Set-up errors in radiotherapy for oesophageal cancers--is electronic portal imaging or conebeam more accurate?

Radiother Oncol. 2011 Feb;98(2):249-54. doi: 10.1016/j.radonc.2010.11.002. Epub 2010 Dec 6.

Abstract

Purpose: To compare kV computed tomography (CBCT) with electronic portal imaging (EPI) and evaluate set-up variations in the anterior-posterior (AP), right-left (LR) and cranio-caudal (CC) directions and rotational variations: pitch, roll, and yaw, for oesophageal cancer patients treated with radical radiotherapy.

Methods and materials: Twenty patients with locally advanced oesophageal cancer treated with chemoradiation were consented for this prospective ethics approved protocol. Patients were positioned using skin marks/tattoos, kV-CBCT scans (XVI) and EPI's were performed prior to treatment and registered to the planning CT scans and digitally reconstructed radiographs, respectively. XVI data was used to adjust patient setups before treatment delivery. A total of 122 EPI pairs and 207 CBCT scans were analysed. The systematic and random errors were calculated.

Results: The systematic and random errors (mm) for XVI were 1.3, 1.7, 1.4 and 2.6, 3.9, 2.0 in RL, CC and AP direction, respectively, with EPI of similar magnitude. There was no correlation between the 2 modalities of imaging as 31.7% of all image pairs were discordant >3 mm and 12.5% >5 mm. XVI identified rotations >3° in 44 images.

Conclusions: EPI results in different position correction for verification of radiotherapy in oesophageal malignancies when compared with CBCT. CBCT verification offers adequate 3D volumetric image quality to improve the accuracy of treatment delivery for oesophageal malignancies in radiotherapy and should be used for image guidance.

Publication types

  • Research Support, Non-U.S. Gov't

MeSH terms

  • Adult
  • Aged
  • Aged, 80 and over
  • Cone-Beam Computed Tomography / methods*
  • Esophageal Neoplasms / diagnostic imaging
  • Esophageal Neoplasms / pathology
  • Esophageal Neoplasms / radiotherapy*
  • Female
  • Humans
  • Male
  • Middle Aged
  • Prospective Studies