Comparison of Model-Based Iterative Reconstruction, Adaptive Statistical Iterative Reconstruction, and Filtered Back Projection for Detecting Hepatic Metastases on Submillisievert Low-Dose Computed Tomography

The aim of the study was to compare the diagnostic performance of model-based iterative reconstruction (MBIR), adaptive statistical iterative reconstruction (ASIR), and filtered back projection (FBP) on submillisievert low-dose computed tomography (LDCT) for detecting hepatic metastases. Thirty-eight patients having hepatic metastases underwent abdomen CT. Computed tomography protocol consisted of routine standard-dose portal venous phase scan (120 kVp) and 90-second delayed low-dose scan (80 kVp). The LDCT images were reconstructed with FBP, ASIR, and MBIR, respectively. Two readers recorded the number of hepatic metastases on each image set. A total of 105 metastatic lesions were analyzed. For reader 1, sensitivity for detecting metastases was stationary between FBP (49%) and ASIR (52%, P = 0.0697); however, sensitivity increased in MBIR (66%, P = 0.0035). For reader 2, it was stationary for all the following sets: FBP (65%), ASIR (68%), and MBIR (67%, P > 0.05). The MBIR and ASIR showed a limited sensitivity for detecting hepatic metastases in submillisievert LDCT.