21 - Evaluation of the size of micrometric/nanometric dosimeters for use in radiotherapy and medical physics

Abstract

When treating tumors with radiotherapy, it is of utmost importance to ensure that the prescribed dose is accurately delivered to the target volumes. In that sense, in-vivo dosimetry in real time was recently implemented in radiotherapy departments. Dosimeter performance depends necessarily on physical and geometrical parameters (e.g. beam energy and distance from source to skin), which implies the use of correction factors. Implantable dosimeters are corrections. They should be as small as possible, but still they should provide reliable measurements to comply with the requirements of clinical practice in routine radiotherapy. The state-of-the-art of these kind of dosimeters was the subject of a review elsewhere (1), which reported that implantable detectors of submillimetric size are currently available. The purpose of this study is to assess by Monte- Carlo simulations how much the size of such dosimeters can be decreased without jeopardizing their performance in a clinical environment. First, the interaction of photons from a 60Co source with water was simulated with a Monte-Carlo tool (2). The calculations were performed for 0.3, 0.1, 1, 3 and 10 Gy. Then, the distributions of specific energy were obtained for volumes representing dosimeters at nanometric and micrometric scales. Cylinders with equal radii of 0.3, 0.1, 1, 3 and 10 μm were used for this purpose. The mean specific energy was calculated for each case. To evaluate how the dosimeter size would impact its performance in a clinical scenario, the probability p that a dosimeter measurement falls outside a given interval defined around was estimated. Intervals were defined as [ -γ ; +γ ] with γ equal to 3%, 5% and 10%. The pattern of the distributions of specific energy evolves with dosimeter size and irradiation dose. Fixing the irradiation dose and decreasing the dosimeter radius or fixing the radius and decreasing the irradiation dose strongly widened the range in measured values of specific energy, but also increased the probability of yielding a non-null measurement. In turn, for higher doses and radii, distributions tend to Gaussian curves. Concerning the probability of obtaining a measurement outside the defined interval, the larger the interval, the irradiation dose, and the dosimeter radius, the smaller this probability became (see figure above). The simulation results showed that dosimeters at a nanometric scale are not able to yield statisticallyreproducible measurements and are therefore unfit for use in clinical practice. Increasing the size to micrometric scale led to a decrease in the statistical fluctuations. Nevertheless, to have enough accuracy at routine clinical doses (approximately 2 Gy in the tumor volume), a dosimeter radius of at least 10 μm is required.