Purpose: To characterize respiratory motion and four-dimensional computed tomography (4D-CT) artefact reduction using a volumetric CT (vCT) scanner.
Methods: 4D-CT images of a Quasar Respiratory Motion Phantom (Modus Medical Devices, London Canada) with a cedar insert containing four polystyrene spheres (5-30mm diameter) and two lung cancer patients were acquired on a GE Revolution 256-slice vCT scanner, and a clinical 16-slice Philips Brilliance Big Bore CT simulator. The phantom study included four programmable breathing conditions: 1) sinusoidal (2cm amplitude, 15 breaths-per-minute (bpm)), 2) Sinusoidal (2cm amplitude, 12bpm) scanned assuming 15bpm to simulate incomplete sampling (IS), 3) baseline drift (BD), and 4) irregular amplitude (IA). vCT was acquired using cine mode, 0.28s/revolution, 160mm axial field-of-view (aFOV), 120kV, 100mA for 10s, and 10mA for 45s. The images acquired at 10mA were recombined to recover signal-to-noise ratio (SNR). The motion of the spheres was automatically tracked and compared to the known trace using Pearson correlation. Clinical 4D-CT simulations were acquired using helical mode, 0.5s/revolution, 24mm aFOV, 120kV, 97mA, and pitch adjusted for respiratory rate. Motion artefacts were qualitatively assessed for all studies.
Results: Phantom motion for each 45s vCT acquisition was strongly correlated to the known motion trace (sinusoidal: r=0.996,p<.001; BD: r=0.81,p<.001; IA: r=0.90,p<.001). SNR improved from 5.8 to 12.8 when images were retrospectively recombined, while having a lower imaging dose (CTDIvol=29.5mGy) compared to the 100mA scans (CTDIvol=64.6mGy). SNR was also higher than the clinical 4D-CT acquisition (SNR=8.8). 4D-CT images of both the phantom and lung patients had clearly visible motion artefacts when scanned on the clinical CT simulator, but no visible artefacts when scanned on the vCT scanner.
Conclusion: vCT scanners can capture respiratory tumor motion without motion artefacts that are commonly present in conventional 4D-CT scanners. Improved characterization of respiratory motion can have significant implications for targeted radiation therapy.