Feasibility study on using low dose CT for adaptive proton therapy

In adaptive proton therapy, imaging is imperative for daily patient setup, monitoring anatomical changes, dose prediction, and replanning. However, utilizing cone-beam computed tomography (CBCT) is limited by artifacts and inaccurate Hounsfield units (HUs). Although, in-room CT resolves these limitations, its accumulated imaging dose across multiple fractions is concerning for secondary radiation-induced malignancies, particularly when it comes to pediatric patients. This work comprehensively investigated the feasibility of using low dose CT images for online proton therapy adaptation. Two quality assurance and three anatomical phantoms were scanned using 18 different CT scan techniques. 10 proton therapy patients were included in this study. Patients were imaged daily using different techniques: normal clinical dose (conventional) CT scan and 2 low dose CT scans. The Signal-to-noise ratio (SNR) of the CT images with 3 imaging dose levels were investigated. Treatment plans were generated to study dosimetric impact of reduced imaging dose. With decreasing imaging dose, SNR was reduced in all cases. Dose calculation on the CTs showed minor variation of dose distributions in patient and phantom scans. Neither the target coverage nor organs-at-risk (OAR) sparing showed clinically considerable differences between conventional, low, and ultra-low dose CT scans. This study showed the feasibility of using low dose CT scans for replanning in adaptive proton therapy. Optimized CT protocols with low imaging dose should be determined based on each institution's practice and treatment sites.