The relative effect of vendor variability in CT perfusion results: a method comparison study

Area of Science:

• Radiology
• Medical Imaging
• Neurology

Background:

• Quantitative CT perfusion values are known to be affected by interoperator, intraoperator, and intervendor software differences.
• These variations can influence the reproducibility of diagnostic imaging results.

Purpose of the Study:

• To determine the relative impact of operator and software differences on CT perfusion variability.
• To compare the variability introduced by different software vendors versus different operators.

Main Methods:

• CT perfusion imaging data from 11 patients with suspected ischemic stroke were analyzed.
• Three radiologists independently postprocessed the data twice using four different vendor software applications.
• Cerebral blood volume (CBV), cerebral blood flow (CBF), and mean transit time (MTT) were recorded and analyzed using multivariate analysis of variance and Bland-Altman analysis.

Main Results:

• Vendor software choice was significantly associated with variability in CBV, CBF, and MTT (p < 0.001).
• Interoperator and intraoperator differences did not show significant variability.
• Mean differences in CT perfusion values were greater between vendors than between operators.

Conclusions:

• Different vendor software applications produce non-equivalent quantitative CT perfusion results.
• Intervendor differences are the primary source of variability in perfusion results compared to operator differences.
• Caution is advised when interpreting quantitative CT perfusion values due to potential software-induced variations.