Detection of low-contrast objects: experimental comparison of single- and multi-detector row CT with a phantom.

Abstract

To compare single-detector row computed tomography (CT) and multi-detector row CT by using an experimental phantom containing a contrast-detail modulus for detection of low-contrast structures to optimize acquisition protocols. The parameters milliampere seconds, reconstructed section thickness, and elementary collimation for multi-detector CT were varied for two pitches with single- and multi-detector CT. For objective assessment of image quality, contrast-to-noise ratio (CNR) was calculated for an 8-HU low-contrast 15-mm-diameter object. Subjective assessment of image quality was performed by means of visual detection of low-contrast objects of various sizes by four independent observers. For each acquisition protocol, the effective doses required to obtain the CNR thresholds allowing 100% detection of 5-, 7-, and 9-mm-diameter objects were compared at single- and multi-detector CT at comparable section sensitivity profile with analysis of variance. Significant correlation was found between CNR measurements and subjective object detection (r = 0.95, P <.05). CNRs of 1.0, 0.8, and 0.6 were required to detect 100% of the 5-, 7-, and 9-mm-diameter objects, respectively. For reconstructed section thickness of 5-10 mm, comparable x-ray doses were required with single- and multi-detector CT to detect objects. For reconstructed sections thinner than 5 mm, single- and multi-detector CT allowed detection of only the 7- and 9-mm-thick objects, but a higher x-ray dose was required for multi- than for single-detector CT (P <.05). Multi-detector CT is less effective than single-detector CT in detection of small low-contrast objects if sections thinner than 5 mm are used. Results for single- and multi-detector CT were similar for sections 5 mm or thicker.