Practical dose delivery verification of craniospinal IMRT

Craniospinal irradiation (CSI) using IMRT allows dose sparing to organs-at-risk (OAR) whilst conforming the dose to the target volume. Due to the complexity of treatment involving several isocenters, the dose distribution created by the inverse-planned segmented fields must be verified prior to treatment. We propose and test methods to verify dose delivery using commonly available dosimetry equipment for commissioning and routine patient-specific dose verification of craniospinal IMRT. Ten patients treated with conventional CSI were retrospectively planned with a 3-isocenter (cranial, upper spine, and lower spine) IMRT technique. The isocenters were placed 25-27 cm away from each other in the longitudinal direction but in the same lateral and anterioposterior positions. The planning target volume (PTV) was defined as the brain with a 0.5 cm margin and spinal canal with a 1.0 cm margin. The plans were prescribed to 36 Gy in 20 fractions to the PTV mean dose. Eleven beams (five cranial, three upper spine, and three lower spine) were optimized simultaneously. The dose delivered by the IMRT plans was then recalculated on several different phantoms and measured using the following methods: 1) ionization chamber inserted in a cylindrical phantom, positioned in the junction regions between cranial/upper-spine isocenters and upper-/lower-spine isocenters; 2) MapCHECK centered in the overlap regions; and 3) ArcCHECK measurement with beams from each isocenter. For 1) ± 3% dose difference and for 2) and 3) ≥ 95% of measured points with a γ-index < 1 for 3% dose difference and 3 mm distance-to-agreement were deemed clinically acceptable. The median (range) planned to measured dose differences for the ionization chamber is 0.4% (-1.5% to 3.0%) for the cranial/upper-spine field and 1.8% (-0.8% to 2.6%) for the upper-/lower-spine field overlap region. The median (range) percentage of MapCHECK diodes with a γ index < 1 for 3%/3 mm criterion is 98.0% (95.3% to 99.7%) for the cranial/upper-spine and 97.3% (95.0% to 99.6%) for the upper-/lower-spine field overlap regions. The median (range) percentage of ArcCHECK diodes with a γ index < 1 for 3%/3 mm criterion is 99.7% (97.1% to 100%). Three different methods of verifying craniospinal IMRT were compared and tested. All techniques offer different benefits and together can be used for 1) commissioning the treatment technique and, separately, for 2) patient-specific pretreatment verification measurements.