Analysing methods of animal irradiation experiments with deviations from prescribed dose


Analysing methods of animal irradiation experiments with deviations from prescribed dose

Beyreuther, E.; Eger Passos, D.; Karsch, L.; Löck, S.; Pawelke, J.

Introduction: The development of new radiotherapy technologies is a long time process which requires proving the general concept although clinical requirements with respect to beam quality and controlled dose delivery may not yet be fulfilled. Exemplarily, the necessary radiobiological experiments with laser-accelerated ion beams, which promise to compact ion radiotherapy facilities, are challenged by low particle energy and fluctuating beam intensities delivered by currently available laser systems. The first issue was handled by establishing a small tumour model on mouse ear that allows full penetration by ~25 MeV proton beams [1], whereas the latter, i.e. the subsequent deviations of the delivered from the prescribed dose, should be considered mathematically.

Methods: Based on tumour growth data and dose values obtained in a preceding in vivo trial comparing the biological efficacy of laser-driven and conventional LINAC electrons [2], different mathematical approaches to determine corresponding dose-response relationships were compared. During this experiment, the beam intensity fluctuations were not fully gathered by online dosimetry, which results in deviations of more than 10 % from scheduled dose as measured by retrospective absolute film dosimetry. Instead of classical averaging-per-dose point, which excludes animals with high dose deviations, multivariate linear regression, Cox regression and a Monte Carlo based approach were tested as alternatives to include all animals in statistical analysis.

Results: The application of different mathematical approaches to the same set of experimental tumour growth data and dose values led to similar results, revealing a comparable radiobiological efficacy of laser-driven and conventional LINAC electrons. Although the inclusion of those animals that were previously excluded because of more than 10% dose deviation did not change the experimental conclusion, the new mathematical approaches allowed for including all animals in the analysis. Comparing the different approaches, multivariate linear regression and Cox regression were considered as most feasible for future analysis, since they were already implemented in commercial statistical software, like SPSS (IBM).

Conclusion: The previously established small animal tumour model on mouse ear [1] together with the recently tested regression methods enable the investigation and evaluation of beams at new accelerators relative to their conventional equivalents despite their still limited beam stability, like laser-driven particle beams. The tested mathematical approaches allow for increasing the number of animals in analysis and therewith reduce the total number of animals in experiment with respect to the 3R of animal experimentation.

Acknowledgement: The work was supported by the German Government, Federal Ministry of Education and Research, grant nos. 03ZIK445 and 03Z1N511.

[1] Beyreuther et al. An optimized small animal tumour model for experimentation with low energy protons. PLOS One,2017; 12: e0177428.
[2] Oppelt et al. Comparison study of in vivo dose response to laser-driven versus conventional electron beam. Radiat Environ Biophys, 2015; 54:155-166.

  • Contribution to proceedings
    ERRS & GBS 2017, 17.-21.09.2017, Essen, Germany
  • Lecture (Conference)
    ERRS & GBS 2017, 17.-21.09.2017, Essen, Germany

Permalink: https://www.hzdr.de/publications/Publ-25607
Publ.-Id: 25607