2295 - Treatment Planning Feasibility Study for Lattice SBRT of Locally Advanced Head and Neck Cancer

R. D. Hall¹, E. Czarnecki¹, A. R. Bhatnagar¹, K. Snyder¹, F. Siddiqui¹, and K. Thind²; ¹Department of Radiation Oncology, Henry Ford Cancer Institute, Detroit, MI, ²Department of Radiation Oncology, Henry Ford Health, Detroit, MI

Purpose/Objective(s): Lattice Radiation Therapy, a form of spatially fractionated radiation therapy (SFRT) delivers non-uniform dose distributions comprised of characteristic high- and low-dose spheres of 100% and 30% of prescription, respectively. Neoadjuvant SFRT of locally advanced head and neck cancer (HNC) has shown increased local control and possible reduction of distant metastasis with clinically acceptable long-term toxicities. This study evaluates three treatment planning approaches with varying levels of automation for clinical goal compliance and time efficiency. These use automated and manual beam geometry and dosimetric optimization tools to evaluate tradeoffs in clinical compliance and efficiency. Materials/

Methods: Five patients with locally-advanced HNC (N3+) were prescribed 15 Gy in 1 fraction. High-dose spheres (PTV_High) 1 cm in diameter were placed 1.8 cm apart (center-to-center) within rGTV (5mm inner margin from GTV), avoiding critical organs. Low-dose spheres (PTV_Avoid) 1cm in diameter were placed equidistant between the high-dose spheres in the four cardinal directions. Treatment planning approaches of automated SRS (HA), manual SBRT (mSBRT), and hybrid (HA-mSBRT) were utilized. The hybrid approach utilized the automatically optimized beam geometry obtained in the HA process but forewent the use of an automatic SRS NTO for manually determined NTO parameters and dose optimization structures. Plans were normalized for identical target coverage (PTV_HighD95% = 100%). Clinical goal compliance including OAR dose, and dose gradients were evaluated, in addition to total planning time for evaluation of time efficiency.
Results: Total average planning times were 185 minutes (HA), 183 minutes (mSBRT), and 167 minutes (HA-mSBRT). OAR doses exceeded a 2 Gy max dose goal 20, 17, and 17 times for HA, mSBRT and HA-mSBRT for 5 patients each. Additional to the increased frequency of OAR tolerance breaches, HA lowered the minimum target dose and significantly lengthened the optimization time. The ratio of PTV_High to PTV_Avoid mean dose was 4.09 (HA), 3.76 (mSBRT), and 3.92 (HA-mSBRT). Dose drop-off, as measured by the ratio of median dose to standard deviation of dose within a 5 mm ring surrounding PTV_High was 2.608 (HA), 2.489 (mSBRT), and 2.610 (HA-mSBRT). The above metrics quantify the dose heterogeneity necessary for SFRT plan quality by evaluating the dose difference between high and low regions and dose gradient steepness between prescribed high and low-dose regions. The mSBRT method was the most MU-efficient, using only 1.5 MUs per cGy-degree, in contrast to approximately 2.7 for both the HA and HA-mSBRT methods.
Conclusion: The HA-mSBRT strategy demonstrates the highest clinical goal compliance with high OAR sparing, high dose gradient, and dose uniformity, while being both MU- and time-efficient. This semi-automated approach should provide a feasible solution for prospective use of SFRT in HNC.