Detection of simulated pulmonary nodules by single-exposure dual-energy computed radiography of the chest: effect of a computer-aided diagnosis system (Part 2)

Abstract 

Objective: To evaluate the performance of the computer-aided diagnosis (CAD) scheme on the detection of pulmonary nodules (PNs) in single-exposure dual-energy subtraction computed radiography (CR) images of the chest, and to evaluate the effect of this CAD scheme on radiologists' detectabilities. Methods and material: We compared the detectability by the CAD scheme with the detectability by 12 observers by using conventional CR (C-CR) and bone-subtracted CR (BS-CR) images of 25 chest phantoms with a low-contrast nylon nodule. Results: Both in the CAD scheme and for the observers, the detectability of BS-CR images was superior to that of C-CR images (P<0.005). The detection performance of the CAD scheme was equal to that of the observers. The nodules detected by the CAD did not necessarily coincide with those by the observers. Thus, if observers can use the results of the CAD system as a ‘second opinion’, their detectabilities increase. Conclusion: The CAD system for detection of PNs in the single-exposure dual-energy subtraction method is promising for improving radiologists' detectabilities of PNs.

Keywords:  Dual-energy subtraction, Computed radiography, Computer-aided diagnosis (CAD), Pulmonary nodule