Current estimates indicate that roughly 200,000 patients are diagnosed with brain metastases annually in the United States, and 10-30% of patients with cancer receive a diagnosis of brain metastases during their disease course (1–4). These estimates may be conservative, as the true incidence is likely higher, due to a multitude of factors, such as undiagnosed brain metastasis identified on autopsy and underreporting with national registries (e.g. The National Cancer Database and Surveillance, Epidemiology, and End Results) (5, 6). Historically, patient prognosis was poor with a median overall survival of 3-4 months in patients who did not undergo surgical intervention (7). However, advancements in systemic therapy, surgery, and radiation therapy have resulted in survival advantages across multiple malignancies, thus less common and aggressive histologies are increasingly metastasizing to the brain (e.g., gastrointestinal primary cancers) (1, 8–13). Additionally, the widespread availability of MRI imaging has enhanced detection of subclinical disease.

Whole-brain radiation therapy (WBRT) is a treatment modality that has been used since the 1950s for patients with brain metastases (14). It is commonly delivered to a total dose of 30 Gy over 10 sessions. Clinicians traditionally favored WBRT due to its efficacy in providing palliation and ability to target unknown microscopic intracranial disease. WBRT has been shown to result in improved intracranial tumor control in multiple randomized trials; however, WBRT has also been shown to result in significant cognitive decline, which has been observed in up to 50% of patients following treatment (15–20). These patients can present with one or multiple cognitive domains affected, such as executive function, learning and memory, processing speed, and verbal fluency. As the prognosis in patients with brain metastases continues to improve, treatment has increasingly focused on preservation of quality of life (QOL) and cognitive function. Multiple studies have suggested that there is a correlation between neurocognitive functioning and QOL (21, 22). In recent years, the addition of memantine and hippocampal avoidance to WBRT have demonstrated significant preservation of cognitive sequelae and, in the setting of hippocampal avoidance, better preservation of patient-reported QOL, and now represent current standard of care in appropriately selected patients (17, 23, 24).

Moreover, the efficacy of brain-directed radiotherapy in providing adequate palliation has come under question in recent years. In 2016, the QUARTZ trial, which randomized 538 patients with non-small cell lung cancer brain metastases and poor prognosis, to dexamethasone with WBRT or dexamethasone with supportive care alone was published (25). This trial concluded that WBRT did not offer any benefit in QOL or survival (median survival in both arms was approximately 2 months) over supportive care, thus calling into question the efficacy of brain-directed radiotherapy in this setting. However, while broad, indiscriminate use of WBRT has fallen out of favor, it continues to be a commonly used modality in patients with a high intracranial burden of brain metastases (26, 27).

Another important advancement in the treatment of brain metastases is stereotactic radiosurgery (SRS), which is defined as the delivery of a high dose of very conformal radiation in 1-5 fractions with marked sparing of nearby healthy tissues (28). SRS has been shown to result in fewer cognitive side effects than conventional WBRT (15, 16, 18). However, WBRT has been shown to provide superior rates of intracranial control, especially by decreasing the risk of development of new brain metastases, when compared to SRS (15, 16, 18, 29). As a result, there is some controversy regarding the use of SRS in patients with large numbers of brain metastases (30, 31). Multiple studies have reported a lower risk of cognitive decline with SRS than conventional WBRT; studies comparing SRS to contemporary WBRT with neuroprotective strategies, such as memantine and hippocampal avoidance remain ongoing (32, 33).

In this article, we review the pathogenesis, diagnosis, and evaluation of cognitive decline following cranial irradiation. Additionally, we review the role of SRS, hippocampal avoidance, and memantine as risk mitigation strategies in patients undergoing cranial irradiation for brain metastases.

Despite evidence showing that radiation causes cognitive impairment, the pathophysiological understanding of this common clinical scenario is poorly understood. The cause is believed to be multifactorial with changes in brain vasculature, stem-cell depletion, and changes to the brain’s microenvironment all being implicated. While damage to the hippocampus has been implicated in cognitive decline following cranial irradiation, recent evidence suggests that damage to white matter and other cortical territories, such as the frontal cortex also play a significant role (34–37).

