Physics of Ophthalmic Plaque Brachytherapy

Transcript of a lecture from the late 1980s

(the beginning of Plaque Simulator)

Gary Luxton Ph.D and Melvin A. Astrahan Ph.D.
Department of Radiation Oncology
University of Southern California Keck School of Medicine
Los Angeles, California.
Prof. Luxton is now emeritus at Stanford University
Prof. Astrahan is now emeritus at USC

ABSTRACT

Ophthalmic plaque brachytherapy is a widely used treatment for ocular melanoma, offering a possibility of cure with retention of useful vision. At present, the most commonly used isotope for these procedures is I-125 in seed form. It is potentially of great clinical importance in ophthalmic radioactive plaque therapy to accurately model dosimetry in the close proximity of a seed, but dosimetry in plaque therapy is complicated by the low energy of I-125 photons. Dosimetric uncertainties remain, such as in the value of the specific dose constant for the different model I-125 seeds in water, and in the effects of the materials used in the plaque construction. There are complexities associated with the anisotropy of the dose distribution from single seeds, and with the dependence of dose on tissue type. With a known specific dose constant and with seeds calibrated for air kerma rate, dose rate to the apex of an eye tumor can be found with sufficient accuracy from a point source model calculation. Plaque backscatter and seed carrier transmission factors should be taken into account. Clinical dosimetry does not presently take into account the dependence on tissue type, although this dependence is substantial. Dose to critical structures, particularly near the edge of the plaque, require an anisotropic source model calculation, together with information on how the plaque is positioned three-dimensionally on a model of the eye. Exploitation of seed anisotropy and the shielding afforded by the rim of a plaque to reduce dose to the macula and optic nerve is described.



1.Ocular anatomy

The eye (Figs. 1,2) is a complex organ, nearly spherical in shape, with an anterior to posterior external length between 22 and 27 mm. Surrounding the anterior pole there is a transparent protuberance, 0.5 - 0.7 mm. thick, called the cornea. The external surface of the cornea has a radius of curvature of about 8 nun, and terminates posteriorly in a circle with diameter of about 12 mm referred to as the corneoscleral limbus, which joins the cornea to the sclera (1).

Figure 1.1.1 The standard ocular geometry.

Figure 1.1.2. Anatomy of the eye, showing choroid, optic disk and macula.

The sclera is a dense, opaque, fibrous structure which is perforated posteriorly by the optic nerve and by arteries supplying blood to the uveal tract. It covers the posterior five-sixths of the eye. Maximum scleral thickness is estimated to be 1.0 mm. The uveal tract is the middle coat of the eye, a thin (0.1 - 0.25 mm) layer between the sclera and the retina. The uvea is the pigmented, vascular area of the eye comprising the iris, ciliary body and choroid. The choroid, the site of ocular melanoma, is a vascular layer of the eye located between the sclera and the retina, extending from the edge of the optic nerve posteriorly to the ciliary body anteriorly.

The fovea centralis, or fovea, is a 1.5 mm diameter capillary free thinned area of the retina located about 0.8 mm below the horizontal meridian at the posterior pole, about 3 mm temporal to the edge of the optic disk (Fig. 1.1.3). The photoreceptors of the fovea are exclusively cones, and the fovea, which is centered in a region of about 1.5 mm in diameter known as the macula, is the site of the most acute color vision.

Figure 1.1.3 The normal fundus


2. Uveal melanoma

Melanoma of the uvea accounts for approximately 80% of all primary eye tumors (2). Surgical treatment for uveal melanoma consists of enucleation of the eye if the tumor thickness is greater than 2 - 3 mm. Enucleation is often preceeded by external beam xray radiotherapy of the whole eye and orbit. For small tumors, a conservative treatment involving only a part of the eye may be applied by either heavy-particle ion beam (3,4,5) or radioactive plaque brachytherapy (6,7,8). Radioactive plaque therapy is an attractive alternative to enucleation for primary treatment of melanoma of the eye. Compared to surgery, radioactive plaque brachytherapy appears to offer equally good survival with the substantial added benefit of preservation of the eye (9).


2. Treatment Toxicity

Complications subsequent to all forms of ocular radiation therapy may take several years to become apparent, but have occurred in a high percentage of cases. With a mean followup of 38 months Packer et al. (10) report 34% complications following I-125 plaque therapy, while Lean et al. (11) report 61% complications with 39 month followup. In an 18-year study, Davidorf et al. (12) report complications in 87% (13/15) of the patients whose tumors were controlled with radon seed therapy, while 35% (8/23) required subsequent enucleation. Complications include retinopathy, vitreous hemorrhage, perimacular exudate, neovascularization, cataract and glaucoma. Visual acuity usually decreases as a result of plaque therapy. In the recent study by Lean et al. (11), some loss of visual acuity occurred in 75% of patients treated by radioactive plaque therapy. The likelihood of radiation complications is dosedependent, and the need for precise detailed dosimetry is clear.

Primary malignancy of the eye is usually discovered while the lesion is still quite small compared to tumors at other sites. Maximum basal chord length is usually less than 15 mm at diagnosis. Tumor height is often 5 mm or less. The small size of ocular melanoma allows successful tumor destruction with the highly localized fields of episcleral radioactive plaque therapy. The radiation field must be precisely positioned, and the location of sensitive ocular structures, such as the fovea centralis, optic nerve and lens must be identified relative to the treatment field. There are major questions of clinical interest that require input from dosimetry. Specifically, under what circumstances can vision be preserved? What is the relationship between loss of vision and dose to the macula, optic nerve and lens? What are the dose-response relationships for complications such as retinopathy, optic nerve atrophy and cataract? The important clinical question of long-term survival in radioactive plaque therapy as compared to surgery is the focus of a randomized multi-institutional trial, called the Collaborative Ocular Melanoma Study (COMS) (13).

