Gliomas and meningiomas are two representative primary brain tumors (1). Most gliomas are treated using surgery followed by radiotherapy (RT) (2). The treatment of atypical and anaplastic meningiomas are also similar (3). In the long-term survivors with brain tumors after RT, cognitive dysfunction associated with adverse effects of RT is one of the most concerning complication (4). A representative cause of the radiation-induced neurocognitive decline is known to the damage of neural progenitor cells located in subgranular zone of the hippocampus, and RT sparing the hippocampus has been attempted (5). Recently, several prospective cohort studies have reported an association between hippocampal dose-volume parameters and memory decline in patients with primary brain tumors treated by RT (6-8). Gondi et al (6) showed that equivalent dose in 2 Gy fractions (EQD2) to 40% of the bilateral hippocampi was associated with long-term memory impairment. Ma et al (7) reported that D_50% of the bilateral hippocampi of 22.1 Gy was associated with 20% risk of memory decline: D_n% was irradiated dose to n% of the volume of the structure. Okoukoni et al (8) mentioned that the hippocampal V_55Gy was a significant predictor for memory impairment: V_nGy was the percent of the volume of the structure at least irradiated n Gy. However, in daily clinical practice, these dosimetric parameters that are easily achieved and the factors that interfere with reaching these parameters remain unclear.

Therefore, we retrospectively evaluated the feasibility of the hippocampal dose-volume parameters that have been associated with memory decline for intensity-modulated radiotherapy (IMRT) in patients with supratentorial tumors in daily clinical practice.

Tumor characteristics are listed in Table I. D_40% of 13.1 Gy, D_50% of 29.6 Gy, and V_55Gy of 5.0% were achieved in 17, 67 and 33% of patients, respectively. We accomplished all parameters in 3 patients (100%) with atypical meningioma, while no parameters were achieved in 5 of 6 patients (83%) with PTV2 ≥500 cc. The accomplishment of the parameters is listed in Table II. For D_40% of 13.1 Gy, non-meningioma (P=0.001) and tumor spread to the subventricular zone (P=0.025) were significantly unfavorable. For D_50% of 29.6 Gy, PTV2 ≥500 cc (P=0.004) and tumor located in the temporal, corpus callosum, or basal ganglia (P=0.009) were significant unfavorable factors. Non-meningioma was a significant unfavorable factor for V_55Gy of 5.0% (P=0.025).

In this study, we identified that D_50% of 29.6 Gy were easily achieved but large PTV and tumor location interfered with reaching the parameter.

The hippocampus has long been implicated in the acquisition of new memories, with visuo-spatial memory predominantly associated with the right, and verbal or narrative memory with the left hippocampi; the right hippocampi appears to be particularly involved in memory for locations within an environment and the left hippocampi more involved in context-dependent episodic or autobiographical memory (10).

The following mechanisms for radiation-induced neurocognitive dysfunction have been proposed (11). First, pro-inflammatory changes following RT cause an increase in the numbers of microglia, which produce tumor necrosis factor-α and interleukin-1β. And then, this contributes to an ongoing inflammatory state and alteration in the microenvironment, which preferentially drives differentiation of neural precursors to an astrocytic lineage. Moreover, radiation disrupts the vascular niche of the neural precursors and additionally leads to ischemia and toxic neuroexcitation among mature neurons. Finally, radiation exposure reduces the number of dendritic spines on mature neurons, which in turn disrupts synaptic efficiency.

In the clinical course, radiation-induced neurocognitive decline begins with a transient cognitive decline at approximately 4 months posttreatment, followed by an improvement, and then a progressive, irreversible deterioration in cognitive functioning at 12 months or later after irradiation (12).

Recent technical advances have facilitated the evaluation of the hippocampal dose-volume parameters at the time of RT. Several prospective cohort studies have reported an association between hippocampal dose-volume parameters and memory decline in patients with primary brain tumors treated by RT (6-8). Gondi et al (6) reported that EQD₂ to 40% of the bilateral hippocampi was associated with long-term impairment in list-learning delayed recall after RT, and concluded that modern IMRT techniques can reduce the dose to the bilateral hippocampi below the dosimetric threshold. Ma et al (7) reported that D_50% of the bilateral hippocampi of 22.1 Gy was associated with 20% risk of decline for delayed recall, and concluded that their data support a potential benefit of hippocampal sparing. Okoukoni et al (8) reported that the hippocampal V_55Gy was a significant predictor for impairment in immediate recall, and concluded that a limiting dose below 55 Gy may minimize radiation-induced cognitive impairment. These findings encouraged us to spare the hippocampus in daily clinical practice, but we have experienced some cases that have been difficult to reduce the dose to the bilateral hippocampi below the threshold. And then, we explored to the dosimetric parameters that are easily achieved and the factors that interfere with reaching these parameters.

As a result, among the previous reported dose-volume parameters of the bilateral hippocampi with 20% risk of memory decline for 30 fractions, we showed that D_50% of 29.6 Gy was most achievable, but even D_50% of 29.6 Gy may be difficult to be reached in patients with PTV2 ≥500 cc or tumors located in the temporal, corpus callosum, or basal ganglia. The large target volumes could become easy to be close to the bilateral hippocampi. Moreover, tumors in the temporal, corpus callosum, or basal ganglia are also near the bilateral hippocampi anatomically. In the treatment planning, if there were overlaps between the PTVs and the hippocampi, we gave the priority to irradiate the PTVs rather than to spare the hippocampi to aim for cure. Therefore, even the most achievable parameter may be difficult to be reached in patients with the targets near the hippocampi.

Because of its retrospective nature, our study has certain limitations, such as the single institutional design and the small number of samples analyzed.

In conclusion, in IMRT with a dose of 60 Gy in 30 fractions for supratentorial tumors, D_50% of 29.6 Gy was most likely to be achieved in the dose-volume parameters of the bilateral hippocampi associated with 20% risk of memory decline. In daily clinical practice, we may had better primarily try to achieve D_50% of 29.6 Gy of the bilateral hippocampi. However, even the most achievable parameter may be difficult to be reached depending on the target size or tumor location. The further larger study is needed to support our findings.