Program Information
Reported Clinical Outcomes in the Treatment of Monte Carlo-Based Lung SBRT Patients: Results of Heterogeneity Corrected Pencil Beam Vs X-Ray Voxel Monte Carlo Algorithms
D Pokhrel*1, S Sood2, R Badkul2 , H Jiang2, H Saleh2, F Wang2, 1University of Kentucky, Radiation Medicine, Lexington, KY, 2The University of Kansas Cancer Center, Kansas City, KS
Presentations
SU-I-GPD-T-580 (Sunday, July 30, 2017) 3:00 PM - 6:00 PM Room: Exhibit Hall
Purpose: Heterogeneity corrected pencil-beam (PB-hete) and X-ray Voxel Monte Carlo (XVMC) dose calculation algorithms have shown huge dosimetric difference when used in the treatment of lung SBRT. Although, direct correlated of clinical outcomes have rarely been reported previously with PB-hete vs XVMC algorithms for lung SBRT. The aim of this study was to report dependence of the local-control on dose calculation algorithms used for lung SBRT.
Methods: Total of 160-NSCLC patients treated with SBRT were clinically followed up. All patients were treated with 50-60Gy in 5 fractions of 10-12Gy each. 76-patients were planned in BrainLab iPlan TPS with PB-hete algorithm (2009-2012) and treated at Novalis-TX (6MV-SRS;1000MU/min). Other, 84-patients (109 tumors) were planned using same TPS with XVMC algorithm and treated at same machine (2013-2015). Both groups used hybrid (3D-conformal-dynamic-arcs plus static-beams) plans that were prescribed to 75-80% isodose line. Numerically, similar treatment planning objectives were used for both groups. Median follow-up times were 33-month (8-77) and 19-month (7-40) for PB-hete and XVMC groups, respectively. Primary endpoints were local/loco-regional control-rates of treated lesion.
Results: For the same nominal-prescription, identical beam geometry, and total number of monitor units (MU), PB-hete algorithm overestimated target coverage by 10-20%, on average, up to 35% in some cases–significantly underdosing tumor with PB-hete algorithm. 12 (15.8%) out of 76-patients treated with PB-hete and 10 (9.2%) out of 84-patients (109 tumors) treated with XVMC algorithms to same nominal-prescription had local recurrence. Statistically significant difference in recurrence rate (p=0.04) between PB-hete and XVMC groups was observed while using Kaplan-Mayer analysis.
Conclusion: Local/loco-regional control in patients who were treated to nominal-identical prescriptions with PB-hete and XVMC algorithms were statistically significantly different, although, delivered total number of MU were higher (average~15%) with XVMC group. While clinically available, Monte Carlo-based dose calculation is highly recommended for lung SBRT for better local control.
Contact Email: