Abstract

Purpose: The goal of this study, the first of its kind in Morocco, was to evaluate 3D set-up errors and to propose optimal margins for (Planning Target Volume) coverage in head and neck cancers. We investigated if others imaging frequency protocols were as effective as the daily one.
Methods: The alignment data from 20 patients, with daily orthogonal kilovoltage images, were collected and analyzed. The population systematic and random errors and the 3D vector of displacements were calculated. The (Clinical Target Volume) to PTV margin was generated with mathematic formulas (van Herk, Stroom, and ICRU). Other imaging protocols were simulated to evaluate the effect of imaging frequencies on the set-up errors: daily, No Action Level 5, extended No Action Level, weekly.
Results: Ninety eight percent of the set-up deviations were within 5 mm in all three directions. The population systematic and random error ranged from 0 to 0.35 mm and from 4.6 to 5.6 mm respectively. The CTV-PTV margin was, using van Hersk formulae, 4.23, 4.23, 3.96 mm in AP (Top), ML (Center) and SI (Bottom) directions respectively. There was no difference between the daily protocol and eNAL protocol (p=0.08) with 98% and 96.4% of 3D vectors within 3D vector PTV respectively. The eNAL protocol provided the best coverage of residual set up errors by the calculated CTV to PTV margin especially in the AP direction (95.8% with p<0.001).
Conclusion: A 5 mm extension of CTV to PTV margin seems to be the optimal margin. The eNAL could be strongly considered as a verification imaging protocol; however more studies are needed to confirm this result. 

Keywords: Set up errors; PTV margin; Verification imaging protocol