A fast convolution-based method for microdosimetric comparison of ²²⁵Ac, ²¹¹At, ¹⁷⁷Lu and ¹⁶¹Tb at the cell cluster scale

OBJECTIVE: In targeted radionuclide therapy, differences in particle range and radiation quality produce distinct spatial energy-deposition patterns within cell clusters and thereby influence population response. This study aimed to compare the dose deposition, microdosimetric characteristics, and associated population responses of ²²⁵Ac, ²¹¹At, ¹⁷⁷Lu and ¹⁶¹Tb at the cell-cluster scale. Approach. A cubic lattice cell cluster was constructed by replicating a realistic single-cell PC-3 mesh-type model. PHITS was used to calculate distance-grouped nuclear S values and microdosimetric quantities for different subcellular source localizations. Using fast Fourier transform (FFT)-based convolution together with the saturation-corrected microdosimetric kinetic (MK) model, the dose distribution and population response under different labeling fractions, cell cluster sizes, and lognormal activity heterogeneity were evaluated. Main results. Convolution-reconstructed cluster-averaged S values agreed well with direct PHITS simulations, with deviations below 5% in most cases. The four radionuclides showed distinct spatial microdosimetric patterns. For ²²⁵Ac and ²¹¹At, the saturation-corrected dose-mean lineal energy, y*, decreased from about 60-70 keV μm⁻¹ in the near layers to below 4 keV μm⁻¹ in the distant layers, whereas the variations for ¹⁷⁷Lu and ¹⁶¹Tb were more gradual. Across subcellular localizations, labeling fractions, cluster sizes and activity distributions, the mean activity per cell required for ²²⁵Ac to achieve a tumour control probability (TCP) of 0.9 remained approximately 3 orders of magnitude lower than that for ¹⁷⁷Lu and ¹⁶¹Tb . Activity heterogeneity further increased the mean nuclear absorbed dose per cell required to achieve the same TCP, with the strongest amplification observed under nuclear localization and for ¹⁶¹Tb. Significance. This study establishes a rapid convolution-based framework for cell-cluster-scale analysis. Under a unified set of physical inputs and model assumptions, it enables systematic comparison of spatial dose deposition, microdosimetric variation, and population response among radionuclides and across subcellular localization, labeling fraction, cluster size, and activity heterogeneity. .