Kv Intrafraction Verification for SBRT Amplitude-Gated Rapidarc Treatments: An Initial Experience

Abstract

To report the workflow and the quality assurance program established for the first kV-IGRT intrafraction verification gated RapidArc (RA) treatment successfully delivered in our institution. One patient with stage I NSCLC was treated using the gating technique on a medical linear accelerator. Prescription dose was 48 Gy in four fractions and one fiducial marker was implanted near the tumor. During simulation a gated CT in a Stable Exhale (SE) phase (breath hold pause = 20 sec) was acquired. A 6 MV FFF plan was generated on the gated CT using four partial arcs; the fiducial and its 3 mm PRV were delineated in the planning CT. During each fraction a gating amplitude window of 2.5 mm was set, a gated CBCT on the SE was acquired and positioning was performed according to a tumor fusion. Two orthogonal kV images were then collected to verify the position of the marker. Triggered kV images were acquired immediately before beam-on to verify the intrafraction fiducial position during treatment and to perform an a posteriori evaluation of the marker displacements in the superior-inferior (SI) direction with respect to DRR images. Pretreatment quality assurance was performed using a Delta4 phantom (static delivery) and with EBT3 Gafchromic films coupled with a Quasar moving phantom (gated delivery) using three breathing pattern (BP): 1) FB: free BP of the patient, 2) EX1: SE obtained by modifying the FB pattern and 3) EX2: a random SE. A film irradiated in static delivery was taken as a reference to evaluate the gating deliveries and compared with TPS dose as additional static QA. Beam-on time was 1.8 ± 0.5 min per arc and 7.2 ± 1.4 min per fraction. The duty cycle averaged over all arcs was 42 ± 3.9 % and 56 ± 17 kV triggered images per fraction were acquired. The SI distance between the marker and its expected position was 1.7 ± 0.4 mm on average and 2.9 ± 0.7 mm at 95th percentile (range: 0 - 4.2 mm). During treatment the fiducial was always displayed inside its PRV, except during the delivery of arc 3 in fraction 3. This is in agreement with shifts evaluation, where we found an average displacement of 3.6 ± 0.3 mm. 2%/2mm gamma agreement for the static delivery was 97.8% and 100% for the Delta4 phantom and films, respectively. The gated deliveries were dosimetrically equivalent to the static procedure, with dose differences at isocenter below 1% and gamma pass-rates above 98.2%. Good agreement was also found with TPS (gamma scores > 98.9%). The system capability to superimpose marker PRV on real-time kV images made the clinicians confident during the delivery and the evaluation of the fiducial shifts has proven to be consistent to what was observed during the treatment. A 3 mm margin was required to account for 95% of the gating intrafraction uncertainties. Gated treatments were delivered to a moving phantom and no significant differences were found with respect to a static delivery. The use of a small gate coupled with a SE BP allows delivering acceptable dose distributions with reasonable duty cycles.