Effect of radiologists' experience with an adaptive statistical iterative reconstruction algorithm on detection of hypervascular liver lesions and perception of image quality.

Abstract

Purpose: To prospectively evaluate whether clinical experience with an adaptive statistical iterative reconstruction algorithm (ASiR) has an effect on radiologists' diagnostic performance and confidence for the diagnosis of hypervascular liver tumors, as well as on their subjective perception of image quality.

Materials and methods: Forty patients, having 65 hypervascular liver tumors, underwent contrast-enhanced MDCT during the hepatic arterial phase. Image datasets were reconstructed with filtered backprojection algorithm and ASiR (20%, 40%, 60%, and 80% blending). During two reading sessions, performed before and after a three-year period of clinical experience with ASiR, three readers assessed datasets for lesion detection, likelihood of malignancy, and image quality.

Results: For all reconstruction algorithms, there was no significant change in readers' diagnostic accuracy and sensitivity for the detection of liver lesions, between the two reading sessions. However, a 60% ASiR dataset yielded a significant improvement in specificity, lesion conspicuity, and confidence for lesion likelihood of malignancy during the second reading session (P < 0.0001). The 60% ASiR dataset resulted in significant improvement in readers' perception of image quality during the second reading session (P < 0.0001).

Conclusions: Clinical experience using an ASiR algorithm may improve radiologists' diagnostic performance for the diagnosis of hypervascular liver tumors, as well as their perception of image quality.