Pseudoenhancement effects on iodine quantification from dual-energy spectral CT systems: A multi-vendor phantom study...

To measure the effect of pseudoenhancement on spectral CT iodine quantification as a function of lesion size, lesion iodine level, background iodine level, helical versus axial scanning, and spectral CT scanner type in a phantom model.

A custom-built water-filled cylindrical phantom contained either six small vials (8 mm diameter) or six large vials (27 mm diameter) of aqueous iopamidol solutions (0, 0.5, 1.0, 2.0, 4.0 and 6.0 mg iodine/mL). The background iodine concentration was 0, 5, or 10 mg iodine/mL. Helical and axial scans were taken on three different dual-energy spectral CT scanners (two image-based and one projection-based) with the scan parameters consistent between the systems. ROIs were used to measure the average iodine concentration of the vials in the 36 individual scans. Linear fits of the true versus measured iodine values were used for pvalue statistical analysis. Having a y-intercept or slope p-value less than 0.05 implied statistically significant iodine quantification errors.

Iodine quantification pseudoenhancement effects are inversely proportional to lesion size and lesion enhancement and are directly proportional to background attenuation level. No significant differences between helical and axial scans were observed. 100% and 88% of the slope and y-intercept p-values were below 0.05 for the two image-based systems, while 13% of the slope and y-intercept p-values were below 0.05 for the projection-based system.

Pseudoenhancement can artificially increase spectral CT iodine quantification levels most notably for small low-enhancing lesions (<5.0 mg iodine/mL) surrounded by a high attenuating background (10 mg iodine/mL). In this study we found iodine quantification to be more accurate on projection-based spectral CT systems than image-based systems.