Program Information
Estimating Patient Water Equivalent Diameter From Localizer Radiographs -- An Automated Calibration Approach
D Zhang*, G Mihai , M Palmer , Beth Israel Deaconess Medical Center, Boston, MA
Presentations
TH-CD-601-5 (Thursday, August 3, 2017) 10:00 AM - 12:00 PM Room: 601
Purpose: Water equivalent diameter (Dw) is the preferred metric for characterizing patient size and attenuation in CT. Two methods exist for estimating Dw: from axial images or from localizer radiographs. The latter method requires careful calibration to correlate patient attenuation and localizer pixel values, but has the benefits of: (1) being less affected by image truncation due to small field of view, (2) capability of estimating Dw before actual scans start, (3) capability of estimating Dw at different body positions from one image. This study reports an automated approach to simplify the complexity of the calibration process.
Methods: A Python program was developed to automatically pair each axial image with the corresponding row or column in the localizer radiograph, and collect calibration data based on a user-selected polygon-based ROI in the axial image and the matched localizer-line. The pairing process was achieved by projecting pixel locations from axial image space to the patient coordinate system, and then to the scout image space. Calibration data were collected from routine CT QC exams (ACR phantom, dose phantoms, water phantoms, etc.) on a GE CT750 HD scanner, and combined to determine the calibration coefficients.
Results: The program accurately pairs the corresponding localizer line and axial images. Calibration data from phantoms of different sizes fall on a single linear regression line with R^2 > 0.99. For a randomly selected test case, per-slice Dw's calculated from scout and from axial images agree within 3% for the lateral localizer, and within 5% for the posterior-anterior (PA) localizer, due to larger patient mis-centering in the AP/PA direction.
Conclusion: With the proposed automated approach, localizer-to-attenuation calibration can be conveniently conducted from routine QC scans, allowing clinically practical estimations of patient Dw.
Funding Support, Disclosures, and Conflict of Interest: Da Zhang receives research grant support from Toshiba America Medical Systems.
Contact Email: