Impact of Realistic 3D Tumor Microarchitecture on Cellular Dosimetry and Radiobiological Response in Targeted Radionuclide Therapy

Radiation therapy using radionuclides has become more precise with the aid of quantitative imaging; however, the models used to estimate cellular doses do not account for the complex microarchitecture of tumors. Traditional models treat cells as uniform spheres, which may not reflect the actual tumor structure. This research explored how a detailed, 3D representation of tumor microarchitecture can affect the absorbed doses in individual cells, the expected biological response, and the overall treatment strategy.

Researchers created realistic 3D cellular models using 3D cyclic immunofluorescence imaging data. These models allowed for the simulation of radiation effects using Monte Carlo methods for different radionuclides such as Lu-177, Pd-103, Tb-161, and Y-90. The simulations focused on the localization of radiation activity within cancerous cells and in healthy tissues.

The absorbed dose per decay event was computed for each cell nucleus, and this was contrasted with the predictions made by simpler spherical tumor models. The individual survival rates of cells were estimated using a linear-quadratic model, and the tumor control probability was derived from these survival rates.

The findings indicated that the simpler models often overestimated the average dose to cancer cell nuclei by 9.54% to 14.51% when the dose was localized in the nucleus, and by a larger margin of 33.23% to 44.88% when the dose was in the cytoplasm or cell membrane. In more extensive tissue models, the simplified and accurate models sometimes yielded similar average doses, but the accurate models consistently showed a wider range of doses across individual cells, including some that were not lethal under the simplified predictions.

This discrepancy in dose distribution led to a broader range of survival rates and altered the predicted tumor control probability (TCP).

Specifically, the realistic model required 1.11- to 1.33-fold more activity to achieve the same TCP for nuclear localization, and up to 5.93-fold more for cytoplasmic and membrane localization. To achieve the same level of tumor control, the choice of radionuclide and how its source was localized in the tumor differed between the simplified and realistic models. This detailed analysis also revealed that assuming uniform activity within each voxel might hide important microscopic variations in dose.

In conclusion, the intricate structure of the tumor has a significant impact on how cells absorb radiation, how they respond biologically, and what treatment conclusions can be drawn. Employing 3D models that reflect the actual tumor architecture can provide a more accurate representation of cellular dose heterogeneity, potentially leading to better decisions about which radionuclides to use and how to target subcellular components in tumors with variable structures.