Abstract
Purpose: A breathing motion model was developed that mapped the positions of lung and lungtumortissues to the tidal volume and airflow of the patient. This model is tested using a state‐of‐the art CT scanner. Methods and Materials: CTimages were acquired on a Philips Brilliance 64‐slice CT scanner using cine mode with 25 images per couch position and 0.68 × 0.68 × 0.625 mm3 voxels (4.0 cm longitudinal coverage). Simultaneous quantitative spirometry‐based tidal volume measurements were also acquired. The positions of the internal lungtissues were tracked by subdividing the lungtissues into 1 × 1 × 1 cm3 cubic regions and determining where those regions went in each of the 25 images. Registration was conducted using cross‐correlation maximization. The resulting positions r, tidal volumes and airflows (derivative of the tidal volume) were fit to the linear motion equation r = p 0 + α v r v + β f r f , where v and f are the volume and flow, rv and rf are the unit volume and flow vectors, and α and β are the volume and flow fitting parameters. Results: The values of images α and β varied smoothly across the lungs. In specific transverse slices, α (the ratio of motion to tidal volume) was smaller near the anterior of the lungs, increasing to a maximum near the center of the lungs and decreasing slightly near the posterior. β (the ratio of motion to airflow, a measure of hysteresis) was greater in the lateral portions of the lungs than the medial portion with little anteroposterior variation. Conclusions: Quantitative mapping the “5D” breathing model is feasible using a 64‐slice CT scanner. Quantitative mapping using fewer slices will require validated deformable registration techniques. This 5D model provides a quantitative model of the free‐breathing motion throughout the lungs. Supported in part by NIH R01CA96679.
Published Version
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call