Room: Track 1
Purpose: To characterize the impact of motion artifacts on ventilation maps generated from clinical 4-dimensional computed tomography (4D-CT) images and compare them to ventilation maps generated from volumetric 4D-CT (v4D-CT).
Methods: Four patients with non-small cell lung cancer who received a clinical helical 4D-CT scan for treatment planning on a Philips Brilliance Big Bore CT scanner were also imaged using a research protocol on a GE Revolution volumetric CT scanner. Clinical 4D-CT images were acquired using: helical mode, 0.5s/revolution, 120kV, 45-60mA, pitch=0.056-0.075, and 24mm axial field-of-view (aFOV). On the volumetric scanner, images were acquired at a single couch position using the following parameters: cine mode, 0.28s/revolution, 100kV, 100mA, and 160mm aFOV for 45-50s. The end-inhale and end-exhale phases were selected from each image set, and the lung volume was semi-automatically segmented in 3DSlicer using a fast-marching technique to aid in registration. The images were non-rigidly registered in Elastix using a b-splines transformation and Mattes mutual information. Specific ventilation was then calculated using an existing density-based method. Finally, the v4D-CT and conventional 4D-CT ventilation maps were registered using Elastix, and the maps were compared visually.
Results: Motion artifacts were visible in all clinical 4D-CT images, while none were visible in the v4D-CT images. For all patients, ventilation measured by clinical 4D-CT differed from v4D-CT in regions where motion artifacts were present. For two patients, motion artifacts led to mis-registration of the inhale and exhale phases, which resulted in visually obvious artifacts in the final ventilation image.
Conclusion: The motion artifacts and resulting ventilation artifacts that are known to exist using conventional 4D-CT may be eliminated by v4D-CT. Artifact reduction with v4D-CT may enable the wider application of CT ventilation in clinical decision making.