Quantitative evaluation of visual detection performance in medicine: ROC analysis and determination of diagnostic benefit. [Radiographic image evaluation]
An ROC curve provides an empirical description of the trade-offs which are possible among the various types of correct and incorrect decisions as the human decision-maker varies one or more confidence thresholds. Conventional ROC curves measured in simple decision-making situations can, in some cases, be used to predict human decision performance in more complex situations. By considering both the consequences of the various types of diagnostic decisions and the overhead cost of a diagnostic study, one can use the ROC curve to evaluate the diagnostic usefulness of a study in any particular clinical context. Since the ROC curve describes the possible relationships among the probabilities of the various types of correct and incorrect decisions, it plays a central role in optimizing diagnostic strategies using the general techniques of decision analysis. Applications in radiographic image evaluation are described. (CH)
- Research Organization:
- Chicago Univ., Ill. (USA). Dept. of Radiology; Franklin McLean Memorial Research Inst., Chicago, Ill. (USA)
- OSTI ID:
- 7360140
- Report Number(s):
- FMI-1000-766; CONF-760452-1
- Country of Publication:
- United States
- Language:
- English
Similar Records
Applications of ROC analysis in diagnostic image evaluation
Quantitative evaluation of human visual detection performance using empirical receiver operating characteristic curves