Correlation Between Metabolic Uptake in Planning PET and End-Of-RT Tumor Change in Lung Cancer
P Zhang*, S Alam, M Thor, A Rimner, J Deasy, W Lu, Memorial Sloan Kettering Cancer Center, New York, NY
Presentations
PO-GePV-M-350 (Sunday, 7/10/2022) [Eastern Time (GMT-4)]
ePoster Forums
Purpose: To investigate spatial and temporal patterns of tumor response to radiotherapy among locally advanced non-small-cell lung cancer (LA-NSCLC) patients using longitudinal images acquired during radiotherapy (RT), which would facilitate the design of optimal dosing strategies and adaptive radiotherapy to improve the therapeutic ratio.
Methods: This retrospective pilot study included 11 LA-NSCLC patients who were treated with concurrent chemoradiotherapy to 60Gy in 30 fractions. The longitudinal image dataset for each patient included a free breathing planning PET-CT, and consecutive weekly CBCTs acquired during RT. The gross tumor volume (GTV) was contoured on all images by thoracic radiation oncologists. Patterns of spatial tumor changes were derived by deformable registration between the planning CT and each of the weekly CBCTs using a large deformation diffeomorphic metric mapping algorithm. Local tumor volume change was quantified by a Jacobian map. Voxel-level correlations between the Jacobian map and SUV were calculated.
Results: The volume of primary GTV ranged 23 cc to 173 cc with a median of 91 cc. On average, GTV shrinkage was 10±14%, 21±20%, 27±18%, 29±19%, and 33±20%, at week 2-6. The Pearson correlation coefficient between the global tumor shrinkage observed on the last week image and the mean and max SUV on the planning PET was -0.63 and -0.67, respectively. On a voxel level, the average correlation between the histogram of the Jacobian map and SUV was -0.69.
Conclusion: The moderate correlation indicates that larger tumor shrinkage, measured by both global volume change and the local Jacobian, is associated with higher metabolic uptake measured by PET. Such information is helpful to select candidates who may benefit from adaptive RT, and to further guide the optimization of biological dose painting. This promising pilot study is supporting a larger patient study to validate the findings.
Keywords
Not Applicable / None Entered.
Taxonomy
Not Applicable / None Entered.
Contact Email