Feasibility study of low-dose CT colonography using model-based iterative reconstruction: preliminary findings in 20 patients

Abstract

Purpose: To compare image quality on computed tomographic (CT) images acquired at standard dose (SD) and low dose (LD) using adaptive statistical iterative reconstruction (ASIR) and model-based iterative reconstruction (MBIR) techniques. Methods and Materials: 20 patients were prospectively recruited for the study with informed consent. They underwent standard dose (SD) and low-dose (LD) CT colonography. Low-dose parameters were based on our phantom study with using vendor-specific automatic tube current modulation known as noise index. Both sets of scans were reconstructed with ASIR and MBIR. Objective and subjective image qualities were assessed as well as diagnostic accuracies for significant lesions (e.g. polyps, cancer, etc). Effective doses for each scans were recorded. Results: Objective image analysis supports significant noise reduction and superior contrast-to-noise ratio with low-dose scans using MBIR technique (p < 0.05) despite being acquired at lower doses. Subjective image parameters were equivalent for LD MBIR and SD ASIR for both colonic and extra-colonic findings. Diagnostic accuracies for polyp detection and other significant lesions were comparable. Dose recorded were substantially lower for MBIR (range 45-75% reduction compared with ASIR) with mean average being 4.2 mSv in our population group. Conclusion: MBIR shows superior reduction in noise whilst maintaining image quality and most importantly substantial dose reduction can be achieved. More patients are being recruited to substantiate diagnostic accuracies data with full results to be presented at the ECR 2013.