It is our hope that the work presented in this report will result in incremental improvement in the complex process of optimization of EPT. This optimization process has already resulted in: the use of I-125 or 103Pd rather than 60Co or 192Ir, which lowered the incidence of complications associated with EPT (3, 4, 6, 14, 18, 20); addition of episcleral hyperthermia to EPT which has helped to reduce tumor doses by 30% while maintaining efficacy of this brachytherapy (7, 17) and the use of 3D image reconstruction and 3D dosimetry in the treatment planning process (1,2).
Referring again to Fig. 3, the slotted plaque used in this study is roughly comparable to the conventional 20 mm diameter COMS plaque. The slotted plaque holds 25 sources whereas the COMS design holds 24 sources. The slotted plaque, however, is substantially thinner than the COMS plaque, being only 2 mm compared to 4 mm high. This makes the slotted plaque easier to work with surgically. It also leaves substantial room for modification of the slotted design. It is probable that if the slots were made a bit deeper, (the plaque would need to be a corresponding amount thicker), dosimetric homogeneity would be further improved.
The slotted plaque is faster and easier to load (and unload) with I-125 sources than conventional plaques. The plaque is positioned concave side up, and sources are simply dropped into a slot and a drop of cyanoacrylate adhesive placed on each source. The beveled sides of the slot automatically center the source. To remove the sources, the plaque is placed in a beaker of acetone for about 15 min. The acetone dissolves the cyanoacrylate adhesive, and the plaque is then turned concave side down. The sources fall to the bottom of the beaker where they are easily retrieved with forceps or tweezers. Conventional plaques are more difficult to work with. In the COMS design, for example, the sources must be pushed into (and later removed from) tight slots on the convex side of a silicone carrier, a tricky procedure with long forceps. The carrier is then inserted into the gold shell. If the carrier is glued into the shell with a silicone adhesive, removal is difficult since silicone adhesives are not easily disolved.
The dosimetric model appears to be adequate for clinical use. There remains room for improvement in the model, however, particularly with regard to accounting for deviation from a homogeneous, full scatter geometry. We have previously determined (12) that a small correction is necessary at short distances to account for the presence of a simple gold shell. The more complex structure of the slotted plaque may require a different, perhaps greater correction.
The determination of source, phantom, and dosimeter coordinates is subject to numerous uncertainties. Measurements at 2 mm from the plaque surface, in a region of steep dose gradient, using TLD rods of 1 mm diameter would be expected to be sensitive to small localization errors. A small improvement of the calculational model would result from volume-averaging dosimetry over the detector for comparison with physical measurements, rather than simply calculating dose to a single point at the center of a dosimeter, or averaging a few points along a line crossing the detector.
Treatment planning for ophthalmic tumors typically includes about a 2 mm margin surrounding the base of the visible tumor. With the slotted plaque, however, the dose gradient outside this margin is steeper than with other plaque designs. Since error is less tolerable with regard to plaque placement, the slotted plaque does place greater demands on the skills required of the ophthalmic surgeon. This has not presented a problem in our clinical experience. The sharp edge of the dose distribution also means that the slotted plaque may require a few more isotope sources to cover the tumor base (and margin) than conventional plaque designs. This may slightly increase the cost of treatment when using the slotted plaque. The dosimetric benefits, however, appear to greatly outweigh the additional cost.
The prototype plaque used for this study was empirically designed to yield good results for tumors roughly 6 to 7 mm tall. It is likely that different slot dimensions will yield better results with tumors of different heights. One of the more important constraints limiting the design of slotted plaques, however, is to not have so much collimation that "cold spots" are produced within the tumor. A "family" of plaques based on the slotted design, including notched plaques designed for tumors close to the optic nerve are logical extensions of this work.
The goals of conformal therapy are to improve dose homogeneity within the tumor and reduce dose to uninvolved structures outside the tumor. Our simulation model and measurements indicate that the slotted plaque achieves these goals. The dosimetric benefits of the slotted plaque may be particularly valuable for patients with taller tumors and patients with tumors which lie close to the fovea or optic nerve.