OBJECTIVES: The objective of this study was to compare image quality on abdominal/pelvic computed tomographic images acquired with filtered back projection (FBP), adaptive statistical iterative reconstruction (ASIR), and novel model-based iterative reconstruction (MBIR) techniques with varying levels of automatic tube current modulation and tube voltages. METHODS: A 2-phase study was performed. In phase 1, a torso phantom was scanned at 17 different noise levels of automatic current modulation (selected using noise index [NI]) at 120 kilovolt (peak) (kVp). Images reconstructed with FBP, ASIR, and MBIR underwent objective analysis. In phase 2, additional scans were performed at 3 different kVp (80, 100, and 120 kVp at 3 different NIs (33, 50, and 70). Objective and subjective image qualities were assessed. Computed tomography dose index and dose-length products were recorded. RESULTS: The objective image analysis supports significant noise reduction with MBIR compared with ASIR and FBP (P < 0.05) at all 17 NI tested at 120 kVp. When lowering the kVp, the subjective image quality was improved, but when this is performed in conjunction with increasing NI, image quality was maintained only at moderately high NI of 50 but was degraded at higher NIs despite improving contrast-to-noise ratio. CONCLUSIONS: Our results represent the first exploration in the utility of MBIR technique with alteration of kVp in conjunction with tube current modulation in comparison with traditional methods. Objective image noise for MBIR is superior. Subjective image quality is only moderately improved. Scanning at low kVp and moderately high NI with MBIR can ensure that a balance of improved image noise and contrast can be achieved as well as reducing dose.
OBJECTIVES: The objective of this study was to compare image quality on abdominal/pelvic computed tomographic images acquired with filtered back projection (FBP), adaptive statistical iterative reconstruction (ASIR), and novel model-based iterative reconstruction (MBIR) techniques with varying levels of automatic tube current modulation and tube voltages. METHODS: A 2-phase study was performed. In phase 1, a torso phantom was scanned at 17 different noise levels of automatic current modulation (selected using noise index [NI]) at 120 kilovolt (peak) (kVp). Images reconstructed with FBP, ASIR, and MBIR underwent objective analysis. In phase 2, additional scans were performed at 3 different kVp (80, 100, and 120 kVp at 3 different NIs (33, 50, and 70). Objective and subjective image qualities were assessed. Computed tomography dose index and dose-length products were recorded. RESULTS: The objective image analysis supports significant noise reduction with MBIR compared with ASIR and FBP (P < 0.05) at all 17 NI tested at 120 kVp. When lowering the kVp, the subjective image quality was improved, but when this is performed in conjunction with increasing NI, image quality was maintained only at moderately high NI of 50 but was degraded at higher NIs despite improving contrast-to-noise ratio. CONCLUSIONS: Our results represent the first exploration in the utility of MBIR technique with alteration of kVp in conjunction with tube current modulation in comparison with traditional methods. Objective image noise for MBIR is superior. Subjective image quality is only moderately improved. Scanning at low kVp and moderately high NI with MBIR can ensure that a balance of improved image noise and contrast can be achieved as well as reducing dose.