The small size of the target tissue in ophthalmic plaque therapy evokes dosimetry questions of clinical interest regarding tissue dose in close proximity to individual seeds. I-125 seeds are most commonly used. Surprisingly, perhaps, the question of dose rate close to a calibrated I-125 seed as used in an eye plaque, and indeed, the more basic question of dose rate in an infinite water medium from an exposure-calibrated I-125 seed are still somewhat unsettled. Recent work has resulted in new information, and hopefully both accurate dosimetry and consensus are near. Questions of dosimetry accuracy assume added relevance in ophthalmic plaque therapy, considering the high reported complication rate from radioactive plaque treatments due to late effects from radiation injury.


2. OPHTHALMIC PLAQUE BRACHYTHERAPY TECHNIQUE


1. Radioisotopes for episcleral plaque therapy

The first applications of ophthalmic brachytherapy employed a technique of permanent implantation of single radon seeds (14). The use of removable episcleral radioactive plaque therapy as an alternative to enucleation however, was developed by Stallard (15,16), who originated a set of plaques that conformed to the eye. The Stallard plaques, which are commercially available (The Radiochemical Centre Ltd, Arnersharn , England), are circular and semi-circular in cross-section (Fig. 2.2.1) containing Co-60 foil in a ring pattern sheathed in platinum to absorb the 0.3 MeV beta radiation. The 60CO-foil plaque technique has been adopted in several centers in the United States (6,17,18).

Figure 2.1.1. Stallard-type Co-60 foil plaques, showing a cut-away view, and two plaque designs. The semi-circular plaque is notched to accomodate placement of the plaque next to the optic disk.

A variety of other radioisotopes have been used for brachytherapy of eye tumors. Lommatzsch (19) has pioneered the use of Ru-106/Rh-106 beta sources for melanomas less than 5 norn in height and Chenery, Japp and Fitzpatrick (20) introduced permanent implant grains of 198Au. Luxton et al. (8) described the use of removable 1921r seed plaques. The most widely-used isotope for radioactive plaque therapy other than the 60CO-foil plaque, however, has been the I-125 seed plaque, first discussed by Packer and Rotman (7). I-125 is also the isotope designated for use by the COMS (13).

The use of I-125 seed plaques has several advantages compared to Co-60 and Ir-192, most notably, greater ease of radiation protection due to the low energy of I-125 photons (28.5 keV average). Also, there is 20-25% less dose to the opposite surface of the eye for the same dose to the tumor apex (8), and a high degree of customization in treatment planning is possible through selection of seed strengths, seed 'Orientations and differential plaque loading (21). The plaque is an effective shield for the tissue on the concave side of the plaque away from the tumor. The use of a plaque design with a collar at the edge (Fig. 5) provides a degree of collimation that is effective for I-125 seeds within 1-2 mm of the edge (21), and can, be exploited to reduce dose to the macula without changing the dose to the tumor. The major disadvantage to the use of I-125 seed plaques is the complexity of dose calculation. Also, seeds need to be replenished on a regular basis, owing to the 59.6 day half-life (22).


2. Tumor assessment; plaque and seed strength selection.

Plaques (Fig. 2.2-1) are selected to cover the basal dimensions of the tumor with a margin of 1-3 mm. The requirement of a margin can be relaxed somewhat if the plaque is designed with no collar at the edge, an important consideration for tumors located close to the optic nerve The required activity of radio active material is calculated from the height of the tumor or other dose prescription point, based on prescribed treatment goals of total dose and dose rate. A point source model for seeds is adequate for this purpose (8). Tumor height is measured with ultrasound, although it can also be obtained from thin-section CT or MRI. Ultrasound gives the height to a precision of 0.1 mm. Basal dimensions are obtained from indirect ophthalmoscopy. This is a procedure of viewing light reflected from the fundus (the posterior portion of the interior of the eye) through a hole in a mirror used to illuminate the dilated pupil (the aperture through the iris), using a high power lens to form a real image. Melanoma appears as a pigmented area against a red background. Image magnification is not uniform over a 2-D image of the fundus, however, since the fundus is a curved surface that approximates the inside of a sphere. Nevertheless, the outline of the tumor is often defined by a drawing on a fundus diagram (Fig. 2.2.2). The fundus diagram shows the optic disk and posterior pole, and displays the equator and ora serrata (the anterior margin of the retina) as concentric circles.

Figure 2.2.1 Examples of radioactive seed plaques, showing notched, circular, and rim-type COMS plaques with carrier.

For tumors that are anteriorly located, basal dimensions can be identified by a combination of direct and indirect illumination. Margins can be marked on the sclera and measured with calipers. Tumor base dimensions can also be estimated from MRI and CT.


3. Plaque construction.

Plaques of different designs have been developed by various groups (7,8,14,15,16). The plaque itself generally consists of a section of a metallic spherical shell, usually made of gold alloy. The shell contains the seeds, either directly on its surface, or in a closefitting carrier that provides a separation between the seeds and the sclera. The plaque may have a collar (lip) of 2 - 3 mm height around it for the purpose of adding collimation to the dose distribution near the plaque edge. I-125 seeds are usually glued or press-fit into pre-defined locations. Thus, the plaque assembly has definite slots intended to receive the I-125 seeds, and thereby to also specify position coordinates and orientation for each seed. The carrier may be acrylic (23,24) or "silastic" (silicon-based plastic) (13), or may consist of grooves cast into the gold shell itself (8). There are suture eyelets at the outer perimeter of the plaque. The eyelets can be positioned at defined distances from visible structures, and used to reference the seed coordinates to the anatomy of the eye.

Figure 2.2.2. Fundus diagram showing a 10 mm diameter tumor near the equator.