It is important to note that patients with brain metastases typically have cognitive impairment at baseline before radiotherapy: a phase 3 trial with prospective cognitive testing found greater than 90% of patients had impairment on one of more cognitive tests at baseline (20). Following completion of radiotherapy, patients can present with cognitive decline as early as 1- to 6-months following treatment. Symptom presentation during this time window is potentially reversible and is believed to be due to transient demyelination (43). Patients who present at 6-months or later generally have irreversible cognitive dysfunction with multiple affected cognitive domains. These patients frequently present with deficits in attention, information processing, executive function, and learning and memory. Multiple radiographic findings, such as white matter abnormalities and changes in fractional anisotropy on diffusion tensor MRI have shown an association with cognitive decline (79–91). Consultation with neuropsychologists can be very helpful in quantifying and trending cognitive changes. Additionally, it is essential for clinicians to rule out other possible causes, such as dementia, delirium, metabolic and endocrinologic disturbances, and disease progression.

The diagnosis of neurocognitive decline following cranial irradiation requires neuropsychological assessment. Earlier clinical trials utilized the screening test, the Mini-Mental Status Exam (MMSE); however, its use in this setting has fallen out of favor due to its limited sensitivity in diagnosing cognitive impairment (92). More commonly, clinical trials now employ neuropsychological testing that assesses multiple cognitive domains, such as executive function, learning and memory, verbal fluency, and attention. Commonly utilized cognitive assessments on randomized phase 3 clinical trials are presented in Table 1 .

Due to the role the hippocampus plays in learning and memory, there has been a great deal of interest in sparing this region of the brain during WBRT (23, 24, 102). In 2014, Gondi et al. published the results of RTOG 0933, which was a phase 2 multi-institutional study where patients with brain metastases outside a 5 mm margin around the hippocampi received WBRT with hippocampal avoidance to a dose of 30 Gy in 10 fractions ( Figure 1C ) (24). There were 100 patients enrolled and all underwent formal neuropsychological testing. Enrolled patients were compared to the control arm of PCI-P-120-9801, which was a phase 3 study utilizing WBRT with identical eligibility criteria to RTOG 0933 (103). At 4 months, the mean relative decline in the modified HVLT-R Delayed Recall compared to baseline was 7%, which was significantly improved from the historical control (p < 0.001). In addition, cognitive results were comparable to what had been observed in prior studies of SRS (18). Similar to the findings of RTOG 0614, the benefits in cognitive preservation were seen in patients who were able to complete neuropsychological testing at 4 months. Thus, hippocampal avoidance is likely more beneficial in patients with a better baseline prognosis and life expectancy.

In 2020, Brown et al. published the results of NRG CC001, which was a phase 3 trial where patients with brain metastases were randomized to: (1) hippocampal avoidance WBRT with memantine; or (2) WBRT with memantine (23). There were 518 patients enrolled with a median follow-up of 7.9 months. All patients completed neuropsychological testing at regular intervals. The primary endpoint was time to cognitive failure, as shown in Table 1 . The risk of cognitive failure favored the hippocampal avoidance arm (HR: 0.76; 95% CI: 0.60-0.98; p = 0.03). Additionally, at 6-months patients in the hippocampal avoidance arm had less memory complaints (p = 0.01), fewer cognitive symptoms (p = 0.01), and less symptom interference (p = 0.008). At 6-months, approximately 80% of patients in each arm died. This suggests that patients with a better baseline prognosis and life expectancy are likely to benefit the most from hippocampal avoidance. However, not all patients with brain metastases were eligible for inclusion on NRG CC001, such as patients with ventricular system distortion or hydrocephalus, the presence of leptomeningeal disease, and brain metastases arising from primary germ cell tumors, small cell carcinoma, an unknown primary or lymphoma. In 2021, a phase 2 randomized trial conducted in China compared WBRT with or without the use of hippocampal avoidance in patients with brain metastases (104). This trial demonstrated that hippocampal avoidance as associated with better memory preservation at 6-months compared to conventional WBRT.

