(Tuesday, 7/16/2019) 9:30 AM - 10:00 AM
Room: Exhibit Hall | Forum 7
Purpose: Minimizing radiation-induced acute esophagitis (AE) in locally advanced non-small cell lung cancer is important given the proximity between the esophagus and the tumor. The esophagus expansion is due to inflammation/edema during radiotherapy (RT). The purpose of this pilot study is to quantify local esophagus expansion and to provide an early prediction of AE for use in adaptive RT.
Methods: Weekly MRIs were acquired during RT under a prospective IRB approved study. Thus far, nine patients have been treated using conventionally fractionated intensity-modulated RT to 60-70Gy with concurrent chemotherapy. Six patients developed moderate AE, at a median of 4 weeks after RT start. The weekly MRIs were registered using B-spline regularized diffeomorphic registration and Jacobian of the transformation was computed to measure the maximum local esophagus expansion. The weekly associated esophagus contours were propagated onto the planning-CT and the delivered esophagus dose was accumulated on a voxel-by-voxel basis. A logistic regression model was built to predict AE using mean esophagus dose (MED) and maximum esophagus expansion (MEex) of week 1-3 (w1-3) up to which no patient developed AE. The performance of the models was compared to a published MED only model.
Results: The population median accumulated MED at the end of RT was significantly higher than the planned dose (24Gy vs. 23Gy, p=0.01). The overall best model included MED and MEex from w1-3 (AUC=1.00) vs. w1-2 (AUC=0.92) and w1 (AUC=0.75), respectively. The performance of the published MED only model was considerably worse using MED w1-3 (AUC=0.67).
Conclusion: Combining esophageal accumulated dose and maximum expansion information during the first three weeks of RT generated better predictions of AE compared to using dose or expansion only information within the first two weeks after RT start. Routine collection of both sources of information could facilitate adaptive RT in order to further reduce AE.
Keywords
Dose Response, Registration, MRI
Taxonomy
IM- Radiation dose and risk: Models
Contact Email
