Micro CT calibration accuracy for pre-clinical studies in ion therapy

Accurate estimation of tissue stopping power ratios (SPR) from computed tomography (CT) data is essential in charged particle therapy (CPT), yet it remains a major source of range uncertainties. While various SPR calibration methods have been developed for clinical CT scanners, they are not tailored to cone beam type Micro CT scanners commonly used in pre-clinical studies involving small animals. This work aims to assess how different clinical SPR calibration methods influence range predictions in small animal CPT pre-clinical trials, evaluated at both the voxel and organ levels. Approach: A small animal calibration phantom was imaged with a Micro CT at X-ray tube voltages between 45 kV and 70 kV. The images were used in two single energy (SECT), a stoichiometric and a simplified method, and one stoichiometric dual energy (DECT) calibration method. The accuracy of SPR predictions for each method was evaluated using simulated CT images of murine head, thorax, and abdomen virtual phantoms derived from in vivo data. Main results: Both SECT methods were found to yield comparable accuracy in the head and abdomen areas. Root mean squared (RMS) errors to the nominal SPR values were found to be 5.04% and 4.22% in the head phantom and 2.31% and 2.61% in the abdomen phantom, for the stoichiometric and simplified SECT method, respectively. RMS errors in the lung phantom were found to be 16.42% and 35.80%. The stoichiometric DECT method was found to improve accuracy in the lung phantom to 12.37% while in the head phantom it achieved 19.37% and in the abdomen 6.42%. Significance: This is the first report of treatment plan agnostic low-kV Micro CT SPR calibration accuracy. It can provide a reference for expected SPR and range uncertainties in pre-clinical CPT trials.