RESUMO
BACKGROUND: The improvement of in vitro assessment of targeted alpha therapy (reproducibility, comparability of experiments ) requires precise evaluation of the dose delivered to the cells. To answer this need, a previous study proposed an innovative dosimetry method based on α-spectroscopy and a specific deconvolution process to recover the spatial distribution of 212 Pb isotopes inside in vitro culture wells. Nevertheless, although promising, the deconvolution method was time consuming and only tested for a simple isotope decay chain. PURPOSE: The purpose of this work is to propose a new matrix deconvolution method of α spectra based on a constrained-non-negative-maximum-likelihood decomposition, both faster and offering a greater modelling flexibility, allowing to study independently the kinetics of each of the daughter nuclides of complex decay chains (illustrated here with 223 Ra) in in vitro culture wells. METHODS: Firstly, the performance of the new method was fully evaluated through Monte Carlo simulations of in vitro irradiations. Different spatial distributions of 212 Pb and 223 Ra, the corresponding α spectra measured by a silicon detector and the doses delivered to the cells were simulated with Geant4. The deconvolution results were then compared to the simulation results. Secondly, measurements were carried out in culture wells without cells containing 15 kBq of 212 Pb or 9.3 kBq of 223 Ra, placed above silicon detectors recording α spectra in real time. The matrix deconvolution was then applied to determine the spatial and temporal distribution of all α-emitting daughters of studied isotopes. RESULTS: The matrix deconvolution was proved to recover the simulated distribution gradients, ensuring simulated doses within 3 % for both tested radionuclides, with errors on dose normally distributed around the reference value (consequently not exhibiting any bias), even in the case of complex decay chains as 223 Ra. The experimental study of 212 Pb and 223 Ra showed highly inhomogeneous distributions and time evolution of the concentration gradients, consistent with the previous study. Furthermore, it highlighted the complex kinetics of 223 Ra with different distributions of its α-emitting daughters (219 Rn, 215 Po, 215 At, 211 Bi, 211 Po). CONCLUSIONS: This study validates a new deconvolution method, fast and flexible, that proved to be accurate and reliable. This method allowed to reveal the complexity of isotopes kinetics in in vitro experiments, especially with complex decay chains. Experimental dosimetry, necessary to improve reliability of in vitro studies in targeted alpha therapy, is demonstrated to be feasible with the proposed method.