The plaque is sutured to the sclera. The dose rate is calculated to the apex of the tumor. Based on empirical experience, Stallard (3) recommended at least 80- 100 Gy total dose to the apex. Consistent with the practice of brachytherapy at other tumor sites, dose rate to the apex is generally planned to be >= 40 cGy/h.


3. I-125 PLAQUE DOSIMETRY

Accurate calculation of dose rate to the structures of the eye require the following: (i) dosimetry factors for calculating tissue dose from a single calibrated seed (ii) seed position location and orientation within the plaque, (iii) plaque orientation and positioning with respect to the tumor and critical eye structures and (iv) seed calibration. This section will address item (i).


1. Point-source model calculation of treatment time.

Dose rate calculation is simply a matter of addition of dose rate from each of the seeds at each point of interest. The dose distribution from each of the different design I-125 seed (Models 6711 and 6702, 3M Corp., Medical Products Division, St. Paul, MN) (Fig. 3.1.1) is sharply anisotropic (25,26,27), the anisotropy is most pronounced along directions that are close to parallel with the seed axis (Fig. 3.1.2), and seeds can be treated as point sources for calculations at points nearly transverse to the seeds. In particular, to estimate the total activity required, which is based on dose rate to the tumor apex, a point source calculation is sufficient. The angular dependence of the dose distribution from a single seed varies slowly with polar angle for a point in directions within 30° of perpendicularity to the seed axis. According to the data of Ling et al. (26) for model 6711 and Schell et al. (27) for model 6702, the angular anisotropy correction factor varies between 1.0 and 0.95 for angles between 0 and 30 degrees of the transverse direction and for distances between 0.5 and 3 cm. For a tumor apex 5 mm deep within the eye along the plaque central axis, the lines joining the apex of the tumor to the centers of the nearest seeds are inclined at relatively small angles to the perpendicular to the aids of the individual seeds, since the seeds are tangential to the eye. Thus, the estimate of dose rate at the apex of the tumor (normally taken to be along the plaque central axis) can assume an angular factor of 1.0 for the various seeds. Beyond 5 mm from the center of the plaque along the central axis, there is little sensitivity to the extended nature of the seed sources (8).

Figure 3.1.1. Current models of I-125 seeds: model 6711 (silver wire and model 6702 (three resin spheres).

Fig. 3.1.2 Polar plot of r² * D(r,θ) from ref. 27.


2. Anisotropic model calculations.

An extended source, i.e., anisotropic, model calculation is required for precise calculation of dose to critical structures. This is particularly true for a posteriorly placed plaque, since the fovea and optic nerve are near the surface of the eye, and the direction to a nearby seed may form only a. small angle with the seed axis. For example, it may be possible to orient the carrier so that the long axis of the seed nearest the macula is aligned with its axis pointing toward the fovea. In such a case, the dose rate at the fovea (center of the macula) can be 50% or less of what would be calculated from a point source model. The anisotropy of the I-125 seed is due to the extended nature of the physical source and to the strong self-absorption of the radiations, due to the low energy of the x and gamma rays. Treating the seed as an unfiltered line source will not explain the angular dependence of the dose rate, since the major effect contributing to the anisotropy is radiation self-absorption. Nevertheless, adopting a model based on the linear shape of the source, we write the dose rate from a single seed as

D(x,y) = S * g(r) * Ao * (1/y) * (1/L) * B(r) * T * F(θ) * ([arctan((2x+L)/2y) - arctan((2x-L)/2y)] / r²);

where

D(x,y)
is the dose rate (Gy/h)
y
is the perpendicular distance (cm) from the line source of length L
x
is the parallel distance along the length of the source (cm)
S
is the source activity expressed in air kerma strength (Gy-cm²/h)
g(r)
is a dimensionless scatter-attenuation factor, a function of the radial distance r = sqrt(x² + y²), normalized such that g(1) = 1 at r = 1 cm (g(r) is analogous to the quantity introduced by Dale (28) for a point source)
Ao
the specific dose constant, is the dose rate at 1 cm in water per unit air kerma rate
B(r)
is a factor that represents the effect of the (gold) backing material as compared to a full-scatter water phantom
F(θ)
is an anisotropy factor as a function of angle θ = arctan(|x|/r)
T
is a factor that represents the transmission through the carrier material relative to water.

T is unity if the carrier is water equivalent. In principle, F(θ) can be a function of r as well, and B(r) can depend on both plaque geometry and seed location within the plaque. The first reported measurement of the effect of gold backing, i.e., B(r), was by Weaver (29) who, using slab geometry in polystyrene with a model 6702 seed, found an average decrease in dose rate of 8% relative to full scatter in the 5 - 10 mm range.


3. Measured values for the seed factors.

Values for g(r) and B(r) measured along the direction transverse to the seed within 2 cm of the seed axis are given in Table 1 for a model 6711 seed (30-35). For the effect of the backing material relative to the full-scatter geometry, Wu et al. (35) find a value greater than 1 for B(y) for all distances between 1 and 5 mrn. There are several differences between the measurement setup of Wu et al (35) and that of Luxton et al (34). Perhaps the most notable is that Wu et al employ a slab geometry, while Luxton et al use a gold plaque on a hemisphere embedded in a forward-scatter block. Thus in Wu et al. there is more forward-hemi sphere tissue phantom scatter for the goldbacking case, which may account for the differences. The "gold" is actually only 18-karat gold alloy (75% gold) in both cases, with the balance being about two-thirds silver and one-third copper. According to measurements and calculations by Cygler et al. (36), the effect of pure silver backing would be greater by about a factor of 3. Cygler et al. who use a diode detector and a model 6702 seed, also find an enhanced response up to 5 mm from the seed in a gold-backed slab geometry, but a decrease below the full-scatter geometry between 5 mm and 25 mm transverse to the seed. By contrast, Harnett and Thomson (37) state that they find no effect from the gold, but details of the geometric setup are sketchy, and they do not show their data. Cygler et al.(36) are careful to distinguish their relative response from a true relative dose rate because of variations in diode response per unit dose for different energy. The goldscatter effect, due to L-shell fluorescent x-rays between 9 and 14 keV (38), is geometrydependent, and will undoubtedly be different for a seed placed peripherally than for a seed on the central aids. For now, one can choose whether or not to incorporate a B(r) type of factor different from 1, since the effect is <10%, and complex to calculate. Our preference is to use the values in the third column of Table 1 augmenting these with a value of 1.05 at 1 mm, and 1.0 at 3 mm, for the curved geometry of the ophthalmic plaque. This is plotted in Figure 3.3.1.

