Backscatter measurements from a single seed of I-125 for ophthalmic plaque dosimetry

Med. Phys. 15 (3), May/Jun 1988, Copyright 1988 Am. Assoc. Phys. Med.
Gary Luxton, Melvin A. Astrahan, and Zbigniew Petrovich
Department of Radiation Oncology, University of Southern California, Los Angeles, California 90033

(Received 11 November 1987; accepted for publication 17 March 1988)

To determine the dosimetric effect of a gold plaque applicator used in I-125 ophthalmic irradiation, relative dose rates at points 2-18 mm transverse to the axis of a single seed of I-125 were measured in an acrylic phantom under three different measurement conditions. The detectors were 1-mm diameter X 3-mm length LiF thermoluminescent dosimeters (TLD's). Conditions corresponded to the following: (i) full scatter, (ii) the presence of an ophthalmic gold plaque, and (iii) no scatter material on the side of the seed opposite to the TLD's. The dose rate with the gold plaque is less than that with full scatter phantom. There is no significant decrease in dose rate at 2.2 mm from the seed. Dose rate is significantly reduced at greater distances. The does rate decrease ranges from 4% at 5 mm to 10% at 18 mm. The I-125 seed in the gold plaque gives 3%-5% higher dose rate than in the absence of backscatter material.

Key words: brachytherapy, ophthalmic irradiation, iodine-125, dosimetry, choroidal melanoma


Temporary placement of an episcleral plaque containing radioactive material has gained acceptance as a treatment for choroidal melanoma and other tumors of the eye since the work of Stallard with Co-60 (1). interest in the technique has grown in recent years as seen, for example, from the report of Shields et al. (2). The use of radioactive materials other than Co-60 has also been implemented, as reported, for example, by Packer et al. (3) and Lommatzsch (4). It is advantageous to replace Co-60 for episcleral plaques with a less penetrating radiation. There is unnecessary exposure to both patients and staff with Co-60, since it is difficult to shield against its radiation.

Dose specification for ophthalmic plaques has been based on measurements and calculations for full scatter geometry. Full scatter assumes, in effect, that the plaques compensate for missing or shielded tissue. The plaques generally consist of foil shells of high atomic number (Z) material containing the radioisotope. As pointed out by Weaver (5), the assumption of full scatter geometry is suspect for I-125 because of the low (approximately 28 keV) average energy of radiations from this isotope. At this energy the absorption cross section is much larger in high-Z materials than in tissue. Measurements with gold plaques containing Ir-192 by Luxton et al. (6) do not show any substantial deviation from full scatter calculations at distances as close as 2 mm from the plaque.

In measurements using planar slab geometry and gold foil backing next to an array of I-125 seeds, model 6702 (3M Corp. Medical Products Division, St. Paul, MN), nominal activity 1.5 GBq (40 mCi), Weaver (5) observed a dose rate decrease of 8%±3% in the 5-15 mm range. Weaver's results (5) showed no difference between gold sheet backing and complete absence of backscatter. Indeed, at 5-mm depth, Weaver (5) found 4% less dose rate with gold foil backing compared to no backing material present. However, with an uncertainty stated to be ±3% for each measurement, this difference was not significant. No clear pattern emerged as to the spatial dependence of the dose decrease. While the measured dose rate decrease relative to full scatter reported by Weaver' was independent of distance at 8%±3% over the range 5-15 mm, the largest effect actually observed was 11% at 5 mm.

In order to independently confirm the effect reported by Weaver (5) as well as to quantify its radial dependence, we examined experimentally the apparent dose rate decrease from a single I-125 seed in a gold plaque relative to the case of full scatter geometry. Measurements were made at points perpendicular to the seed axis. Instead of the planar geometry used by Weaver, the measurement geometry of the present report simulated that of an ophthalmic treatment, positioning the seed on a tissuelike hemisphere representing the eye. The dosimetric effect of a conforming gold plaque applicator was studied. An I-125 seed model 6711 (3M Corp. Medical Products Division, St. Paul, MN) nominal initial activity 180 MBq (4.86 mCi) was used. This seed, which is described by Ling et al. (7), differs internally from the model 6702 used by Weaver (5), and is often used in ophthalmic plaque therapy.


