Investigation of heterogeneity corrections for accelerated Monte Carlo-based treatment planning of plaque ocular brachytherapy.

OBJECTIVE

To develop and validate a practical method for accurate dose calculations of low-dose-rate brachytherapy treatment of eye tumors.

METHODS

Dose calculations of 16 mm Collaborative Ocular Melanoma Study (COMS) and 20 mm notched eye plaques with 125I seeds were performed using the egs_brachy package of the EGSnrc Monte Carlo (MC). Several eye phantoms were modeled, including: (1) uniform phantoms made of only tumor (T), water, or vitreous body (VB) medium; (2) scaled water (SW) and scaled VB (SVB) distributions were derived by scaling the precomputed distributions for all-water and all-vitreous geometries by the relative energy absorption coefficient or each ocular structure to that of water or the vitreous body, respectively; (3) fully defined eye phantom with all structural materials defined; and 4) TG-43 calculations. Two contrasting tumor locations were modeled in this work. For investigation of dose distribution with standard plaques, the tumor was symmetrically located on the medial side of the eye. For the notched plaque, dose distributions were investigated with the tumor located posteriorly (abutting the optic nerve).

RESULTS

Compared to the fully-defined phantom calculations, all-water phantom calculations underestimated the tumor dose by 15-20%, while overestimating the lens and sclera dose. The optic nerve dose was predicted correctly. VB phantoms underestimated the tumor coverage by ∼3.5%, while T phantoms accurately predicted the target coverage. Both VB and T phantoms overestimated the organ-at-risk (OAR) doses more than all-water phantom. SW and SVB phantoms accurately predict both tumor and OAR doses within 4% of the complete phantom values for small volume metrics. Dose differences decreased to < 2% to volumes > 20%.

CONCLUSION

The dose scaling methodology provides a practical approach to account for ocular heterogeneity effects. This leads to accurate dosimetry, without computational penalties associated with patient-specific MC calculations. The methodology is found to be robust and applicable to both standard COMS and notched eye plaques of different sizes.