Program Information
Estimating Patient Water Equivalent Diameter From Localizer Radiographs -- Stability of Calibration Across Scanners, Vendors, Beam Directions, and Image Filters
D Zhang*, G Mihai , M Palmer , Beth Israel Deaconess Medical Center, Boston, MA
Presentations
SU-K-201-4 (Sunday, July 30, 2017) 4:00 PM - 6:00 PM Room: 201
Purpose: Water equivalent diameter (Dw) is widely accepted for characterizing patient size and attenuation in CT, and can be estimated from axial images or localizer radiographs. The latter has two benefits: being less affected by image truncation due to small FOV, and capable of providing pre-exam Dw at different body positions. However, this approach requires careful calibration to correlate patient attenuation and localizer pixel values, presumably on a per-scanner basis. This study reports a cross-scanner comparison of the calibration results, to answer how the calibration parameters differ: (1) for different scanners; (2) with different localizer image filters; (3) with different beam directions.
Methods: Different sized phantoms were centered and scanned to obtain localizer and axial images. A program was developed to pair each axial image with the corresponding line in the localizer to collect calibration data, which were analyzed using linear regression. The calibration was conducted on five GE scanners (3 CT750HD, 1 VCT, 1 Revolution), two Siemens scanners (Definition-AS), and three Toshiba scanners (Acquilion-One, Acquilion-Prime80, and Acquilion-64), using combinations of clinical localizer techniques (PA vs LAT for all CTs; Sharp vs Standard vs Smooth localizer filters for Toshiba CTs).
Results: Scanners from the three vendors demonstrated distinct calibration parameters. The calibration slope clustered closely around their respective means regardless of beam directions (PA vs LAT), without statistical significant difference: p=0.95 (GE), 0.12 (Siemens), 0.86 (Toshiba-Sharp/Standard), 0.46 (Toshiba-Smooth). The calibration intercept, for GE and Siemens groups, clustered regardless of beam directions [p=0.68 (GE), 0.63 (Siemens)], but differed statistically in PA vs LAT for Toshiba scanners. Standard- and Sharp-filter calibration data showed no statistical difference, and are statistically different from Smooth-filter data.
Conclusion: The study investigated the stability of localizer-Dw calibration across multiple CTs, and evaluated the impact from cross-scanner variation, scanner models/vendors, localizer beam directions, and image filters.
Funding Support, Disclosures, and Conflict of Interest: Da Zhang receives research support from Toshiba America Medical Systems.
Contact Email: