Evolution of dose distribution for a 125I eye plaque using MCNPX and Plaque Simulator software

Abstract

Objective

Eye plaque radiation therapy is the treatment of choice for small- and medium-sized choroidal melanomas. This study investigated the dose distribution around eye plaques containing ¹²⁵I seeds to treat ocular melanoma using Monte Carlo N-Particle eXtended (MCNPX) and Plaque Simulator (PS) software.

Methods

Dosimetry evaluation and comparison of the resulting isodose curves for ¹²⁵I COMS plaques were performed using the MCNPX code. The isodose curves and dose distributions were calculated using PS treatment planning for a ¹²⁵I COMS plaque.

Results

In the validation, the maximum relative difference between the results of this study and those reported in other literature was approximately 9%–10% for the COMS plaques. The dose distributions of MCNPX were lower than those of PS with a relative difference of approximately 27.7%–35.4%. The dose distribution may differ depending on the ¹²⁵I source spectra and seed design used in the two methods. In addition, the dose algorithm used in PS made a major contribution to the relative differences between the results.

Conclusion

The PS did not provide accurate details of dose distribution near the surface of the plaque insert. The source parameters used in each program should be studied more carefully to determine the source of the differences in the estimated dose values.

INTRODUCTION

Episcleral plaque radiation is the most popular treatment for choroidal melanoma management to preserve the globe, vision, quality of life, and cosmetic outcomes.^1,2 This therapy is the treatment of choice for small- and medium-sized choroidal melanoma tumors, whereas enucleation is the most commonly used treatment for large choroidal melanoma tumors.³ Ocular melanoma brachytherapy can provide patients with good long-term results in terms of local tumor control. Physicians have used several radionuclides such as ¹²⁵I and ¹⁰⁶Ru to treat tumors.^4–7 In radiotherapy, the risk of normal tissue toxicity limits the dose that can be safely directed toward the tumor. Therefore, precise dosimetry is fundamental for both tumor dose prescription and the determination of the risk of normal tissue toxicity. The problem of large ocular tumors in Iran is attributed to the dispersion of the population and limited access to treatment using radioactive plaque, such that the dimensions of most patients’ tumors (apical height) are approximately 5 mm or more.⁸ For larger ¹²⁵I plaques, 85 Gy can be delivered in three, four, five, six, or seven days using 8.6, 6.5, 5.2, 4.4, or 3.8 U, respectively.

Currently, the methods of plaque brachytherapy dosimetry are limited by the size and spatial resolution of detectors such as thermoluminescent dosimeters, where steep dose gradients create significant challenges. This limitation leads to incomplete dose distributions and, inevitably, inaccurate treatment planning for quality assurance of eye plaques before clinical use. Dose planning for ¹²⁵I COMS plaque brachytherapy is performed using Bebig software in our therapy center.

This study evaluated the dose distribution around an eye plaque containing a new ¹²⁵I seed. Dosimetry evaluation and comparison of the isodoses for ¹²⁵I-LDR plaques were performed using Monte Carlo N-Particle eXtended (MCNPX) and a Plaque Simulator (PS) software system.

MATERIALS AND METHODS Monte Carlo calculation of COMS plaques

The dose profile around ocular plaques was examined using the Monte Carlo method with MCNPX code, considering the recommendations of the report of Task Group 129 by the American Association of Physicists in Medicine and American Brachytherapy Society.⁹ MCNPX is a Monte Carlo radiation transport computer code capable of tracking gamma rays and particle types at nearly all energies and is also used for modeling the interaction of radiation with matter. ¹⁰ The RMESH Tally used in MCNPX generates the amount of energy absorbed for the transport of mesh cells or the voxel volume.¹¹ All components used in this simulation (such as surface, cell, material, source, and number of particles per source [NPS] card) are described in the input file and shown in Figure 1.

FIGURE 1. MCNPX modeling including of COMS plaque and 125I seed in xy plan. Material type (density in gr/cm3) corresponding to cells1to 9: eye ball (1.03), sillastic insert (1.12), gold cover of plaque (15.8), silver rod of seed (10.5), radioactive coating on silver rod (6.425), air inside seed (0.000120479), titanium capsule and end welds (4.54), skull bone around eye (1.6) soft tissue out sides skull bone (1.0).

Cubic mesh tabulation with 0.1cm³ cells was used to derive the dosage profile. The data were viewed and the values for plaques containing ¹²⁵I seeds were extracted using the ORIGIN program. For ocular plaques, isodose diagrams related to the deep dose percentage were created for all three plans (X, Y, and Z). According to these graphs, the dose at the top of the tumor in each case was equal to 100% of the deep dose. The reference dose was 85 Gy.⁴

The F6 tally (MeV/g) was used to calculate the depth dosage rate. Small spheres (between 0.1 and 0.2 mm) were defined for this purpose and are shown in the modeling in Figure 1, which provides further details on each cell, particularly the coordinate system and the names, materials, and dimensions of the seed source cell.

Considering the suggestions of Task Group 129, the tissue-equivalent phantom in MCNPX modeling was taken to be a 30 cm box on one side of the eyeballs containing liquid water.^12–15 Both the plaque and phantom were created in accordance with Thomson's guidelines.¹⁶ Figure 1 illustrates the eye, which was a sphere with a radius of 1.23 cm in the middle of the modeling box. The x-axis depth of the air section was 13.77 cm, whereas the y- and z-dimensions of the air section were 30 cm× 30 cm. The eye model additionally contained a cornea composed of a second sphere with a radius of 0.727 cm situated on the x-axis 0.663 cm from the origin of the eye. To simulate the muscle and other supporting tissues, liquid water was poured into the space between the eye and bone and outside the plaque. Plaques with diameters of 16 mm and 20 mm were placed on the surface of the eyeball to measure the dose that entered the eyeball along the plaque's central axis. The F6 tally was used to measure dose variations along the plaque's axis inside 0.2 mm-diameter spheres. The dose contribution from the photons was considered. To accelerate the dose calculation process, the F6 tally may be used instead of the *F8 tally. The NPS was set to 10⁸ and the cut-off energy was 1 keV. The uncertainty of the results was approximately 4%.

In this study, the ¹²⁵I seed was designed in a manner similar to that of a previously investigated ¹²⁵I source.¹⁷ The source was AgI with a density of 6.2 g/cm³, which was coated on a silver wire (3.2 mm in length (Figure 1)). It was encapsulated in a titanium tube (4.7 mm in length, 0.8 mm in outer diameter, and 4.54 g/cm³). The energy spectrum of the source is shown in Table 1.

TABLE 1 Photon spectrum for the decay of ¹²⁵I to ¹²⁵Te according to the ICRP-38 report.

Energy (KeV)	% Intensity
4.17	0.04
4.83	0.11
3.61	0.11
31.24	0.14
3.34	0.18
4.83	0.18
4.57	0.45
4.07	0.50
4.12	0.82
4.30	0.99
4.03	3.58
31.71	4.30
3.76	6.14
35.49	6.68
30.94	7.20
31.00	14.00
27.20	39.78
27.47	74.08
Total	%159.28

Plaque simulator software system

To calculate the dose distributions, dedicated treatment planning systems (TPSs) are available for individual tumor/seed configurations used in episcleral plaque therapy.^18,19 The Bebig PS is used at our partner institute (Farabi Eye Hospital) for dose calculation of critical structures in ¹²⁵I plaque implantation. This TPS was specifically designed for ocular plaque radiation therapy and was developed by BEBIG GmbH. The latest version of the PS was released in 2005 and was designed to work on Mac OS PCs. This treatment planning system offers a 3D treatment simulation and modeling package for different types of radioisotopes, such as brachytherapy sources and several different commercial plaques. The 24 seeds were arranged on the COMS plaque, with an activity of 3.6 mCi for each ¹²⁵I seed. The planning parameters for the PS are recommended to be listed in a table slot parameter of the 20 mm plaque, and the Cartesian coordinates of the center of the seed are listed in Table 2.

TABLE 2 Slot parameters of the 20 mm plaque and Cartesian coordinates of the center of the seed.

