TH-C-18A-03: Novel Technique for Dynamic Lung Ventilation Imaging Based On Wide Coverage 4D CT

Purpose: An emerging lung ventilation imaging method based on conventional 4D CT has a high translational potential and advantages over other competing modalities, however suffers from binning artifacts and irregular breathing during a scan. We propose a novel technique to compute dynamic ventilation based upon wide coverage 320-slice 4D CT, which provides high spatial and temporal resolution and binning artifact-free images. Methods: Wide coverage 4D CT images were acquired using a commercial 320-slice CT scanner, with 160 mm craniocaudal coverage and 0.35 s rotation time, during tidal breathing for five patients with thoracic cancer. The image at each of 51 time points was deformably registered to a peak-exhalation phase image using B-spline deformable image registration (DIR). For each time-point, the Jacobian determinant of deformation acted as a surrogate for regional ventilation. Dynamic ventilation was then computed by comparing time-resolved ventilation images. The accuracy of DIR was quantified by calculating target registration errors (TREs) of 100 anatomic landmarks per patient. Results: The proposed technique provided regional ventilation information of high spatial and temporal resolution. The average TRE of five patients was 1.4±1.3 mm. Dynamic properties of regional ventilation were found to be spatially heterogeneous. For example, ventilation increased during inhalation and decreased during exhalation in well-ventilated normal lung regions. However, in poorly-ventilated emphysematous regions, ventilation showed relatively minor changes with time around zero. Furthermore, ventilation resolved at this timescale demonstrated a signal that appeared to come from the heartbeat, especially in proximity to the heart. Conclusion: 320-slice 4D CT-based ventilation imaging provides regional ventilation information of high spatial and temporal resolution. Dynamic ventilation information obtained by this technique may be useful in creating a model of dynamic ventilation that could be used to compensate for errors in conventional 4D CT-based ventilation imaging, and may also provide new insights into respiratory physiology. Supported in part by Free to Breathe Young Investigator Research Grant.