A. Phantom and I-125 seed

Dose rate was measured using LiF thermoluminescent dosimetry (TLD) in an acrylic phantom of dimensions 25 X 25 X 25 cm. Acrylic (C5H8O2,) was used because it is a readily machined solid at room temperature, with a linear attenuation coefficient (8) within 5% of that of water at the average I-125 energy of approximately 28 keV (7). The solid form of acrylic simplifies the problem of precise repositioning of seed and detectors required to obtain statistical significance in the measurement results.

The phantom was made up of several pieces which fit together along machined surfaces (Fig. 1). A 25-mm-diameter cylinder was centrally located in the phantom. One end of the cylinder was flat, while the other end was hemispherical. Parallel holes 1 mm in diameter were drilled through the rounded end of the cylinder, passing through the axis. Hole diameter was just large enough to hold LiF TLD-100 detectors [TLD-100 (LiF) supplied in crystalline form by Harshaw/Filtrol, Solon, OH], 3 mm X 1 mm diameter. At the center of the hemispherical surface an indentation a few tenths of a millimeter deep and approximately 1 mm wide was cut to enable reproducible positioning of a single I-125 seed model 6711. This seed is 4.5 mm in total length, with 3 mm active length. The seed was located at the approximate geometrical center of the composite acrylic block phantom. Its repositioning uncertainty was estimated to be ±0.5 mm along its length. The same seed was used for all measurements.

Figure 1

FIG. 1. Schematic diagram of the central part of the full scatter acrylic phantom. An I-125 seed is positioned at the apex of the dome. TLD's are inserted into slots in line with the seed to measure dose rate at various distances from the seed axis. An acrylic backscatter block is placed on the side of the seed distal to the TLD's. The full phantom consists of acrylic sections that extend more than 10 cm in each direction from the seed.

The full scatter acrylic phantom shown in Fig. 1 was one of three experimental setups for the measurements. The others are shown schematically in Fig. 2. The measurement conditions corresponded to either the presence of a gold alloy ophthalmic plaque or to the complete absence of scattering material on the side of the seed distal to the TLD's. No measurements were made with scatter material present on the other side of the gold plaque. No measurable differences are expected from the presence or absence of scattering material on the other side of the gold plaque since there is less than 0.1% transmission of I-125 radiations through the gold. The plaque was a spherical shell approximately 1.5 mm thick, 18 X 12 mm chord length. Its radius of curvature was 12.5 mm, the same as that of the hemispherical end of the acrylic cylinder.
1. Detector location

TLD rod detectors were positioned at the centers of their respective holes between nylon spacers. Longitudinal reproducibility was estimated to be + 0.5 mm. Six holes were used, the centers of which were 2-18 mm from the center of the I-125 seed. The transverse distance from the I-125 seed to each of the six TLD locations was determined from 3:1 magnified radiographs of I-mm-diameter steel needles placed at the TLD sites. The seed location was marked by a thin lead foil. The distances were determined with a measurement lens to a precision estimated to be ±0.05 mm.

Figure 2

FIG. 2. Schematic diagram of the scatter measurement conditions. Dose rate was measured in the absence ofbackscatter as well as with a gold plaque placed on the side of seed distal to the TLD's.

B. TLD measurements

1. Detector selection, annealing, and readout

Upon receipt from the manufacturer, TLD rods were washed in alcohol, annealed at 400°C for 1 h, and then at 100°C for another hour. The same anneal procedure was followed after each use of the TLD's. Prior to readout, rods were annealed at 105°C for 15 min. Detectors were sorted into groups of equal sensitivity to within ±3.5% according to their readouts following exposure to 1 Gy of 4-MV x rays. The sorting procedure was repeated twice over the 70 day course of the measurements. Frequent readouts of fixed interval exposures to a stable light source (built into the TLD reader, Harshaw model 3000, Harshaw/Filtrol, Solon, OH) were used to monitor the reader, permitting corrections for continuous small changes in sensitivity that occurred over the course of reading out a batch of TLD's.

2. Detector calibration

Each TLD exposure to the I-125 seed was separately calibrated by exposure of 6-10 other TLD's from the same group to a dose 1 Gy to water of 4-MV x radiation. Calibration TLD's and measurement TLD's were read out at the same time. Several rods from the same group, but which had not been exposed to radiation, also were cycled at that time through readout, including prearmeal. This provided a value for system noise, or instrument background, which was subtracted from the sample readings. Instrument background expressed as a fraction of measurement readings varied from less than 0.1% at 2 mm from the seed to as much as 7% for measurements at the most distant location (17.7 mm), due to the rapid falloff of dose rate with increasing distance. All TLD's, including those used for measurement of background, passed through a complete anneal cycle before reuse.

