KEY POINTS
- The in silico study included 25 patients with bulky solid tumors >100 cm³. Each patient had four plans generated: jaw-partitioned and conventional full-open-jaw VMAT for both 3D-GRID and lattice treatments.
- All plans prescribed 18 Gy in one fraction, with 95% of the GRID or lattice peak volume receiving 100% of prescription. Jaw-partitioned VMAT divided the X-jaw field into lateral and central subfields across three arcs.
- Peak-dose coverage was maintained, but dose fall-off improved substantially. Gradient index fell from 11.18 to 9.48 for GRID and from 14.06 to 11.03 for lattice plans, both p < 0.001.
- Peak-to-valley separation also improved. PVDR D50 increased from 2.68 to 2.86 for GRID and from 2.90 to 3.08 for lattice plans, both p < 0.001, alongside lower dose in non-peak GTV regions.
- The advantage became greater as tumors became larger. GTV volume strongly correlated with improvement in gradient index for both GRID (ρ = 0.807) and lattice (ρ = 0.809) planning, both p < 0.001.
- The major cost was treatment complexity. Monitor units increased from 7,975 to 15,844 for GRID and from 8,815 to 18,922 for lattice, while the plan-complexity metric also increased significantly.
- All 50 jaw-partitioned plans passed ArcCHECK QA with gamma passing rates ≥95% using 3%/2 mm criteria. Conventional plans were not measured, so comparative delivery accuracy cannot be inferred.
CLINICAL TAKEAWAY
Jaw partitioning appears to strengthen the fundamental peak-valley geometry sought in spatially fractionated radiotherapy, particularly for very large tumors. Whether the dosimetric improvement is worth roughly doubling treatment output and potentially increasing delivery time or motion sensitivity remains a practical question.