The purpose of the study was to examine whether CT imaging can be used to quantify radiation-induced injury to the esophagus. Weekly CT images for 14 patients receiving proton therapy for thoracic tumors were retrospectively reviewed. The images were registered with the original treatment planning CT image using deformable registration techniques, and the esophageal contours from the treatment plan were automatically mapped to the weekly images. The relative change in the size of the esophagus was calculated for each CT slice as the ratio of the cross-sectional area of the esophagus (minus air) in the weekly CT image to the same area in the planning CT image. The maximum relative change in cross sectional area for each CT image was calculated and examined for correlation with the clinical toxicity score for all the patients. The average maximum relative expansion of the esophagus at the end of treatment was 1.41 ± 0.26, 1.68 ± 0.36, and 2.10 ± 0.18 for patients with grade 0, 2, and 3 esophagitis, respectively. An unpaired t-test, with the level of significance corrected with a Bonferroni correction, showed that the difference between grade 3 and 0 was significant, but the differences between grade 0 and 2, and 2 and 3 were not. The timing of changes in esophageal expansion closely matched that of clinically noted changes in patient symptoms. Expansion of the esophagus on CT images has potential as an objective measure of toxicity. The ability to quantify objectively the spatial distribution of radiation-induced injury will be a useful tool in understanding the impact of partial esophageal sparing on the probability of esophagitis.