Animal tissue-based quantitative comparison of dual-energy CT to SPR conversion methods using high-resolution gel dosimetry. (24th March 2021)

• Record Type: Journal Article
• Authors: Niepel, K B; Stanislawski, M; Wuerl, M; Doerringer, F; Pinto, M; Dietrich, O; Ertl-Wagner, B; Lalonde, A; Bouchard, H; Pappas, E; Yohannes, I; Hillbrand, M; Landry, G; Parodi, K
• Is Part Of: Physics in medicine & biology. Volume 66:Number 7(2021)
• Publication Date: 2021-03-24
• Subjects: dual-energy CT -- proton therapy -- dosimetry -- stopping power; Biophysics -- Periodicals; Medical physics -- Periodicals; 610.153
• DOI: 10.1088/1361-6560/abbd14
• Languages: English
• ISSNs: 0031-9155

Abstract

Abstract: Dual-energy computed tomography (DECT) has been shown to allow for more accurate ion therapy treatment planning by improving the estimation of tissue stopping power ratio (SPR) relative to water, among other tissue properties. In this study, we measured and compared the accuracy of SPR values derived using both dual- and single-energy CT (SECT) based on different published conversion algorithms. For this purpose, a phantom setup containing either fresh animal soft tissue samples (beef, pork) and a water reference or tissue equivalent plastic materials was designed and irradiated in a clinical proton therapy facility. Dosimetric polymer gel was positioned downstream of the samples to obtain a three-dimensional proton range distribution with high spatial resolution. The mean proton range in gel for each tissue relative to the water sample was converted to a SPR value. Additionally, the homogeneous samples were probed with a variable water column encompassed by two ionization chambers to benchmark the SPR accuracy of the gel dosimetry. The SPR values measured with both methods were consistent with a mean deviation of 0.2%, but the gel dosimetry captured range variations up to 5 mm within individual samples. Across all fresh tissue samples the SECT approach yielded significantly greater mean absolute deviations from the SPR deduced using gel range measurements, with an average difference of 1.2%, compared to just 0.3% for the most accurate DECT-based algorithm. These results show a significant advantage of DECT over SECT for stopping power prediction in a realistic setting, and for the first time allow to compare a large set of methods under the same conditions.