2. Optimization of CT-scan acquisition parameters and choice of the breathing amplitude margins for non-synchronized freebreathing technique

Abstract

Introduction Spirometry allows an indirect control of moving targets in breathing hold (BH) or their monitoring in freebreathing (FB). FB technique requires a synchronization between the CT-scan acquisition (4D-CT) and the breathing cycles. This technique is not included in every system. The Slow-CT scan is a possible alternative to optimize the monitoring of moving targets. The purpose of this study is to evaluate the feasibility of non-synchronized FB using Slow-CT scan in terms of image quality, breathing amplitude and respiratory rate. Methods Standard-CT, Slow-CT and 4D-CT acquisitions are made with a Siemens Somatom Scope Power scan, with a voltage of 130 kV and modulated intensity (CARE Dose 4D®). UH, low contrasts, slice thickness and spatial resolution are analyzed on Catphan® 504. The moving target, a lung equivalent cylindrical insert including a 3 cm diameter sphere, is set in motion by Quasar™ Respiratory Motion. The breathing cycles are created by Trace Generator. Reconstruction performances are evaluated for breaths from 10 to 30 breaths per minute (BPM) and an amplitude of 2 cm. The radiation dose is estimated by the volumetric computed tomography dose index (CTDI). The choice of the FB amplitude margins is made from a retrospective analysis of 15 patients treated in BH conditions. The image quality of the 4D-CT Average and Slow-CT sequences is examined on patients. Results Noise is reduced up to 30% in Slow-CT (σ = 3.0 UH) in comparison with 4D-CT Average (σ = 4.9 UH). A systematic error of +0.35 mm of slice thickness reconstruction occurs for 4D-CT Average. The low contrasts score (visual) increases in 4D-CT Average. In 4D-CT Average (>50 s), the amplitude reconstruction is faithful (49.8 mm instead of 50.0 mm in theory) with the 10 BPM/2 cm amplitude couple. In Slow-CT (>60 s), for BPM between 12 and 18, the mean reconstructed amplitude is 47 mm. It is 41 mm for BPM ⩽10 or ⩾30. The amplitude reconstruction is optimized for 15 BPM (49.5 mm). In Standard-CT (6–20 s), it appears distortions in the 3 reconstruction planes and relatives errors up to 60 %. In Slow-CT and 4D-CT, the mean CTDI are respectively 16.4 mGy and 15.2 mGy. The optimized Slow-CT acquisition parameters are 130 kV, collimation of 16x0.6 mm, 2 mm reconstructed slice, pitch of 0.4, rotation time of 1.5 s and a ratio (Acquisition time (s))/BPM⩾4. The retrospective analysis of patients shows a maximum interfraction amplitude difference of 100 mL. The spirometer calibration includes a 60 mL uncertainty (2% of 3 L). A margin of 100 mL is chosen as upper limit (inhalation); the baseline, allowing an automatic stabilization of the system display, is the lower limit (exhalation). On clinical images, Slow-CT shows a motion blur reduction (heart, mediastinum) and a noise reduction in a range of 10–15%. Conclusions The Slow-CT sequence offers an alternative to 4D-CT for incompatible systems. Parameters such as collimation, pitch, rotation time and the ratio (Acquisition time (s))/BPM⩾4 must be adapted. CTDI are similar for Slow-CT and 4D-CT. Margins optimization and further clinical cases are ongoing. A study about dose optimization while obtaining the same amount of signal will be carried out using different level of noise reduction algorithm.