1145 - Fully Flexible Lattice Position Optimization and Motion-Robust Single-Field-Each-Peak Optimization for LATTICE Therapy

Purpose/Objective(s): Lattice therapy (LATTICE) is a form of spatially fractionated radiation therapy that delivers high-dose peaks to tumor sub-volumes while sparing surrounding healthy tissues. Traditional LATTICE relies on rigid vertex arrangements, limiting adaptability for irregularly shaped tumors or those near critical organs. Proton LATTICE therapy (pLATTICE) also faces challenges due to proton range uncertainties and intra-fractional motion, which can misalign high-dose peaks. This study introduces FlexLRT, a novel planning method that simultaneously optimizes fully flexible vertex placement and dose distribution. Additionally, it presents a single-field-each-peak (SFEP) approach to enhance motion robustness in pLATTICE.

Materials/Methods: FlexLRT integrates vertex positioning with treatment parameters, such as proton spot weights or photon fluences, within a constrained optimization framework. Vertex positions are treated as variables subject to spatial constraints to improve dose conformity. For pLATTICE, SFEP assigns each high-dose peak to a single, optimally selected field angle from a predefined set, reducing motion-induced range uncertainties. Field selection and dose optimization are formulated as a mixed-integer programming problem, solved using the alternating direction method of multipliers and iterative convex relaxation techniques.

Results: FlexLRT and SFEP (NEW) were validated against the heuristic searching (HS) method. HS plans were created using conventional LATTICE methods with varying vertex placements for FlexLRT and randomly preselected field configurations for SFEP. Three plans were selected for comparison, termed BEST, MID, and WORST, corresponding to the smallest, median, and largest total optimization objective f, respectively. In photon LATTICE abdomen plans, the peak-to-valley dose ratio (PVDR) increased from 3.00 (WORST) to 5.88 (NEW), a 96.0% improvement. For pLATTICE, SFEP ensured single-field delivery per peak, improving organ-at-risk (OAR) sparing while maintaining comparable PVDR and conformity index relative to the best HS plan.

Conclusion: This study introduces a novel LATTICE planning approach that integrates fully flexible vertex positioning with dose optimization and a motion-robust SFEP method for pLATTICE. Compared to conventional techniques, the proposed framework enhances PVDR, improves target coverage, and provides superior OAR sparing, making LATTICE therapy more adaptable and effective for complex tumor geometries and motion-sensitive treatments.