TABLE 1. Ratio of dose rate measurements for the experimental conditions of Fig. 2 relative to the full scatter phantom (Fig. 1). Columns 3 and 5 are the same, corrected for the inhomogeneities in the phantom due to the TLD's (see text).

Transverse distance
from seed (mm)
Gold shell backscatter
Measured dose compared
to full scatter
Gold shell backscatter
Ratio corrected
to water phantom
No backscatter
Measured dose compared
to full scatter
No backscatter
Ratio corrected
to water phantom
0.962 ±0.019

3. Repetition of measurement and correction to reference date

To reduce statistical uncertainty, each measurement of dose rate from the I-125 seed was repeated at least ten times. Dose rate was obtained relative to 4-MV x-ray calibration. The mean number of repetitions was 14.1, averaged over all data points. Exposures to the I-125 seed lasted from 1 to 18 h. All measurements were corrected to the reference date using an I-125 half-life of 59.6 days (9). In all exposures, the six slots most proximal to the I-125 seed each contained a TLD rod.


Ratios of dose rate for the experimental conditions of Fig. 2 relative to dose rate of full scatter geometry with the same seed are given in Table 1, and plotted in Fig. 3. At 2.2-mm distance, there are only slight differences among the three cases. The dose rate is greater with a gold plaque than with no scatter (p = 0. 15). The differences of the two cases from full scatter are not statistically significant, with p values of 0.28 and 0.22, respectively.

Figure 3

FIG. 3. Ratio of dose rate measured from a single I-125 seed with no backscatter or with a gold plaque present (Fig. 2) to the dose rate measured with full backscatter (Fig. 1).

Over the range 5-18 mm, there is a pattern of increased relative effect on dose rate with increasing distance. Dose rate with the gold plaque is consistently less than that for full scatter geometry, and is consistently larger than dose rate in the complete absence of backscatter. For example, at 13.8 mm, dose rate with a gold shell is 9.4% less than full backscatter (p < 0.001) and 5.3% larger than with no backscatter (p < 0.02). At 17.7 mm, dose rate with no backscatter and dose rate with the gold plaque are 16% and 11% less, respectively, than dose rate with full scatter. Table 1 summarizes the measured ratios of dose rate relative to measurements with full scatter.

Compared to a homogeneous water phantom, the measurements overestimate dose rate differences among full backscatter, gold backscatter and no backscatter, as discussed below. The inhomogeneity in the phantom from the TLD rods overattenuates the direct photon dose component compared to a homogeneous acrylic phantom. Because of the geometry of the phantom, the direct photon component at each measurement location, and which is the same for all three cases, must have traversed the TLD rod detectors that are more proximal to the seed. On the other hand, due to the changes in direction that accompany scatter at low energy, scattered radiation has low probability to pass through intervening TLD's. The scatter component, which accounts for dose rate differences among the phantoms, is thereby overestimated slightly as a fraction of the total dose. Using a simple primary plus scatter dose model and using the Monte Carlo calculations of Dale (10). for a point source in water to estimate the primary dose fraction in water, and also assuming that effect of the TLD's on scattered radiation is negligible, a correction factor was calculated for each case to obtain relative dose rate for dose to a homogeneous phantom. The average correction was 0.9%. The resulting ratios of dose rate in water, i.e., corrected for the inhomogeneity of the TLD rods in the phantom, but not for the use of acrylic, are also given in Table 1.

The result evident in Fig. 3 and Table 1 is that there is a significant decrease in I-125 dose rate in the range 5-18 mm transverse to the seed if backscatter material is entirely absent or if a gold shell is present. Over the entire 2-18 mm range, the dose rate with a gold plaque is approximately 3%-5% larger than with no backscatter material. There is possibly a slight dose rate increase from the plaque relative to full scatter at 2 mm from the seed. A very localized dose rate increase, if real, could be due to 9-14 keV (11) L shell fluorescent x rays from the gold.

