Investigation of compact multi-layer ionization chamber detector designs for quality control of CBCT-based synthetic CTs for adaptive proton therapy.

To define an optimal compact multi-layer ionization chamber (MLIC) detector design for proton radiography (PR) acquisition as quality control (QC) of cone-beam CT (CBCT)-based synthetic CTs (sCTs) in the context of adaptive proton therapy (APT). Approach:In 10 patients, 50 PR measurements were simulated. To replicate measured IDDs (= IDDsMeas), noise based on actual PR measurements was added to the raw IDDs (= IDDsRaw). IDDsMeaswere discretized to 1-6mm and limited to 30-50mm to mimic MLIC detector designs with 30 different ionization chamber (IC) pitch and detector length combinations. The influence of noise and space/limited IDD information on the ability to determine range shifts was assessed by comparing IDDRawto IDDMeas. To assess how well different MLIC detector designs identify/detect sCT uncertainties, 43 error scenarios were generated by modifying CT calibration curves of muscle/fat/bone-like density tissues. The range shifts were calculated by fitting nominal IDDs to IDDsMeasof each error scenario using least-square. The normalized comparison of expected and measured range shifts is reported as normalized relative range shifts, ideally equaling zero. Main Results:Each MLIC detector design was investigated based on 506 IDDRaw, with 100 different noise contributions each, resulting in 50600 IDDMeas. The relative range shift and relative normalized range shift increased with a larger IC pitch. An increased detector length allowed for a larger IC pitch. Detector lengths <35mm proved insufficient for sampling enough relevant IDD information. Significance:This study demonstrates that range shifts due to density errors between measured and simulated IDDs can be identified using a compact MLIC detector. A normalized relative range shift uncertainty detection of ≤0.51% can be verified with a minimum 45mm detector length and maximum 3mm IC pitch. The obtained results suggest that a compact MLIC detector can be used to verify the integrity of CBCT-based sCTs, enabling APT based on sCTs.