. (2002). Time Course of Perception and Decision Making During Mammographic Interpretation.American Journal of Roentgenology,179(4), 917-923.