Neural stem cell sparing by linac based intensity modulated stereotactic radiotherapy in intracranial tumors

Background: Neurocognitive decline observed after radiotherapy (RT) for brain tumors in long time survivors is attributed to radiation exposure of the hippocampus and the subventricular zone (SVZ). The potential of sparing capabilities for both structures by optimized intensity modulated stereotactic radiotherapy (IMSRT) is investigated. Methods: Brain tumors were irradiated by stereotactic 3D conformal RT or IMSRT using m3 collimator optimized for PTV and for sparing of the conventional OARs (lens, retina, optic nerve, chiasm, cochlea, brain stem and the medulla oblongata). Retrospectively both hippocampi and SVZ were added to the list of OAR and their dose volume histograms were compared to those from two newly generated IMSRT plans using 7 or 14 beamlets (IMSRT-7, IMSRT-14) dedicated for optimized additional sparing of these structures. Conventional OAR constraints were kept constant. Impact of plan complexity and planning target volume (PTV) topography on sparing of both hippocampi and SVZ, conformity index (CI), the homogeneity index (HI) and quality of coverage (QoC) were analyzed. Limits of agreement were used to compare sparing of stem cell niches with either IMSRT-7 or IMSRT-14. The influence of treatment technique related to the topography ratio between PTV and OARs, realized in group A-D, was assessed by a mixed model. Results: In 47 patients CI (p <= 0.003) and HI (p < 0.001) improved by IMSRT-7, IMSRT-14, QoC remained stable (p >= 0.50) indicating no compromise in radiotherapy. 90% of normal brain was exposed to a significantly higher dose using IMSRT. IMSRT-7 plans resulted in significantly lower biologically effective doses at all four neural stem cell structures, while contralateral neural stem cells are better spared compared to ipsilateral. A further increase of the number of beamlets (IMSRT-14) did not improve sparing significantly, so IMSRT-7 and IMSRT-14 can be used interchangeable. Patients with tumors contacting neither the subventricular zone nor the cortex benefit most from IMSRT (p < 0.001). Conclusion: The feasibility of neural stem cell niches sparing with sophisticated linac based inverse IMSRT with 7 beamlets in an unselected cohort of intracranial tumors in relation to topographic situation has been demonstrated. Clinical relevance testing neurotoxicity remains to be demonstrated.