Taken together, these trials suggest that in patients undergoing WBRT, the use of memantine and hippocampal avoidance reduces the risk of cognitive decline following WBRT. Additionally, these benefits are the most apparent in patients with a better baseline prognosis and life expectancy. However, not all patients are eligible for hippocampal avoidance, as no metastases are permitted within 5 mm of the bilateral hippocampi. Furthermore, hippocampal avoidance requires the use of advanced planning methods, such as intensity modulated radiation therapy and volumetric modulated arc therapy. Similarly, SRS requires specialized planning techniques, as well as specialized radiosurgery platforms to deliver treatment. Therefore, not all centers may have the necessary technical capabilities to deliver these treatments, particularly in underserved areas.

Preserving neurocognition in patients undergoing cranial irradiation continues to be a major area of research focus. In patients with small cell lung cancer (SCLC), prophylactic cranial irradiation (PCI) is frequently administered to a dose of 25 Gy in 10 fractions in patients with no detectable brain metastases (105–107). PCI was historically administered using conventional WBRT techniques; however, recent studies have assessed incorporating hippocampal avoidance in this setting. In 2021, the PREMER study was published, which randomized 150 patients with SCLC to standard or hippocampal avoidance PCI across 13 institutions in Spain (108). At 3-months the investigators observed that the decline in memory favored the hippocampal avoidance arm (5.8% vs. 23.5%; p = 0.003). However, in 2021, a phase 3 trial conducted in the Netherlands did not observe a lower probability of cognitive decline in the hippocampal avoidance PCI arm (109). The NRG CC003 trial is presently investigating this hypothesis in North America and will be completing accrual later this year (110).

While there is strong evidence supporting the role of SRS in the management of limited numbers of brain metastases (15, 16, 18), the use of SRS remains controversial in the management of larger numbers of lesions. In 2020, a study by Rinna et al. observed that in patients undergoing Gamma Knife ^® SRS that 10 or more metastases, and metastases in close proximity to the hippocampi were at an increased risk for excessive hippocampal dosing (111). In 2021, a study published by Burgess et al. evaluating 60 SRS plans with a median distance to the hippocampus of 2.4 cm observed that patients can undergo replanning to decrease the hippocampal dose by > 50% without compromising target coverage (112). Taken together, these findings suggest that limiting dose to the hippocampus during SRS may further decrease the risk of cognitive decline in these patients.

CCTG CE.7 is an ongoing phase 3 trial that is randomizing patients with 5-15 brain metastases to SRS or WBRT with the addition of memantine and hippocampal avoidance (33). There is ongoing prospective investigation into the role that regions outside of the hippocampus play in cognitive decline following cranial irradiation (113). There is presently a trial underway at Johns Hopkins University investigating neurocognitive functioning with sparing of the genu of the corpus callosum during WBRT for brain metastases (114). Another study underway at the University of California San Diego is investigating sparing of white matter tracts during SRS for brain metastases (115).

In recent years, brain metastasis velocity, which describes the recurrence rate of new brain metastases following treatment with SRS had become a validated prognostic metric (116). NRG BN009 is a phase 3 trial comparing salvage SRS to SRS with hippocampal avoidance WBRT with the addition of memantine in patients with a first or second distant relapse following upfront SRS and a brain metastasis velocity of 4 or higher (32).

Cognitive decline is a common clinical manifestation observed in patients who undergo WBRT for brain metastases. Strategies, such as SRS and the addition of memantine and/or hippocampal avoidance to WBRT are excellent treatment options to mitigate this risk. Studies are underway that will allow for further application of these treatments, as well as defining the role that other brain regions play in the pathogenesis of cognitive decline following cranial irradiation.

Conceptualization: EL, DT. Supervision: DT. Writing – original draft: EL, BJ, DR. Writing – review and editing: All authors. All authors contributed to the article and approved the submitted version.

PB reports contribution to UpToDate outside of the submitted work. JW is on the advisory board of Bayer, he serves as a consultant to Angiochem, Bayer, Juno, Novocure, Vanquish Oncology, and GT Medical technologies. JP reports research funding and honoraria from Varian and research funding from Genentech, NIH, and Kroger; he serves on the advisory board of Novocure. IM serves as a consultant to BrainLab and Integra; DT reports institutional support from Novocure Ltd and consulting for Boston Scientific Corporation outside to the submitted work.

The remaining authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.