Treating the invisible target: treatment planning targeting the microscopic tumor spread in gliomas

Glioblastoma (GBM) is the most common primary malignant brain tumor in adults, but treatment outcome is poor, and tumor recurrence is often unavoidable. Evidence suggests that regions with tumor cell densities below the medical imaging visibility threshold are more prevalent than the current treatment methodologies account for. This study aimed to explore different treatment planning strategies that account for the microscopic tumor cell distribution.

A diffusion-proliferation model was used to predict the tumor spread in three patients with GBMs, each presenting different degrees of complexity. The tumor spread was quantified as the invasion volume V_x , defined as the volume encompassed by a simulated cell density isocontour x (cells/mm³). Dose plans were created using targets defined by V ₁₀₀₀, V ₁₀₀, and V ₁₀. Additionally, radiobiological modeling was used to predict the optimal dose per voxel needed to achieve a tumor control probability of 0.95 and treatment plans were made by mimicking the optimal dose distributions as reference. The difference between the reference dose distributions and mimicked dose plans was quantified as the percentage of voxels in the mimicked plan receiving a dose within 95% and 107% of the prescribed dose (quality index, Q _0.95-1.07).

All dose plans generated using the first methodology were clinically acceptable in terms of tumor coverage and sparing of organs at risk. The mean quality index Q _0.95-1.07, evaluated as a function of target volume across all cases, was 95% for the GTV, however, a decreasing trend as the target volume increased was observed.

Incorporating tumor spread models into treatment planning for GBM is feasible and warrants further investigation to potentially enhance treatment outcomes. Dose mimicking approaches based on reference dose distribution require careful planning and become particularly challenging for large, complex targets.