Table I. Measured dosimetry factors for model 6711 seed.
r(cm)
g(r)1
g(r)2
B(r)3
B(r)4
0.1
1.065 ±.032
0.2
1.017 ±.032
1.045 ±.035
0.3
1.075 ±.035
1.026 ±.047
0.5
1.039 ±.034
1.027 ±.046
0.964 ±.035
1.021 ±.022
0.74
0.943 ±.019
1.0
1.0
1.0
0.930 ±.020
0.997 ±.005
1.38
0.916 ±.020
1.77
0.904 ±.025
2.0
0.808 ±.030
3.0
0.622 ±.023

1Ref. 30, 2Ref. 31, 3Ref. 34, 4Ref. 35

Figure 3.3.1. The gold-plaque backscatter factor relative to fullscatter phantom. Points with error bars are from Ref. 34.

Concerning the parametrization of Eq. (1), the anisotropy factor F(O) must be dependent on r also, at some level of precision. In this context, we have reviewed the fullscatter measurements of Ling et al. (26) and Schell et al. (27) for the model 6711 and 6702 seeds, respectively. For both model seeds, the available data exhibit only weak dependence of the quantity F(O) on the radial variable r for values of r up to 2.5 cm, the maximum value of dosimetric interest in ophthalmic plaque therapy. The results, expressed in terms of F(O), are given in Table II. Notice that to obtain these values from the published data, it was necessary to evaluate the geometric factor present in Eq Angle (0 ) 0' (transverse) 300 500 700 900(axial) Model 6702 1.000 .956 .930 .785 .533 Model 6711 1.000 .959 .892 .682 .409 Table II. Mean values of anisotropy factors, F(O) from Eq. (1), obtained from data of Refs. 26 and 27.

Deviations from the means given in Table II for the measured points are less than ±2% for 0 !9 500, within about ±5% for the 0 =700, and within +10% for the 0 =900 (axial) data. The error bars shown in Refs. 26 and 27 for the axial data are about ±10%. From these observations we conclude thatr the parametrization of Eq. (1) is in agreement with existing experimental data for the full-scatter phantom.

Chiu-Tsao et al. (39) have found that for the silicon-based plastic "silastic" (Dow-Corning, material MDX 4-4210, polymethyl siloxane fluid, cures at room temperature in 24 hours, specific gravity - 1.12.) material used as seed carriers in the COMS plaques, T is about 0.90 along the central axis. We have confirmed this result using an exposure rate meter to compare measurements with and without a seed carrier. The seed carrier has a thickness of only 1 mm between the bottom of the seed and the eye, but the low energy of the I-125 radiations does lead to a substantial effect from this small thickness of low-tomedium atomic number material.


4. Specific dose constant Ao

The specific dose constant has been recently the subject of several measurements and calculations. Measurements in PMMA (acrylic, lucite) corrected to water (30) and solid water (31,32,33) have found that the dose constant for both models of I-125 seeds are 10-20% less than what has been until recently, the most commonly accepted values.

Table III. Specific dose constant, Ao 6711 1.032 .932 ±.020 .909 .832 ±.03 .853 .85 ±.03 6702 .962 .929 .932 .90 ±.03 Reference 25, calculation 30, measurement 40, calculation 31, measurement 32, measurement 33, measurement

Recent values for Ao for models 6711 and 6702 are given in Table III. In an appendix to our paper (30), we describe a model to correct for the effect of the measurement phantom differing from water. Applying this to the uncorrected measurements in "solid water" (41) of Refs. 31,32,33 in Table III would increase these values of A. by 3.5% (30).


5. I-125 Dose Constant and seed calibrations.

The quantity Ao in eq. (1) is the dose rate in water from an I-125 seed per unit air kerma rate, the calibration recommended by AAPM Report no. 21 (42). Suggested unit for air kerma strength is gGy m² h-1. Water is chosen because it is a standard tissue-like material. The specific dose constants for an I-125 point source in eye tissues, however, are 10% - 15% less than in water, according to Monte Carlo calculations by Chiu-Tsao et al. (43). At present, these differences are not normally taken into account in clinical calculations.

The unit presently used by the manufacturer to specify seed calibration is "apparent activity in milliCuries". Apparent activity refers to the activity of a hypothetical point source of pure I-125 that would produce the same air kerma rate at a reference distance. Air kerma rate is simply related to exposure through the W-value, the average energy absorbed per ion pair in dry air:

Air kerma rate = Exposure rate x W (2)

