Hippocampus avoidance in pediatric patients Free

See the article by Acharya et al. in this issue, pp. 1175–1183.

Hippocampus avoidance (HA) is widely used in adult delivery of whole brain radiation therapy (RT), most often in the setting of cerebral metastatic disease.¹ The evidence for a similar cognitive-sparing strategy in children, however, has been limited. In this issue, Acharya et al² present an impressive secondary analysis of a prospective cohort of childhood and adolescent survivors of low-grade glioma (LGG) treated with focal radiation with frequent neurocognitive evaluations across nearly 10 years of median follow-up—longer than the most substantial prospective adult trials of HA.^2,3 They report a significant effect of radiation dose to the hippocampus on short and long delay recall, though not on processing speed. These changes can be as substantial as a loss of 2 standard deviations at 10 years for short delay recall in patients with hydrocephalus treated with a minimum of 40 Gy to the entire hippocampus. This effect persists after accounting for age at diagnosis and hydrocephalus, though both appear to influence recall ability more than hippocampal dose.

These findings are significant given the current lack of strategies to minimize the neurocognitive impact of radiation to the healthy, developing brain in children who often present at a sensitive age in their development. The widely recognized cognitive consequences of radiation have implications on the eventual ability of patients, particularly those with radiation “curable” malignancies, to pursue a higher quality of life.³ Arguably, the importance of HA may be greater in children than adults, whose more relatively formed brains are less vulnerable to disruption of development and plasticity.

This study also differs from adult trials in that it explores HA for a malignancy that does not require the delivery of radiation to the entire brain for disease control and for which there should be no additional risk of failure with minimization of hippocampal dose as long it is not directly involved by tumor. In RTOG 0933, which was conducted in adult patients with cerebral metastasis from a variety of malignancies receiving whole brain RT, 3 of 67 eventual relapses (4.5%) occurred in the hippocampus avoidance area.¹ The risk is likely much lower in tumors such as LGG and other localized pediatric brain tumors such as craniopharyngioma, which are not at high risk of tumor spread beyond the radiographically visible tumor. Previous work has shown that these malignancies can be successfully cured with tighter margins (3–5 mm).^2,4

The next step is to explore these findings in the setting of a prospective cohort, potentially as a cooperative group study through the Children’s Oncology Group or as a single or multi-institutional trial. Careful patient selection will be required in order not to compromise tumor outcomes. In our opinion, ideal candidates would be patients similar to those included in this study, LGG, craniopharyngioma, or other curable localized brain tumors. As these patients have little risk of relapse outside of the primary tumor, advanced modality RT delivery may allow complete avoidance or drastic minimization of dose to the hippocampus without increased risk of relapse. In contrast, highly radiosensitive malignancies with widespread tumor seeding like medulloblastoma are at high risk of out-of-field relapse and require a component of whole brain radiation. Though these patients may also benefit from HA, there may be increased risk of relapse with reduced dose, and complete sparing would not be feasible.⁵

These data provide clear dosimetric goals in limiting the volume of hippocampus receiving >40 Gy. This study examined mostly midline tumors, though limiting dose to a single hippocampus may be more difficult in more lateralized tumors being treated to doses between 50 and 60 Gy. The dosimetric goals in these and other cases may be more easily achievable with advanced modalities like proton radiotherapy, which has no exit dose and provides a more conformal or focused dose distribution. Protons are now widely available and commonly used for pediatric tumors, particularly cerebral malignancies in order to minimize cognitive impact.⁶ While proton therapy is often denied by insurers for adults, it is generally approved for children.

In order to utilize these data, it is critical to define the hippocampus accurately. Acharya et al employed the RTOG 0933 atlas, and a single radiation oncologist retrospectively contoured the volume of interest in all patients. In the setting of a prospective trial across multiple institutions, it will be important to collaborate with neuroradiologists to ensure that volumes are accurate, and sequences helpful for delineation are being obtained. Central review of volumes should be considered to ensure consistency and clear guidelines. Atlases should be made available to guide investigators.

Additional work will be needed to identify other ways to minimize intellectual decline in patients. The effect of hydrocephalus, in particular, had a profound effect on short delay recall in this study, though the mechanisms of this effect remain relatively elusive. The number and extent of surgeries did not appear influential. The dataset also did not allow for a clear way to account for the possible impact of daily anesthesia through 4–6 weeks of treatment.⁷ Finally, it will also be important to identify the regions of the brain responsible for the impact of radiation on processing speed, which appears unrelated to hippocampus dose and causes profound long-term effects on children, who often require lifelong adjustments, particularly in the setting of standardized exams in school.⁸ We now have the advanced imaging tools to identify regions of the brain with great accuracy and the RT treatment modalities to allow for selective avoidance of regions we deem most important. We must continue to learn from studies such as this and design prospective trials to understand which areas of the brain are most critical to avoid so that we can improve quality of life for our patients.

This study used a very specific metric of cognitive function—recall of words of varying lengths and relationships to each other after intervening distraction. More global work needs to be done to identify how these decrements impact long-term socioeconomic status, employment status, and education level.⁹ We need to understand more about the relationship between the metrics of these studies and the common question of children and parents—to what degree can they live a normal life?

As we continue to improve rates of overall and progression-free survival across pediatric brain malignancies, it will be increasingly crucial to focus on survivorship. Studies like this one identify a simple and compelling potential technique that may have a significant impact on long-term quality of life without compromising tumor control.

References