Affiliations: Department of Medicine, University of Ottawa, Ottawa,
Canada | Department of Mechanical and Industrial Engineering,
Ryerson University, Toronto, Canada
Abstract: In this paper, we study the reliability of X-ray imaging diagnosis
considering human cognitive abilities (e.g., spatial orientation,
visualization, line orientation, and perceptual speed), which play a vital role
in the clinical decision making that requires classification systems. Also,
this study explores sex influence on X-ray imaging diagnosis based on 176 X-ray
images evaluated by 10 female radiologists and 8 male radiologists. Most
related literature focuses on a binary classification (True or False) that uses
a set of features derived from a given pattern. Also, they utilize the Receiver
Operating Characteristics (ROC) analyses for assessing the accuracy of X-ray
diagnosis. In this study, we use fuzzy benchmarking to construct fuzzy
classification systems required for fuzzy medical decision-making and fuzzy
reliability assessment. The proposed method differentiates the influence of
human cognitive abilities and sex in X-ray diagnosis. The results from this
study shows reliability of X-ray diagnosis is high and male radiologists excel
in spatial and line orientation and female radiologists perform better in
perceptual speed while both are competent in visualization ability.
Keywords: X-ray diagnosis, diagnosis reliability, radiology, human cognitive, decision-making, ROC, human errors
