Impact of intracellular radionuclide distribution in a Monte Carlo biophysical 3D multi-cellular model for targeted alpha therapy

Background

To understand and predict the therapeutic efficacy of targeted alpha therapy (TAT), nano- and microdosimetry are needed to consider the very heterogeneous dose deposition at cellular and subcellular levels.

Purpose

The objective of this study is to theoretically evaluate the importance of cell internalization of alpha-emitters on relevant dosimetric and biological endpoints.

Methods

Isolated cells and realistic 3D multi-cellular geometries (spheroids modeled with CPOP) were generated as well as distributions of alpha-emitters corresponding to various cellular internalization cases. The alpha particles emitted were tracked with Geant4 (Monte Carlo) simulations. We calculated mean specific energies deposited into each cell nucleus ($Z_n$), cell survival fractions using the NanOx biophysical model, values of relative biological effectiveness (RBE) and tumor control probabilities (TCP) for each scenarios. The impact of spheroid compaction and size, alpha particle energy and radionuclide daughter diffusion was studied. The impact of the heterogeneous distribution of a number of alpha particles per cell was also studied, using a lognormal probability law.

Results

For a given activity per cell (APC), the radionuclide distribution had a critical influence on $Z_n$ in isolated cancerous cells or small spheroids (50 $\mu m$ radius), while its impact was relatively low in larger and more compact spheroids, with a maximum variation of 30% between the distributions. For an average 10% cell survival, RBE was found to be approximately between 2.3 and 3.3, depending on the spatial radionuclide distribution and the activity distribution per cell. TCP of 1 was always obtained with an APC larger than 0.534 mBq when a uniform tumoral distribution of radionuclides was considered, and for APC larger than 0.801 mBq with a lognormal distribution. However, below these activities, TCP could strongly depend on the radionuclide distributions up to a factor of 9.5 with a uniform distribution and 1.5 for a lognormal one.

Conclusions

According to these findings, a precise modeling of alpha-emitter intracellular distributions may be required for small micro-metastases or tumors presenting regions with relatively low radionuclide concentration in order to limit the prediction uncertainties on biological outputs. Intratumoral fluctuations of APC were also found to be a critical parameter to consider for therapeutic efficacy prediction in TAT.