Seed No.	Alpha (deg.) °	Beta deg) °	Tilt (deg.) °	Offset(mm)	X	Y	Z
1	39.57	20	270	0	1.70	−2.97	−8.15
2	39.3	60	270	0	1.70	−7.51	4.34
3	39.3	140	270	0	1.70	−8.55	−1.51
4	39.3	180	270	0	1.70	−5.58	−6.65
5	39.3	180	270	0	1.70	0.00	−8.68
6	39.3	220	270	0	1.70	5.58	−6.65
7	39.3	260	270	0	1.70	8.55	−1.51
8	39.3	300	270	0	1.70	7.51	4.36
9	39.3	340	270	0	1.70	2.97	8.15
10	29.5	11	270	0	0.38	−1.29	6.65
11	29.5	62	270	0	0.38	−5.96	3.17
12	29.5	268	270	0	0.38	6.74	−0.24
13	29.5	165	270	0	0.38	−1.75	−6.52
14	29.5	114	270	0	0.38	−6.16	−2.74
15	29.5	216	270	0	0.38	3.97	−5.46
16	29.5	319	270	0	0.38	4.43	5.09
17	20	36	270	0	−0.57	−2.75	3.79
18	20	108	270	0	−0.57	−4.46	−1.45
19	20	180	270	0	−0.57	0.00	−4.69
20	20	252	270	0	−0.57	4.46	−1.45
21	20	324	270	0	−0.57	2.75	3.79
22	9	90	270	0	−1.23	−2.14	0.00
23	9	270	270	0	−1.23	2.14	0.00
24	0	90	270	0	−1.40	0.00	0.00

RESULTS Verification of depth dose calculation for MCNPX

The normalized depth dose on the z-axis for a 16 mm-diameter COMS ¹²⁵I eye plaque was calculated using the MCNPX code and compared with the results of two published studies,^20,21 as shown in Figure 2. The depth doses were normalized to a depth distance of 5 mm along the central axis (z-axis) of the tumor apex. In addition, the depth dose along the z-axis for a 20 mm-diameter COMS ¹²⁵I eye plaque was calculated using the MCNPX code and compared with the results of Melhus et al.,¹⁵ as shown in Figure 2. The location of the origin in this model was considered as the center of the eyeball, whereas the origin of depth distance r in Figure 2 was z = -1.23 along the z-axis as surface of silastic insert. The results of this comparison were in good agreement, except for the very near depth of the plaque. These results helped validate the simulation code because the phantom and source geometries were similar.

FIGURE 2. Comparison of the normalized depth dose in the depth distance of 5mm (as a tumor apex) on the central axis (z-axis) for 125I COMS plaque: (A) COMS plaque with 16mm in diameter, (B) COMS plaque with 20mm in diameter.

Dose distribution and isodose curve using MCNPX

The F6 tally was used to measure the dose variation along the central axis of the plaque. In Figure 3, the isodoses are shown in the y-z plane perpendicular to the x-axis. The z-axis was perpendicular to the plaque and was considered the central axis. Along the z-axis, the 100% isodose at a depth of 11 mm was considered the tumor apex. As shown in Figure 3, the dose distribution was from multiple seeds (24 seeds).

FIGURE 3. A) Isodose lines from MCNPX for 20 mm plaque; B) location of the 20 mm plaque, origin of the eyeball (with r = 12.3 mm) and the 125I seed sources.

Depth dose and isodose curves using PS

The isodose curves for ¹²⁵I seeds located in the COMS design plaque with a diameter of 20 mm were generated using the PS, as shown in Figure 4. The depth rate dose on the z-axis for the eye plaque with a 20 mm-diameter was determined using the PS and compared with the MCNPX result. Figure 5 shows that the dose distributions of the PS were higher than those of MCNPX.

FIGURE 4. PS retinal diagram for the 125I COMS plaque with a diameter of 20 mm.

FIGURE 5. Plaque's central axis dose values for the 20mm plaque using MCNPX and PS.

DISCUSSION

We investigated the dose rate of ¹²⁵I COMS plaque to treat ocular melanoma tumors. We performed MCNPX simulations using the above eye model and eye plaque to calculate the dose distribution in the eye. The PS used source specifications defined by AAMP's TG-43U1 in the ¹²⁵I seed model 6711, whereas the MCNPX simulation used the photon spectrum of ICRP-38 for the decay of ¹²⁵I (Table 3). The dose algorithm used in the PS was an analytical algorithm that contributed to the relative difference between the results. The plaque model employed in the PS was the same as that used in the MCNPX simulation. The seed used in the PS was the 6711 model and its geometry was similar to that of our seed (IRseed model). However, significant differences were observed at the end of the seeds.

TABLE 3 Photon spectrum of the ¹²⁵I source (model 6711) according to the TG-43 and AAPM reports.

Photon energy (keV)	% Intensity
31.710	4.39
35.492	6.68
30.980	20.20
27.202	40.60
27.472	74.70
Total	%147.6

As part of simulation validation, the maximum relative difference between the results this study and those of other literature was approximately 9%–10% for the COMS plaque. The results in Figure 2 indicate that for the COMS plaque with a 16 mm diameter, the maximum relative difference between the results of this study and those of Rivard et al. was 9.1%, and this difference with Bristol et al. was 9.5%. Furthermore, for the COMS plaque with a 20 mm diameter, the maximum relative difference between the results of this study and those of Melhus et al. was 9.5%.

