Click here to


Are you sure ?

Yes, do it No, cancel

Impact of Image Reconstruction Kernel On CT Number to Proton Stopping Power Calibration

M Chacko1,2*, H Grewal1,2, S Rana1, D Wu2, J Sonnad2, (1) The Oklahoma Proton Center, Oklahoma City, OK, (2) University of Oklahoma Health Science Center, Oklahoma City, OK


(Sunday, 7/12/2020)   [Eastern Time (GMT-4)]

Room: AAPM ePoster Library

Purpose: The purpose of this study was to (1) assess the impact of four computed tomography (CT) vendor specific reconstruction kernels on CT number, measured in Hounsfield Units (HU), to proton relative linear stopping power (pRLSP) calibration curves, and (2) quantify the magnitude of the change in pRLSP retrospectively on proton therapy treatment plans.

Methods: Calibration curves (HU vs. pRLSP) were generated using the stoichiometric method. Seventeen tissue substitutes plugs, from -1000 HU to +1300 HU, were scanned in a Siemens SOMATOM Confidence RT Pro placing each in the center of a 203 mm diameter Lucite phantom filled with deionized water imaged at fixed technique, and reconstructed using three head and two body kernels. Calibration curves were retrospectively applied in the XiO proton treatment planning system to assess water equivalent path lengths (WEPL) along the axial isocenter plane for five prostate and five lung treatment plans.

Results: Analysis from -80 to + 1500 HU revealed no difference between pRLSP values resulting from either body kernel (Br38 and Br62), but a 3% difference between Hr38 and Br38 at 1500 HU and 0.03 absolute difference at -900 HU. Between head kernels, the coefficient of variation increased to 2% at 1500 HU. The largest absolute differences in pRLSP were at -900 HU between Hr38 and Hr68. Comparison between Hr38 and Br38 in prostate patient plans revealed no significant differences in WEPL, but significant differences (p<0.0021) in the lung calculated with either CT calibration curve up to 4 mm.

Conclusion: The HU to pRLSP calibration is not interchangeable between Siemens image reconstruction kernels. Kernel selection has a larger impact on the calibration curve between head-specific kernels than body. Incorrect selection between head and body-specific kernels appears to have the largest impact in lung plans, where absolute pRLSP difference translates to clinically significant WEPL differences.


Not Applicable / None Entered.


Not Applicable / None Entered.

Contact Email