In this paper we describe a concept for dosimetric treatment plan verification using two-dimensional ionization chamber arrays. Two different versions of the 2D-ARRAY (PTW-Freiburg, Germany) will be presented, a matrix of 16 x 16 chambers (chamber cross section 8 mm x 8 mm; the distance between chamber centers, 16 mm) and a matrix of 27 x 27 chambers (chamber cross section 5 mm x 5 mm; the distance between chamber centers is 10 mm). The two-dimensional response function of a single chamber is experimentally determined by scanning it with a slit beam. For dosimetric plan verification, the expected two-dimensional distribution of the array signals is calculated via convolution of the planned dose distribution, obtained from the treatment planning system, with the two-dimensional response function of a single chamber. By comparing the measured two-dimensional distribution of the array signals with the expected one, a distribution of deviations is obtained that can be subjected to verification criteria, such as the gamma index criterion. As an example, this verification method is discussed for one sequence of an IMRT plan. The error detection capability is demonstrated in a case study. Both versions of two-dimensional ionization chamber arrays, together with the developed treatment plan verification strategy, have been found to provide a suitable and easy-to-handle quality assurance instrument for IMRT.

The purpose of the present study is to understand the mechanism underlying the perturbation of the field of the secondary electrons, which occurs in the presence of a detector in water as the surrounding medium. By means of 'reverse' Monte Carlo simulation, the points of origin of the secondary electrons contributing to the detector's signal are identified and associated with the detector's mass density, electron density and atomic composition. The spatial pattern of the origin of these secondary electrons, in addition to the formation of the detector signal by components from all parts of its sensitive volume, determines the shape of the lateral dose response function, i.e. of the convolution kernel K(x,y) linking the lateral profile of the absorbed dose in the undisturbed surrounding medium with the associated profile of the detector's signal. The shape of the convolution kernel is shown to vary essentially with the electron density of the detector's material, and to be attributable to the relative contribution by the signal-generating secondary electrons originating within the detector's volume to the total detector signal. Finally, the representation of the over- or underresponse of a photon detector by this density-dependent convolution kernel will be applied to provide a new analytical expression for the associated volume effect correction factor.

No abstract

Shielded p-silicon diodes, frequently applied in general photon-beam dosimetry, show certain imperfections when applied in the small photon fields occurring in stereotactic or intensity modulated radiotherapy (IMRT), in electron beams and in the buildup region of photon beam dose distributions. Using as a study object the shielded p-silicon diode PTW 60008, well known for its reliable performance in general photon dosimetry, we have identified these imperfections as effects of electron scattering at the metallic parts of the shielding. In order to overcome these difficulties a new, unshielded diode PTW 60012 has been designed and manufactured by PTW Freiburg. By comparison with reference detectors, such as thimble and plane-parallel ionization chambers and a diamond detector, we could show the absence of these imperfections. An excellent performance of the new unshielded diode for the special dosimetric tasks in small photon fields, electron beams and build-up regions of photon beams has been observed. The new diode also has an improved angular response. However, due to its over-response to low-energy scattered photons, its recommended range of use does not include output factor measurements in large photon fields, although this effect can be compensated by a thin auxiliary lead shield.

The spatial resolution of 2D detector arrays equipped with ionization chambers or diodes, used for the dose verification of IMRT treatment plans, is limited by the size of the single detector and the centre-to-centre distance between the detectors. Optimization criteria with regard to these parameters have been developed by combining concepts of dosimetry and pattern analysis. The 2D-ARRAY Type 10024 (PTW-Freiburg, Germany), single-chamber cross section 5 x 5 mm(2), centre-to-centre distance between chambers in each row and column 10 mm, served as an example. Additional frames of given dose distributions can be taken by shifting the whole array parallel or perpendicular to the MLC leaves by, e.g., 5 mm. The size of the single detector is characterized by its lateral response function, a trapezoid with 5 mm top width and 9 mm base width. Therefore, values measured with the 2D array are regarded as sample values from the convolution product of the accelerator generated dose distribution and this lateral response function. Consequently, the dose verification, e.g., by means of the gamma index, is performed by comparing the measured values of the 2D array with the values of the convolution product of the treatment planning system (TPS) calculated dose distribution and the single-detector lateral response function. Sufficiently small misalignments of the measured dose distributions in comparison with the calculated ones can be detected since the lateral response function is symmetric with respect to the centre of the chamber, and the change of dose gradients due to the convolution is sufficiently small. The sampling step width of the 2D array should provide a set of sample values representative of the sampled distribution, which is achieved if the highest spatial frequency contained in this function does not exceed the 'Nyquist frequency', one half of the sampling frequency. Since the convolution products of IMRT-typical dose distributions and the single-detector lateral response function have no or very small frequency contributions beyond 0.1 mm(-1), the mathematical approach introduced by Nyquist and Shannon shows that the sampling frequency of 0.2 mm(-1) is appropriate. Overall it is shown that the spatial resolution of the 2D-ARRAY Type 10024 is appropriate for the dose verification of IMRT plans. The insights obtained are also applied in the discussion of other available two-dimensional detector arrays.