(Wednesday, 8/1/2018) 7:30 AM - 9:30 AM
Room: Karl Dean Ballroom C
Purpose: Ventilation can be estimated using deformable image registration applied to 4D computed tomography (4DCT) datasets. Current dose response analysis using 4DCT-derived ventilation is limited to one time point. In this study we investigated whether scans at 3 or 6 months after radiation therapy (RT) showed greater ventilation response to dose.
Methods: Eight patients with non-small cell lung cancer had 4DCT datasets acquired at 3 time points: before RT, 3 months post-RT, and 6 months post-RT. For each dataset, ventilation was estimated using the Jacobian determinant of the transformation given by deformable image registration of the 4DCT dataset. The post-RT maps were deformed to the pre-RT map and a voxel-wise ratio of each post-RT map to the pre-RT map was taken to determine the ventilation response at 3 and 6 months post-RT. All voxels for all subjects were then stratified into high and low function and the average Jacobian ratio was calculated from 0 to 60 Gy in 2 Gy dose bins. Linear regression was performed to estimate the decrease in Jacobian ratio with respect to dose.
Results: All voxels showed a 1.0% decrease in Jacobian ratio per 10 Gy at 3 months (R-squared=0.71) and a 1.4% decrease at 6 months (R-squared=0.91). The 6-month slope is significantly different (p=0.018) from the 3-month slope, suggesting increased ventilation decline at 6 months. Both functional stratifications showed similar significant trends. Low-function voxels had a decrease of 0.1% and 0.4% per 10 Gy at 3 and 6 months, respectively (p=0.004). High-function voxels had a decrease of 2.1% per 10 Gy at 3 months and 2.6% per 10 Gy at 6 months (p=0.037).
Conclusion: Ventilation data 6 months post-RT showed increased sensitivity to dose compared to 3 months post-RT, suggesting 6-month follow-up scans are preferred for longitudinal studies of post-RT lung function changes.
Funding Support, Disclosures, and Conflict of Interest: NIH grant R01 CA166703
Keywords
CT, Ventilation/perfusion, Lung
Taxonomy
IM/TH- Image Analysis (Single modality or Multi-modality): Image registration
Contact Email