Experimentally, W is independent of photon energy at 33.97 JIC (44), corresponding to an air kerma per Roentgen of 0.876 cGy/R. The exposure rate constant (symbol n for a hypothetical point source is conventionally taken by the manufacturer to be 1.45 R-cm² AmCi-h) following Krishnaswamy (45), whose value was also adopted by Ling et al. (25). This value is close to the 1.464 R-cm²/(mCi-h) calculated by Schulz, Chandra and Nath (46) for a point source of pure I-125, but differs somewhat from the 1.59 Rcm2 AmCi-h) calculated by Hashemi et al. (47) for a spectrum modified to take into account the Kemissions of silver. It is important to realize, howev er, that the exposure rate constant is used by the manufacturer merely to define the unit of apparent activity based on a comparative measurement ultimately referenced to standards laboratory measurement of air kerma rate. The actual number used for the exposure rate constant has no effect on dosimetry, and the conventional value of 1.45 R-cm²/(MCi-h) will therefore be used here when referring to the unit of apparent activity. Thus, air kerma rate per unit apparent activity is 1.270 cGy-cm²/(mCi-h). To obtain dose rate to an infinitesimal mass of water, the air kerma rate would be multiplied by the ratio of mass energy absorption coefficients. Using standard tables for a 27.5 keV photon (48), this ratio is 1.012, corresponding to 1.286 cGy cm²/(mCi-h) dose rate per unit of apparent activity [0.9655 x 10-11 cGy cm²/Bq-sl. To obtain the more-familiar historical dose rate constant defined as dose rate to water at a depth of 1 cm per unit of apparent activity, one must multiply AO by 1.270:

Dose rate (water, 1 cm) = A. 0 1.270 (cGy/h) per mCi apparent ... (3)

The concept of apparent activity implies knowledge of an air kerma rate from the seed, and for this reason it is a more practical concept than intrinsic activity. The relationship between the intrinsic activity and air kerma rate depends on details of seed construction and on uncertain calculations. Taking into account the absorption by 0.05 mrn Ti walls, Chiu-Tsao et al. (43) find an effective activity of 85% of the true activity for a point source. For the silver-wire seed, 50% of emitted photons are directed away from the wall and into the silver, resulting in few escaped photons to contribute to the apparent activity. Thus the apparent activity of model 6711 seeds is of the order of 50% of the actual activity (The manufacturer states that there is a total self absorption of approximately 35%, but the quantitative basis for this statement is not clear.). This is confirmed in Monte Carlo calculations by Chiu-Tsao et al. (32), who find an apparent activity of 54% of actual activity for model 6711 seeds.


6. Dependence of calibration factor on seed design.

The specific dose constant for I-125 seeds depends on seed design, even for air kermacalibrated seeds. This occurs because the exact energy spectrum for encapsulated seeds depends upon the seed design, and low-energy photons are strongly attenuated in water. The relationship between dose rate in water and kerma rate in air is not that of pure I-125, nor is it the same for the different seed models. Two models are in present use: 6711 silver-wire and 6702 ion-exchange resin spheres (with no gold marker), both are encapsulated in 0.05 mm thick titanium walls. Low-energy photons of 22 and 25 keV arise from K-shell fluorescence of the the silver substrate for model 6711 seeds (25) and from the Ti encapsulation for both 6711 and 6702 model seeds. The 4.5-keV characteristic L-to-K shell transition Ti x rays give detectable contribution to the exposure rate in air (49), but are 95% attenuated in 0.05 cm. of water. Experimental details of the calibration procedure can also affect the assignment of air kerma strength by the standards laboratory. Williamson performed a Monte Carlo simulation (40) of the I-125 seed calibration procedure of the national standards laboratory which used a (wall-less) free-air ionization chamber. He found that Ti x rays increased the estimate of apparent activity by 5% for both models 6711 and 6702 singleseed calibrations. He also found that A0 calculated from the Monte Carlo for the model 6702 seed was 5% larger than that from the model 6711 (silver wire). The difference between the two energy spectra is that the silver wire model contributes additional Ag K shell x-rays with a mean energy (average 22.7 keV) less than those from pure I-125 (average 28.5 keV). The additional photons contribute a smaller dose in water:at 1 cm per unit air kerma. This implies a smaller specific dose constant of the silverwire seed as compared to the model 6702, whose energy spectrum is closer to pure I-125, in agreement with the measurements of the NCI-contract group (31,32,33).


4. 3-D TREATMENT PLANNING AND OPTIMIZATION

The objective of plaque therapy for ocular tumors is to control the tum - - while retaining as much useful vision as possible. At present, the prognosis for controlling small and medium sized tumors is excellent. Unfortunately, the high incidence of late, radiation-induced, vision-limiting complications continues to pose a problem. This situation suggests that dose to uninvolved, visually critical structures within the eye such as the macula (or fovea), optic nerve and lens should be precisely determined and optimized whenever possible. In addition, surgical placement must be accurately preplanned, executed and documented. Optimization must be based on a full 3D dose calculation.

Fig. 4. 1. 1 Since the tumor apex is approximately normal to the longitudinal axis of a seed. the source can be rotated to reduce dose to a selected site such as the macula without compromising the tumor dose.

Dose to normal structures adjacent to a plaque can be reduced without compromising the tumor dose by taIdng advantage of the anisotropic dose distribution around 1-125 sources. The left side of figure 4.1.1 illustrates an arbitrary isodose contour for a source pointing directly at the macula. The macula falls outside this contour. When the source is rotated 90', dose to the macula increases.

Figure 4.1.2 Shielding by the lip of a plaque is only effective for sources very near the lip. Sources more than about 2 nun from the lip irradiate the enitire eye.

Collimation provided by the gold shell reduces the dose to portions of the orbit and will also shield adjacent ocular structures when low energy sources such as 1-125 are used. When the sources are placed very close to the plaque's lip, shielding of adjacent ocular structures can (potentially) be very effective as illustrated in figure 4.1.2. The effectiveness is greatest when high intensity sources are positioned close to, and approximately parallel to the lip. Shielding is of little value when the sources lie more than about 2 mm from the lip, and is marginal for radially oriented sources.

The dose distribution can also be tailored to the three dimensional tumor volume by using a non-uniform distribution of source activity in the plaque (Figure 4.1.3). This can be accomplished by leaving selected slots empty and/or by using sources of differing activity. Loading a plaque with sources of unequal activities, however, is a potential source of confusion.

Figure 4.1.3 Asymmetrical loading of the plaque results in an asymetrical dose distribution.

