Abstract
The purpose of this paper is to introduce new methodology for the evaluation of diagnostic imaging systems based on the assessment of diagnostic features by human observers. This methodology is intended to supplement the usual receiver operating characteristic (ROC) curve approach to imaging system evaluation. The approach is based on measuring the experimental evidence of a priori scientific propositions. The methodology requires simple calculations that can be implemented by clinical inves tigators. An example, based on diagnosis of sacroiliitis, is given to illustrate the method's usefulness in a diagnostic situation. (Med Decis Making 2:311-321, 1982)
Get full access to this article
View all access options for this article.
