Monte Carlo dosimetry for beta source selection and design for endovascular irradiation

Purpose/Objective: Low dose irradiation is emerging as an exciting and promising new treatment modality for the prevention of vessel restenosis, which occurs in about 40% of the percutaneous transluminal angioplasty cases. The most challenging aspect of this new vessel irradiation modality is the determination of an effective target dose, and selection of an optimal source and delivery system. There is still much work to be done in the areas of radiobiological mechanisms, optimal source selection and design, optimal delivery system design, and precise dosimetric evaluation. This particular study was undertaken to use Monte Carlo techniques to examine and compare dose characteristics of two different beta sources (P-32 and Sr-90/Y-90), and to examine the dosimetric effects of source design considerations and vessel heterogeneities, i.e. calcification buildup. Materials and Methods: All our dose calculations were made using the MCNP4A code (Monte Carlo N-Particle, version 4A). This code was originally developed at Los Alamos National Laboratory and is capable of simulating coupled neutron-photon-electron problems using a three dimensional geometry modeling package. To validate our code, we compared point dose distributions for P32 and Sr90/Y90 with published data. We also compared our Monte Carlo calculations with physical measurements made with LiF TLDs about a custom made P-32 source sealed in the end of a flexible NiTi wire. We precisely modeled the source, the detectors, and the phantom material for our calculations. Once validated, we used both conventional and lattice tally geometries to examine two dimensional dosimetric effects resulting from design considerations such as: source selection (P32 verses Sr90/Y90); source length (3mm verses 10mm); source encapsulation materials (NiTi verses Stainless Steel); and catheter material composition (Nylon12, Nylon12 with BaSO4, and PVC) and thicknesses (0.15, 1, and 2 mm), and centering balloon design. We are also examining the effects of vessel composition and heterogeneities on dose to the critical organ (the vessel wall). Results: Validation of Code: We found good agreement between our calculated point dose distributions and those previously published in the literature. For the custom P32 source, our initial comparison of calculated radial dose with measured radial values, at distances of 1.5 mm to 5.0 mm, also show reasonable agreement (at least within measurement error). Source Type: As expected, comparison of the P32 distributions with the Sr90/Y90 distributions show a greater range (almost two times) for the 2.3MeVmax Y90 betas than the 1.7MeVmax P32 betas. Sr90/Y90 also displays a slower falloff (a flatter radial dose curve) than P32 making it more ideal in efforts for uniform dose. Source Length: Slight differences were observed in dose distributions close to the source due primarily to the geometric effects of a line source. Source Encapsulation: We observed little to no difference between Stainless Steel and Nickel Titanium sheaths for the SrY90 source. Catheter Material: No discernible difference was observed until the thickness of the catheter became significant (>1cm). Balloon Material and Vessel Heterogeneities: Our findings on the effects of the centering balloon material and vessel composition will be presented. Conclusion: Monte Carlo has proven to be an excellent tool for quickly assessing the dosimetric properties for endovascular/brachytherapy sources, especially with respect to costly design implementations not yet made. A quick and reliable decision can be made on calculated values, provided accurate material compositions and densities are available. This is especially useful when working with electron sources and low energy gamma sources which are more sensitive to changes in material heterogeneities. Monte Carlo is not plagued with limited size and positioning constraints of TLD, diode, and ion chamber measurements, and it is not affected by energy calibration considerations needed for radiochromic film. In this particular study we found that the re is little dosimetric advantage in changing material composition for the sheath or catheter. Small variations in thicknesses (<1mm) also make little to no contribution in radial dose