Precise criteria for dose optimization are difficult to establish, since they involve both radiobiological as well as physical considerations. The primary reason that we are interested in optimization is to reduce late, vision limiting complications. This must be accomplished, however, without compromising tumor control. The most important visual structures with regard to dose optimization are the macula, fovea and optic nerve. The lens is no longer considered an important optimization site since cataracts can be surgically corrected by intraocular lens implanation. Reducing dose to these critical structures is expected to reduce late radiation induced complications. Unfortunately, for tumors which underly the macula or fovea, there is little that can be done in terms of optimization. Many tumors, however, do not directly involve these critical structures, but may lie a few millimeters adjacent to them. Dose optimization is possible and desirable for these tumors. The goal of dose optimization is to find a source arrangement for which dose to a designated point (or points) within the tumor volume falls within a specified window, while dose to designated critical points is either minimized or kept below a specified tolerance limit.

To implement dose optimization we have developed an interactive treatment planning system for episcleral plaque therapy (21). The program uses CT to generate a three dimensional model of the eye based on patient specific anatomy (such as a myopic eye). The tumor perimeter is digitized directly from fundus photography. The program employs a real time, highly interactive 3D color-graphic display, and operates on an inexpensive Macintosh IIx microcomputer workstation. The program currently supports 1921r and I-125 sources, and accounts for isotope decay, source anisotropy, and scatter dose modification due to the gold plaque as in eq.(1), and also collimation provided by the gold plaque shell.

Our present optimization strategy is based on the hypothesis that reducing dose to the macula (M) without compromising tumor dose will reduce late, vision degrading complications compared to present practice without affecting tumor response rates. M is calculated at the posterior pole, close to the fovea. Dose to the tumor apex M is presumed to represent the minimum tumor dose. The ratio T:M is used as a practical figure of merit, with the goal of optimization being to maximize the T:M ratio. Values for T, M and T:M are displayed as each source is loaded (or unloaded), rotated, moved within the plaque, or the entire plaque is moved and rotated. In this way, the parameters of distance, activity distribution, source orientation, plaque orientation, and shielding can be interactively manipulated to improve the T:M ratio. For COMS plaques, the carrier can also be rotated within the plaque shell.

Example. A patient with a 5 mm tall posteriorly located tumor is to be treated using a COMS 16 mm diameter plaque. A dose of 100 Gy to the tumor apex with at least 2 rnrn retinal margins is prescribed. In figure 4.1.4 the plaque has been loaded somewhat arbitrarily with 9 sources to deliver 100 Gy to the apex of the tumor in about 7 days. No attempt to optimize this treatment has been included. In this calculation, the I-125 seeds were considered to be isotropic point sources and lip collimation was ignored. The T:M ratio was calculated to be 1.29.

One simple way to optimize this treatment is to be more conscious of which slots in the carrier will be loaded, and the orientation of the carrier within the gold shell. In this way, the isodose contours on the retinal surface will more closely conform to the perimeter of the tumor. Figure 4.1.5a illustrates an alternate loading pattern.

The seed arrangement of 4.1.4 can be improved by moving two peripheral seeds to central slots directed toward the macula, and moving the central seed to a slot more distal to the macula. The new source arrangement is depicted in Fig. 4.1.5a. With this source arrangement isodose contours on the retinal surface above the plaque more closely approximate the tumor perimeter and three of the seeds are a bit further from the macula. With the seeds considered as isotropic point sources and collimation ignored the T:M ratio was calculated to be 1.51. This simple rearrangement reduced the calculated macular dose by 15% without compromising the tumor dose or prescribed margins. With the seeds considered as anisotropic line sources and lip collimation included, the T:M ratio was calculated to be 2.82. Figures 4.1.5a and b are the crosssectional and retinal dose distributions for this seed arrangement.


5. QUALITY ASSURANCE

We define quality assurance in episcleral plaque therapy (EPBT) as a program of testing the correctness of the treatment plan and its implementation. One significant issue, not previously mentioned here, is verification of correct plaque placement. Houdek et al. (50) have reported progress in the development of an MRI technique for verification of correct plaque placement.

The following is a partial list of items routinely checked in EPBT, and the method of their testing.


1. Tumor base dimensions and plaque coverage

Transillumination is performed under anesthesia in the operating room to verify tumor coverage by the plaque. A "cold", i.e., non-radioactive plaque is sutured to the sclera to position plaque without radiation exposure.


2. Seed activity verification and uniformity of activity within a batch

Seeds are individually assayed in a calibrated well-type ionization chamber. The ion chamber is calibrated for I-125 by recording the response to an individually-calibrated seed of the same type that was obtained from the manufacturer. A single calibrated seed may be used over a period of about 3 T1/2 (180 days) to maintain the calibration of the ionization chamber.


3. Seed integrity

Seeds are inspected visually for damage prior to each use. If seeds are being re-used, they should be leak-tested. Seeds are certified as leak-tested by the manufacturer prior to shipment. Leak-testing consists of counting removable activity in a small quantity of scintillation fluid, alcohol or acetone. Removable activity must be less than 0.005 gCi. This may be verified by liquid scintillation counting (more sensitive) or by using a lowbackground thinwindow Geiger.


4. Plaque sterilization

Steam autoclave at 150'C. for 5 min., if no plastic seed carrier is used. Otherwise, chemical and gas sterilization is required, followed by aeration for 12 - 24 hours. For gas sterilization, a porous shielded container should be used that permits gas circulation while providing radiation shielding.


5. Tumor apex dose rate

Calculation should be done by hand using a point source model, to verify computer calculation.


