Poster — Thur Eve — 04: Characterizations of image quality for CT with iterative reconstruction algorithm — development of a routine quality control metric

Abstract

Iterative reconstruction algorithms (IRA) are available on modern CT systems. We are investigating the feasibility of routinely and objectively characterizing the image quality of CT with IRA. Noise‐equivalent‐quanta (NEQ) conceptually and potentially can be the metric of image quality assessment for system performance. CT images of the third module of an ACR phantom were acquired on a GE LightSpeed VCT system using 5 different mA ranging from 25–400 with 6 different ASiR levels ranging from 0–100. MTF and NPS measurements were obtained from these CT images using the method described by Friedman et al1. It was found that MTF acquired using even low mA is reliable and consistent. MTF from data acquired using 400mA provides decreased MTF performance for the reason yet to be determined. There is incremental improvement in MTF by the incremental increase in ASiR level. While ASiR makes no improvements in NPS and NEQ at low spatial frequency, the ASiR increment level of 25 has the same effect of doubling the dose to the NPS and NEQ performances at frequency 0.5 lp/mm. A yet to be defined way of quantifying the spatial spectrum content shift by ASiR should be required for QC when ASiR is used for dose reduction. This study provides a preliminary understanding and characterization of objective image quality of CT with IRA. Further studies on the scene dependency of the image quality by IRA and different implementations of IRA by other CT manufactures are forthcoming.