Use of 4D-CBCT to Assess Target and Organ Motion in Esophageal Cancer Patients

Abstract

Radiotherapy for esophageal cancer is challenged by intra-fractional organ movement. Daily image guidance is used to assess patient set-up errors and correct for positional offsets of the target and organs at risk. There exists a paucity of published data on four-dimensional cone-beam computed tomography (4D-CBCT) as a method of image guidance for these patients. This study utilized 4D-CBCTs to compare the efficacy of different registration volumes (bone, internal target volume (ITV), carina) for image guidance with focus on whether carina is a useful ITV surrogate. The study also assessed our centre's current planning target volume (PTV) margins of 0.5 cm around the ITV. This retrospective study involved 17 esophageal cancer patients treated radically from July 2012-July 2013 using volumetric-modulated arc therapy. Daily CBCTs (402) were reconstructed into 4D Medium Resolution 8-phase 4D-CBCTs using Elekta XVI software and registered to the helical CT simulation scan in voluntary exhale. Two separate dual registration matches were done; the first registered a clipbox around the vertebrae and a mask around the ITV, and the second used carina as the mask. T-test was used to test for significant difference between X,Y,Z-offsets; means and standard deviations were calculated. For each directional shift, the average difference in amplitude (absolute value of maximum inhale plus maximum exhale) between 4DCT and 4D-CBCT was determined. In some comparisons of vertebrae/ITV/carina mean offsets, statistical significance was found (t-test, p=0.05). However, only comparison of vertebrae and ITV mean offsets revealed clinical significance for the Y-dimension with the magnitude being >0.2 cm. For 14 patients (3 excluded due to unavailable data), average difference in amplitude between 4DCT and 4D-CBCT in the X,Y,Z-directions were determined to be 0.12 cm, -0.05 cm, and 0.04 cm, respectively. In 8.4% of the cases, movement of the ITV in the Y-direction was greater than the PTV margin. Carina appears to be a better ITV surrogate versus vertebrae for image guidance on 4D-CBCTs in the Y-dimension. In the majority of cases, PTV margins are sufficient for target coverage. Future research on patient-specific indicators for significant ITV movement in the Y-direction is needed.