Intrafraction variation of target position during cone-beam CT image-guided stereotactic body radiotherapy (SBRT) for early-stage non-small cell lung cancer: A collaborative analysis

Purpose/Objective(s): Long delivery times during lung SBRT lead to concern regarding intrafraction variation (IFV) of the target position.We evaluated the magnitude of positional errors and IFVin a multi-institutional dataset SBRT. Patient, treatment, and tumor characteristics are analyzed with respect to their impact on IFV-MTP. Materials/Methods: 485 tumors underwent 1678 fractions of lung SBRT at 1 of 4 international institutions using daily online CBCT. Prescription dose varied by institution with the most common dose schedules: 18 - 20 Gy x 3, 48 Gy x 4, 50 Gy x 5, 12.5 Gy x 3. Translational target position correction of the MTP was performed via onboard CBCT. Immobilization devices differed by institutional preference, including: body frame, alpha-cradle, chest board, BodyFix, hybrid cradle/BodyFix, or no immobilization device. Multiple CBCT were performed with each fraction according to institutional protocol to potentially include: pre-correction, post-correction, and post-treatment target positional errors. IFV-MTP was measured as the difference in MTP between the post-correction CBCT and the post-treatment CBCT excluding residual error. Results: For 1678 fractions, mean ± SD pre-correction positional error was 0.19mm ± 6.45mm, 0.35 mm ± 6.8, and 1.07mm ± 7.3, in the ML, AP and CC dimensions, respectively (vector 8.97mm ± 7.94). For 1549 fractions, mean post-correction error was 0.01 ± 1.43mm ML, 0.35 ± 1.8mm AP, and 0.70 ± 1.74 mm CC (vector 2.43 ± 1.76mm). For 1501 post-treatment fractions, mean post-treatment error was 0.07 ± 1.74mm ML, 0.56 ± 2.28mm AP, and 1.04 ± 2.14mm CC (vector 3.14mm ±2.09). On univariate analysis of factors predictive of IFV>2mm, the following were significant: higher pre-treatment weight, larger respiratory tumor excursion ML, AP or CC, larger excursion vector, higher baseline FEV1 and DLCO (p<0.01 for all); lower baseline performance was of borderline significance (p = 0.09); age and GTV max dimension were non-significant. Results were similar for an IFV cutoff of 5mm. Body frame immobilization was associated with lowest IFV (p<0.001) for .2 or 5mm, and females had less IFV than males (p = 0.02). Conclusions: Multiple factors influenced IFV during SBRT including patient weight, tumor excursion, pulmonary function, immobilization and gender. Rigid immobilization such as with the Body frame reduces IFV. These data will be used to calculate and verify appropriate SBRT margins. Several patient characteristics did differ according to immobilization device.