Automated Early Identification of an Excessive Air-in-Oil X-ray Tube Artifact That Mimics Acute Cerebral Infarct.

ABSTRACT

PURPOSE: This study aimed to develop an automated, quantitative method to increase the likelihood of identifying and preventing such air-in-oil (AIO) artifact. METHODS: A 1-dimensional radial representation of the 2-dimensional noise power spectrum (NPS) was calculated from AIO artifact images and compared with artifact-free images. A quality control (QC) software program was modified to include measurements of NPS average frequency within the water section of daily phantom scans. Threshold values for each CT system were incorporated into daily QC. RESULTS: Noise power spectrum for AIO artifact images included a large low-frequency peak compared with artifact-free images; NPS average frequencies were 0.197 and 0.319 line pairs per millimeter for AIO artifact and artifact-free images, respectively. Automated QC successfully identified 3 AIO artifacts before detrimental clinical effect occurred. CONCLUSIONS: Serious clinical problems associated with AIO artifact can be detected and avoided by incorporating NPS average frequency measurements of daily phantom images into an automated QC program.