6. Radiation safety

A long list of procedures is followed, including the wearing of gas-sterilized finger ring dosimeters by the ophthalmologist and operating room assistant. The radiation oncologist or physicist should supervise the handling of radioactive materials in the operating room. After the eye plaque is placed on the patient, a plastic or metal eye patch wrapped with at least 0.1 mm thickness of lead foil is placed as an external eye bandage. After removal of the patient, the operating room should be surveyed with a thin-window Geiger counter to verify that no sources have fallen out of the plaque. The patient should be surveyed to verify that maximum exposure rate 1 meter away is less than 2 mR/h (20 gSv/h) The patient-should be labelled. A "caution - radioactive materials" label on a plastic hospital bracelet is sufficient.

If the patient is to be temporarily released in accordance with procedures approved by the appropriate regulating agency (N-RC or agreement-state department of health), then the hazards of radioactive materials should be discussed with the patient or his/her caretaker. The patient should be advised of time, shielding and distance factors, that he/she should sleep in a separate bed, and remain confined to his/her dwelling place while the plaque is attached. Also, the patient should be advised against close approach from children or pregnant visitors An information sheet should accompany the patient in case the patient should require further medical attention while the plaque is in place. The information sheet should detail the number of seeds, their activity, the maximum dose rate at one meter, when the plaque will be removed and telephone numbers for the physician and physicist responsible for the patient's care. A copy of the sheet documenting the fact that these matters were discussed with the patient should be retained at the hospital. Upon return for plaque removal, the operating room should be surveyed with a thin-window Geiger counter after the plaque is removed, and the number of seeds counted to verify the return to inventory of all seeds in the plaque.


6. SUMMARY

We have presented the clinical backgound of episcleral plaque brachytherapy (EPBT), I-125 seed plaque design, and pertinent I-125 seed dosimetry. The dosimetry has included specialized factors related to the low energy of I-125 radiations. A 3-D modelling program that incorporates significant structures of the eye, and a realistic I-125 dosimetry algorithm has been described. The program permits interactive optimization of treatment planning on a desktop computer. A concept of treatment optimization has been proposed. While the system represents a significant improvement in customization and documentation of EPBT, its application has been underway only in the last 2-3 years, and long-term consequences are yet to be determined. Further refinemements of EPBT have yet to be implemented, such as dosimetry which includes tissue-type dose dependence and verification of correct plaque placement using MRI.