We compared our results with those of Bristol et al.²⁰ and found a slight difference, with a relative difference of 12%–15%. This could be due to the small difference in the caps of the two ends of the capsule because the photon spectrum was used; the ICRP-38 photon spectrum for decay was ¹²⁵I, which was the same in both models. The ¹²⁵I seed was modeled in Bristol's study using the Amersham Oncoseed Model 6711. The source was composed of AgBr and AgI in a molecular ratio of 2.5:1 and was coated onto a silver wire with a thickness of 1.0–1.5 µm. Including the radioactive coating, the silver wire was 0.395 cm long with a radius of 0.025 cm. The silver wire was encapsulated in a titanium cylinder with an outer radius of 0.04 cm, walls 0.006 cm thick, and an inner length of 0.375 cm.

In addition, the ¹²⁵I plaque isodose curve was obtained using the PS treatment planning software (Figure 4). The reference dose was 85 Gy. Figure 5 shows that the dose distributions of the PS were higher than those of MCNPX. A factor increase of 1.8–1.9 in the dose calculated from the PS over that of the MCNPX program was observed, which could be related to the source strength specification. The difference in dose values between the PS and MCNPX could be partly attributed to the difference in the photon spectrum of the source. The PS software used the source specification defined by the AAMP report TG-43U1 for our ¹²⁵I seed (model IRseed), which is similar to Model 6711.¹⁷

As shown in Figure 3, the details of the dose distribution near the surface of the plaque could be obtained by MCNPX, whereas the PS did not provide accurate details of the dose distribution near the surface of the plaque insert or in the tissue and bone surrounding the eye. This lack of detail may be important when estimating the areas of the sclera that receive high doses during the treatment of large melanoma tumors.

Zimermann et al. compared eye plaque dosimetry for the Eye Physics EP917 using the PS and Monte Carlo simulation. The Eye Physics EP917 is designed to be semi-elliptical with a notch on the posterior edge, which is different in design. Furthermore, they employed 17 seed sources that differed in design from our ¹²⁵I seeds.²² The relative difference observed in the study by Zimermann et al. was less than that observed in this study.

Figure 5 shows that the dose distributions of MCNPX were lower than those of the PS with a relative difference of approximately 27.7%–35.4%. Although one of the reasons for this difference may be the difference in the source photon spectrum, the difference in the algorithms also contributed significantly. The dose distribution was calculated using MCNPX and the PS with the same spectrum for comparison. The results indicated that the contribution of this relative difference related to the difference in the dose algorithm was approximately 19%, and the contribution related to the spectrum difference was approximately 12%. The MCNPX simulation used the ICRP-38 photon spectrum for the decay of ¹²⁵I (Table 1). The TG-43U1 photon spectrum is a simplified spectrum of photons emitted from a ¹²⁵I seed; therefore, it is a hardened version of the actual ¹²⁵I spectrum (Table 3). In MCNPX, the brachytherapy seed was entirely simulated as a thin coating of silver iodine inside the seed. Therefore, the photon spectrum passing through the titanium capsule of the seed was attenuated. If the complete photon spectrum from ¹²⁵I is considered in the model, the simulation data will be more accurate. However, the difference in the source spectra may cause a difference in the dose distribution.

CONCLUSIONS

We investigated the dose distribution for standard eye plaques including a ¹²⁵I seed (model IRseed) to treat ocular melanoma tumors using MCNPX and the PS. The dose distributions of MCNPX were lower than those of the PS with relative differences of approximately 27.%–35.4%, and the dose gradients from both programs agree well. The results indicate that MCNPX and the PS can be used to estimate dose distributions from COMS eye plaques loaded with ¹²⁵I seeds (model IRseed). The PS dose distributions were greater than those of MCNPX. The difference in the source photon spectrum and dose algorithm used in the PS caused this relative difference in the results. The source parameters used in each program should be studied more carefully to determine the source of the differences in the estimated dose values.

ACKNOWLEDGEMENTS

The authors have nothing to report.

CONFLICT OF INTEREST STATEMENT

The authors declare no conflict of interest.

ETHICS STATEMENT

This is a pure physics paper and there is no patient related data presented. Ethics Statement is not applicable to this paper.

INSTITUTIONAL REVIEW BOARD STATEMENT

Not applicable.

INFORMED CONSENT STATEMENT

Not applicable.