Optimizing CT Acquisition for Prostate Cancer Radiation Treatment Planning and Delivery

G. Chen,A. Tai,F. Liu,C. Lawton,A. Li

DOI: https://doi.org/10.1016/j.ijrobp.2013.06.1878

2013-01-01

Abstract:CT acquisition protocols are usually established based on diagnostic purpose with major emphasis on imaging dose, thus, are not necessarily optimized for radiation therapy (RT) planning and delivery, where slightly increased imaging dose in exchange for improved soft tissue contrast may be justified (e.g., prostate SBRT). In this work, we investigate the impact of CT data acquisition parameters on image quality for treatment planning and delivery guidance in prostate cancer RT. A CT scanner installed inside a linac room (i.e., CT-on-rails) was used. Various CT acquisition parameters (e.g., tube voltage and current, rotation time, pitch, imaging dose control methods) and reconstruction algorithms were first tested on a phantom then on patients. A cylindrical phantom 32 cm in diameter with two plugs of relative electron densities of 1.05 and 0.96 was scanned using a predefined CT protocol (120 kVp, 250 mAs, CareDose off) routinely used for daily IGRT of prostate RT. This protocol leads to a CTDIvol of 17 mGy for a typical prostate case. The phantom was scanned with a series of customized protocols with different parameter sets (e.g., tube voltage of 100, 120, 130 and 140 kVp with effective exposure of 250-1800 mAs). The phantom images obtained were compared and evaluated, and a set of parameters that results in optimal image quality (soft tissue contrast) was used to create an image quality enhancement protocol (IQEP) which was then used to scan selected prostate cancer patients as a part of their regular IGRT procedure. For each patient, no more than two sets of CTs were acquired using the IQEP. The CT data were reconstructed using various available reconstruction algorithms from the vendor. The image quality of the obtained CTs was assessed with (1) accuracy and reproducibility of auto-segmentation for the prostate gland using a tool and (2) intra- and inter-variability of manual contouring of the prostate gland by 10 observers. The phantom data shows that the contrast to noise ratio (CNR) increased with appropriate combinations of mAs and kVp. The dose weighted CNR peaked around 1000 mAs with 120 kVp for the 1.05 electron density plug and 100 kVp for the 0.96 electron density plug. kVp < 100 increased imaging noise. Visual inspection of the CTs obtained with IQEP of 100 kVp and 987 mAs (a CTDIvol of around 50 mGy) showed improved visibility of the prostate gland boundary. This improvement was supported by the reduced intra- and inter-variability of manual contouring. The CT protocols may be optimized for radiation therapy to improve soft tissue contrast, thus, improving the accuracy of target/organ delineation and treatment delivery. If the optimized protocol is used with in-room CT for IGRT, the slightly increased imaging dose is comparable to other IGRT modalities.