Optimal Respiratory Phase Selection Balances Efficiency and Robustness In 4D Proton Therapy for Liver Cancer
Abstract
Purpose
Hepatocellular carcinoma (HCC) treatment with intensity-modulated proton therapy (IMPT) is challenged by respiratory motion, necessitating 4D robust optimization. This study aimed to optimize the number of breathing phases required for 4D robust IMPT planning to reduce computational costs while maintaining plan robustness.
Methods
Fifteen HCC patients with 4DCT data were included. Five 4D optimization strategies with varying numbers of breathing phases (10 phases, 6EX phases, 6IN phases, 3 phases, 2 phases) and one 3D optimization plan were developed using RayStation. The 4D dynamic dose (4DDD) was calculated for all plans, and dosimetric parameters (homogeneity index [HI], conformity index [CI], D98, D2) for the target and organs at risk (OARs) were analyzed. Optimization durations were compared.
Results
The 3D optimization strategy demonstrated superior nominal target coverage and lower doses to organs at risk (OARs), yet exhibited compromised robustness in the 4D dynamic dose (4DDD) scenario. In contrast, all five 4D optimization strategies provided statistically equivalent OAR protection (p ≤ 0.05). The homogeneity index (HI) for the 2 phases and 6IN phases strategies significantly exceeded that of the 10 phases approach (p ≤ 0.05), indicating reduced dose uniformity. For CI, strategies incorporating fewer phases yielded larger CI values, progressively approaching the 10 phases baseline as more phases were optimized. Notably, computational efficiency improved substantially with phase reduction: optimization durations decreased by 36.07% (6EX phases), 32.18% (6IN phases), 61.20% (3 phases), and 69.72% (2 phases) compared to the 10 phases benchmark.
Conclusion
The 6EX and 6IN phases strategies achieved dosimetric robustness comparable to 10 phases while significantly reducing computational costs. These approaches offer a practical alternative to full-phase 4D optimization for HCC IMPT, enhancing clinical feasibility.