Click here to


Are you sure ?

Yes, do it No, cancel

Estimating Patient Water Equivalent Diameter From Localizer Images -- Constancy of Calibration Parameters Across Scanners and Over Time

D Zhang1*, X Liu2 , X Duan3 , J Rong2 , A Bankier1 , M Palmer1 , (1) Harvard, Beth Israel Deaconess Medical Center, Boston, MA, (2) UT MD Anderson Cancer Center, Houston, TX, (3) UT Southwestern Medical Center, Dallas, TX


(Wednesday, 7/17/2019) 10:00 AM - 10:30 AM

Room: Exhibit Hall | Forum 9

Purpose: Water equivalent diameter (Dw) is a sound patient-size descriptor. Localizer-based Dw estimation can minimize truncation errors due to limited FOV and provide pre-scan patient habitus information. This method is considered difficult to implement by the user community due to necessary calibration between localizer pixel values and attenuation, and unknown variations of the calibration results across scanners and over time. We investigated the constancy of calibration results across 16 CTs from 3 hospitals, and the variation over two years (on 9 CTs).

Methods: Localizer and axial images of ACR- and body CTDI-phantoms were acquired. By associating axial images with corresponding localizer lines, calibration parameters were estimated via regression. Experiments were conducted on 7 GE CTs (CT750HD, VCT, Revolution), 4 Siemens CTs (Def. AS, Force), 3 Toshiba CTs (A-One, A-Prime80, and A-64), and 2 Philips CTs (iCT, iCT256) using clinical localizer techniques (LAT and PA 120kV localizers and 120kV axial).

Results: Linear calibration results were observed on all CTs. Slope and intercept data were distinct across-vendors, but clustered closely within the same vendors. Slope, which strongly affects Dw, showed small fluctuations regardless of LAT/PA. Standard error in slope was, GE < 0.7%, Siemens < 0.7%, Canon < 0.6%, and Philips < 3%. Intercept, which corrects for patient bed attenuation, showed larger fluctuations: GE < 1.5%, Siemens < 0.9%, Canon < 8.6%, and Philips < 3.5%. Over time, the variations in slope (SD%) are < 1.3% on GE CTs; < 1.0% on Siemens; < 0.8% on Canon. The variations in intercept (SD%) are < 1.9% on GE; < 0.7% on Siemens; < 10.6% on Canon.

Conclusion: The small variations in calibration results among CTs of the same vendor and over time clearly demonstrated the practicality of implementing Dw estimation by the user community.


Not Applicable / None Entered.


Not Applicable / None Entered.

Contact Email