REFERENCES

1. Newell, FW: Ophthalmology: Principles and Concepts, 4th Ed. Saint Louis, C.V. Moseby Co.; 1978.
2. Chalkley T: Ocular melanoma task force report. Am J Ophthalmol 90: 728-733; 1980.
3. Gragoudas ES, Goitein M, Verhey L, Munzenreider J, Urie M, Suit H, Koehler AM: Proton beam irradiation of uveal melanomas: Results of a 5 1/2 year study. Arch Ophthalmol 100: 928-934; 1982.
4. Saunders WM, Char DH, Quivey JM, Castro JR, Chen GTY, Collier JM, Cartiguy A, Blakeley EA, Lyman JT, Zink SR, Tobias CA: Precision, high dose radiotherapy: Helium ion treatment of uveal melanoma. Int J Radiat Oncol Biol Phys 11: 227-233; 1985.
5. Gragoudas ES, Seddon JM, Egan K, Glynn R, Munzenrider J, Austin-Seymour M, Goitein M, Verhey L, Urie M, Koehler A: Long-term results of proton beam irradiated uveal melanomas. Ophthalmol 94:349-353; 1987.
6. Shields JA, Augsburger JJ, Brady LW, Day JL: Cobalt plaque therapy of posterior uveal melanomas. Ophthalmol 50:1201-1207; 1982.
7. Packer S, Rotman M: Radiotherapy of choroidal melanoma with iodine-125. Ophthalmol 87: 582-590; 1980.
8. Luxton G, Astrahan MA, Liggett PE, Neblett DL, Cohen DM, Petrovich Z: Dosimetric calculations and measurements of gold plaque ophthalmic irradiators using iridium-192 and iodine-125 seeds. Int J Radiat Oncol Biol Phys 15:167-176; 1988.
9. Garretson BR, Robertson DM, Earle JD: Choroidal melanoma treatment with iodine-125 brachytherapy. Arch Ophthalmol 105:1394-1397; 1987.
10. Packer S, Rotman M, Salanitro P: Iodine-125 irradiation of choroidal melanoma: Clinical experience. Ophthalmol 91:1700-1708; 1984.
11. Lean EK, Cohen DM, Liggett PE, Luxton G, Langholz B, Lau R, Astrahan MA, Hyden EC, Petrovich Z: Episcleral radioactive pique therapy: Initial clinical experience with 56 patients. Am J Clin Oncol, in press, 1990.
12. Davidorf FH, Pajka JT, Makley TA, Kartha MK_ Radiotherapy for choroidal melanoma: An 18-year experience with radon. Arch Ophthalmol 105: 352-355; 1987.
13. Earle J, Kline RW, Robertson DM: Selection of iodine-125 for the Collaborative Ocular Melanoma Study. Arch Ophthalmol 105: 763-764; 1987.
14. Moore RF: Choroidal sarcoma treated by the intra-ocular insertion of radon seeds. Br J Ophthalmol 14:145; 1930.
15. Stallard HB: Malignant melanoma of the choroid treated with radioactive applicators. Ann Royal Coll Surg Engl 29:170-182; 1961.
16. Stallard HB: Radiotherapy for malignant melanoma of the choroid. Brit J Ophthalmol 50:147155; 1966.
17. Chan B, Rotman M, Randall GJ: Computerized dosimetry of 60CO ophthalmic applicators. Radiol 103: 705-707; 1972.
18. Char DH, Lonn LI, Margolis LW: Complications of cobalt plaque therapy of choroidal melanomas. Am J Ophthalmol 70:844-851; 1986.
19. Lommatzsch PF, Results after B-irradiation of choroidal melanomas: 20 years'experience. Brit J Ophthalmol 101: 713-717; 1983..
20. Chenery SGA, Japp B, Fitzpatrick PJ: Dosimetry of radioactive gold grains for the treatment of choroidal melanoma. Brit J Radiol 56:415-420; 1983.
21. Astrahan MA, Luxton G, Jozsef G, Kampp TD, Liggett PE, Sapozink MD, Petrovich Z: An interactive treatment planning system for ophthalmic plaque radiotherapy. Int J Radiat Oncol Biol Phys 18: 679-687; 1990.
22. Kubo H: Determination of the half-life of 1-125 encapsulated clinical seeds using a Si(Li) detector. Med Phys 10: 889-891; 1983.
23. Schell MC, Weaver KA, Phillips TC, Char DH, Quivey JM, Barnett C, Ling CC: Design of 1-125 eye plaques for radiotherapy. Endocurieth Hypertherm Oncol In press, 1990.
24. Karolis C, Amies C, Frost RB, Billson FA: The development of a thin stainless steel eye plaque to treat tumours of the eye up to 15 mm in diameter. Australas Phys Eng Sci Med 12:172-177; 1989.
25. Ling CC, Yorke ED, Spiro IJ, Kubiatowicz D, Bennett D: Physical dosimetry of I-125 seeds of a new design for interstitial implant. Int J Radiat Oncol Biol Phys 9:1747-1752; 1983.
26. Ling CC, Schell MC, Yorke ED, Palos BB, Kubiatowicz DO: Two-dimensional dose distribution of I-125 seeds. Med Phys 12: 652-655; 1985.
27. Schell MC, Ling CC, Gromadzki ZC, Working KR: Dose distributions of model 6702 1-125 seeds in water. Int J Radiat Oncol Biol Phys 13: 795-799; 1987.
28. Dale RG: Some theoretical derivations related to the tissue dosimetry of brachytherapy nuclides, with particular reference to iodine-1 25. Med Phys 13: 678-682; 1983.
29. Weaver, KA: The dosimetry of I-125 seed eye plaques. Med Phys 13:78-83; 1986.
30. Luxton G, Astrahan MA, Findley DO, Petrovich Z: Measurement of dose rate from exposurecalibrated 12,51 seeds. In press, Int J Radiat Oncol Phys 18 no. 5, 1990.
31. Weaver KA, Smith V, Huang D, Barnett C, Schell MC, Ling C: Dose parameters of I-125 and 1921r seed sources Med Phvs16: 636-643: 1989.
32. Chiu-Tsao S-T, Anderson LL, O'Brien K Sanna R: Dose rate determination for I-125 seeds. In press, Med Phys 17; 1990.
33. Nath R, Meigooni AS, Meli JA: Dosimetry on the transverse axes of I-125 and 1921r interstitial brachytherapy sources, in press, Med Phys 17; 1990.
34. Luxton G. Astrahan MA. Petrovich Z: Backscatter measurements from a sinele seed of I-125 for ophthalmic plaque dosimetry. Med Phys 15: 397-400; 1988.
35. Wu A, Sternick ES, Muise DJ: Effect of gold shielding on the dosimetry of an I-125 seed at close range. Med Phys 15: 627-628; 1988.
36. Cygler J, Szanto J, Soubra M, Rogers DWO: Effects of gold and silver backings on the dose rate around an I-125 seed. Med Phys 17:172-178; 1990.
37. Harnett AN, Thomson: An Iodine-125 plaque for radiotherapy of the eye: manufacture and dosimetric calculations. Brit J Radiol 61: 835-838; 1988.
38. Bearden JA, Thomsen JS: X-ray wavelengths and atomic energy levels, (In) American Institu f)f Phvsics Handbook 3rd ed. (McGraw- ill New York 1972) n 7-96.
39. Chiu-Tsao S-T, Anderson LL, Stabile L: TLD dosimetry for 1-125 eye plaque. Phys Med Biol 33, Suppl 1:128,1988 (Abstr.)
40. Williamson, JF: Monte Carlo evaluation of specific dose constants in water for I-125 seeds Med Phvs 15:686694; 1988.
41. Constantinou C, Attix FH, Pahwal BR: A solid water phantom for radiotherapy X ray and gamma-ray beam calibrations. Med Phys 9: 436-441; 1982.
42. American Association of Physicists in Medicine (AAPM) Report no. 21, Specification of brachytherapy source strength. New York, NY: Amer Inst of Phys; 1987.
43. Chiu-Tsao S-T, O'Brien K, Sanna R, Tsao HS, Vialotti C, Chang YS, Rotman M, Packer S: Monte Carlo dosimetry for I-125 and 6OCo in eye plaque therapy. Med Phys 13: 678682; 1986.
44. Boutillon M, Perroche-Roux AM: Re-evaluation of the W-value for electrons in dry air. Phys Med Biol 32: 213219; 1987.
45. Krishnaswamy, V: Dose distribution around an I-125 seed source in tissue. Radiol 126:489-491; 1978.
46. Schulz RJ, Chandra P, Nath R: Determination of the exposure rate constant for I-125 using a scintillation detector. Med Pbys 7:355-361; 1980.
47. Hashemi AM, Mills MD, Hogstrom KR, Almond PR: The exposure rate constant for a silver wire I-125 seed. Med Phys 15: 228-34; 1988.
48. Johns H, Cunningham, JR: The Physics of Radiology, 4th Ed. Springfield IL; Thomas Co; 1983.; Appendix A.
49. Kubo,H: Exposure contribution from Ti K x-rays produced in the titanium capsule of the clinical 1-125 seed. Med Phys 12: 215-220; 1985.
50. Houdek PV, Schwade JG, Medina AJ, Poole CA, Olsen KR, Nicholson DH, Byrne S, Quencer R, Hinks RS, Pisciotta V: MR technique for localization and verification procedures in episcleral brachytherapy. Int J Radiat Oncol Biol Phys 17: 1111-1114; 1989.