The fractional effect on dose rate due to missing backscatter increases as distance from the seed is increased. In the close proximity of the seed, the dose rate contribution from direct radiation becomes relatively large because of the geometrical inverse square distance factor. Dose rate due to scatter, however, does not increase so abruptly at small distances, since scatter dose is essentially a nonlocal effect. For example, in a point source Monte Carlo calculation of dose rate in water by Dale (10), the scatter fraction was found to increase montonically from 0% at the seed to 52% at 20 mm from the seed. Thus, the absence of a fraction of backscatter has a relatively larger effect on dose rate as distance from the seed is increased.

One can construct a model to calculate from the present measurements the fraction of scatter that originates in the backscatter block of Fig. 1. This is best illustrated by a numerical example. At 10 mm the dose rate component from unscattered photons is 64% in a full scatter water phantom according to Dale." In the acrylic phantom, at the 1 0-mm measurement point, direct photons traverse an average of approximately 3.0 mm of LiF for an additional attenuation factor of approximately 0. 844 compared to water (8). Thus, the fractional primary dose rate component is reduced to 60%. The remaining 40% is scatter. In the absence of the backscatter block, the dose rate was measured to be 89.0% of the full scatter dose rate. The ratio of the missing 11%, which originates in the backscatter block, to the total scatter is 11% divided by 40%, or 27.5% of the total scatter dose. Similar calculations at other depths show that the dose fraction is roughly constant at 28% ±4% in the region between 7 and 18 mm.


Compared to full scatter, a gold ophthalmic plaque results in a dose rate decrease from an I-125 model 6711 seed transverse to the seed axis. The dose rates are nearly the same at 2 mm distance, and the gold plaque dose rate gradually decreases relative to full scatter, being about 4% less at 5 mm, 7% less at 10 mm and 10% less at 18 mm. This result is in quantitative agreement with the average value found by Weaver (5) for an I-125 model 6702 seed. Actual dose rate decrease may vary somewhat with gold plaque design and particular seed location within the plaque. In clinical calculations for gold plaques with I-125, one should incorporate dose rate reductions relative to full scatter as found from these measurements and those of Weaver (5).


  1. H. B. Stallard, "Radiotherapy for malignant melanoma of the choroid," Br. J. Ophthalmol. 50, 147 (1966).
  2. J. A. Shields, J. J. Augsburger, L. W. Brady, and J. L. Day, "Cobalt plaque therapy of posterior uveal melanomas," Ophthalmology 89,1201 (1982).
  3. S. Packer, M. Rotman, and P. Salanitro, "Iodine- 125 irradiation of choroidal melanoma. Clinical experience," Ophthalmology 91, 1700 (1984).
  4. P. K. Lommatzsch, "Beta irradiation of choroidal melanoma with "'Ru/ 106 Rh applicators," Arch. Ophthalmol. 101, 713 (1983).
  5. K. A. Weaver, "The dosimetry of I-125 seed eye plaques," Med. Phys. 13,78 (1986).
  6. G. Luxton, M. A. Astrahan, P. E. Liggett, D. L. Neblett, D. M. Cohen, and Z. Petrovich, "Dosimetric calculations and measurements of gold plaque ophthalmic irradiators using iridium-192 and iodine-125 seeds," Int. J. Radiat. Oncol. Biol. Phys. (to be published July 1988).
  7. C. C. Ling, E. D. Yorke, I. J. Spiro, D. Kubiatowicz, and D. Bennett, "Physical dosimetry of I-125 seeds of a new design for interstitial implant," Int. J. Radiat. Oncol. Biol. Phys. 9, 1747 (1983).
  8. H. E. Johns and J. C. Cunningham, The Physics of Radiology, 4th ed. (Thomas, Springfield, IL, 1983), p. 723.
  9. H. Kubo, "Determination of the half-life of I-125 encapsulated clinical seeds using a Si(Li) detector," Med. Phys. 10, 889 (1983).
  10. R. G. Dale, "Some theoretical derivations relating to the tissue dosimetry of brachytherapy nuclides, with particular reference to iodine-125," Med. Phys. 10, 176 (1983).
  11. J. A. Bearden and J. S. Thomsen, "X-ray wavelengths and atomic energy levels," in American Institute of Physics Handbook, 3rd ed. (McGraw Hill, New York, 1972), p. 7-96.

Plaque Simulator References | Guide Contents