Management of Complications From Brain Metastasis Treatment

A Narrative Review

Kevin Diao; Alan J. Sosa; Gabriel Zada; Seema Nagpal; Eric L. Chang


Chin Clin Oncol. 2022;11(2):11 

In This Article

Late Complications

Radiation Necrosis

Radiation necrosis refers to necrosis of normal brain tissue secondary to radiation treatment. About 80% of cases occur within 3 years after radiation treatment but in rare cases have been reported up to a decade afterwards.[53,54] The clinical presentation of radiation necrosis is variable and depends on the anatomic location affected. In general, radiation necrosis can be asymptomatic, cause global symptoms such as headache, nausea, or vomiting from increased ICP, seizures or focal neurologic deficits that localize to the region of radiation necrosis.[54] In most cases, tissue is not obtained and therefore imaging and clinical correlates are used to inform diagnosis. The incidence of symptomatic radiation necrosis following SRS ranges from 4–20% and is commonly estimated at 10% overall.[55–58] Conversely, the risk of radiation necrosis is minimal (<1%) following WBRT alone and standard dosing regimens.

The imaging diagnosis of radiation necrosis is challenging, and biopsy or resection is required for definitive diagnosis. Some authors advocate for the use of terms such as adverse radiation effect (ARE) and treatment-related imaging changes (TRIC) as broader terms to capture both reversible and irreversible radiation changes.[19,57] Advanced imaging techniques, including perfusion MRI, diffusion-weighted MRI, MRI spectroscopy, and positron emission tomography (PET) can be useful as diagnostic adjuncts.[59–65] The objectives in the management of radiation necrosis are to palliate symptoms and prevent progressive neurologic deficits. In asymptomatic patients after initial diagnosis of radiation necrosis, close observation with a repeat MRI in 6–8 weeks followed by spacing to every 2–3 months after lesion stability or regression is reasonable as there is no evidence that treatment at this stage will alter disease course. In many patients, the imaging changes will stabilize and improve over the course of weeks to months.

In patients with symptomatic radiation necrosis, systemic corticosteroids such as dexamethasone are the first-line treatment. For patients with mild to moderate symptoms, a starting dose of 2 to 4 mg dexamethasone PO BID is reasonable.[39,40] Patients with severe symptoms should be considered for emergent evaluation and potential inpatient management. Symptom improvement occurs rapidly after initiation of steroids but imaging changes, such as improvement in perilesional vasogenic edema, can lag for several weeks. As such, we typically wait at least 4 weeks prior to obtaining repeat imaging. Steroid dose should be maintained for at least 1–2 weeks, and then gradually tapered afterwards over the course of several weeks. Most patients will not require any additional therapy, but for those with either progressive symptoms or inability to tolerate a steroid taper, more aggressive treatments may be considered.

Bevacizumab is a monoclonal antibody that inhibits vascular endothelial growth factor (VEGF). It has been studied for treatment of radiation necrosis in two randomized controlled trials.[66,67] In both trials, high rates of radiographic response (100% and 66%) and neurologic symptom improvement (100% and 62%) were observed, which were significantly better than with corticosteroid therapy alone. A dose of either 7.5 mg/kg every 3 weeks or 5 mg/kg every 2 weeks for up to 4 cycles, lower than those typically used in anti-cancer regimens, can be used. Imaging response can be detected on MRI as early as after 2 cycles. Follow-up MRI can be obtained 8–12 weeks after initiation of therapy and steroid taper can begin around 72 hours after cycle #1. Retreatment with bevacizumab is feasible and appears efficacious but is not well-studied. Rates of serious adverse events with bevacizumab were low but included pulmonary embolism, sagittal sinus thrombus, and ischemic stroke. The studies excluded patients with active or high risk of bleeding, recent intracranial hemorrhage or major surgery/trauma, significant cardiovascular disease, abdominal fistula or perforation, or poorly controlled hypertension.

Surgical resection may be necessary for cases of refractory radiation necrosis with significant symptoms, contraindication to medical therapy, or uncertainty as to whether a lesion represents radiation necrosis or tumor recurrence. Surgery can offer rapid decompression leading to reduced steroid requirement but is also associated with significant morbidity as represented by one contemporary series where overall morbidity from surgery was 54%.[68] The authors advocated for the use of surgery for radiation necrosis only in cases where all medical therapy had failed. In recent years, laser interstitial thermal therapy (LITT) has also been used successfully to manage refractory radiation necrosis, and offers a minimally invasive alternative guided by MRI. Retrospective data suggests local outcomes comparable to craniotomy albeit with inferior symptom relief.[69]

Hyperbaric oxygen therapy (HBOT) has only been studied in small, retrospective series.[70–72] In one cohort of 10 patients who underwent HBOT, all either had stabilization of improvement of symptoms and/or imaging findings without severe toxicities.[70] Patients were treated at 2.0–2.4 atmospheres for 20–30 sessions lasting 90–120 minutes each. HBOT should not be used in patients with pneumothorax or at high risk for pneumothorax (i.e., chronic obstructive pulmonary disease, lung blebs/bullae, recent thoracic surgery). Its adoption has been limited due to the need for expensive, specialized equipment and the significant time commitment. Anticoagulants such as warfarin and heparin have also been studied in small retrospective series.[73,74] The larger included 8 patients with radiation necrosis, of whom 5 symptomatically improved after anticoagulation for a total of 3–6 months.[73] One small, retrospective study of 8 patients with radiation necrosis found improvement in edema volume after treatment with vitamin E and pentoxifylline.[75] Overall, the evidence to support HBOT, anticoagulation, and antioxidant therapy for radiation necrosis is weak and these therapies cannot be recommended for routine use.

Neurocognitive Decline

Neurocognitive decline is common in patients with brain metastases, both due to tumor progression as well as from brain metastasis therapy. As many as 90% of patients with brain metastases will have one or more impaired neurocognitive functions at baseline.[76] The management of neurocognitive decline has primarily been directed at prevention rather than treatment. The use of fraction sizes >3 Gy with WBRT appeared to lead to higher risk for developing severe dementia.[77] Two randomized controlled trials have found better neurocognitive preservation with SRS alone compared to SRS with WBRT with similar OS in patients with limited[1–3] brain metastases.[8,9] Later reports have validated the use of SRS alone in patients with up to 15 brain metastases.[78,79]

Hippocampal avoidance IMRT (HA-IMRT) has also been demonstrated to have better neurocognitive preservation in a randomized phase III trial, although patients with leptomeningeal disease or metastases within 5 mm of either hippocampus were excluded.[80] Due to the time and resource-intensive nature of the treatments, both SRS and HA-IMRT are best suited for patients who are either asymptomatic or only mildly symptomatic from brain metastases and with good performance status. Another scenario where HA-IMRT may be ideal is for prophylactic cranial irradiation (PCI) in small cell lung cancer (SCLC). Two recent randomized trials of HA-IMRT for PCI in SCLC with a dose of 25 Gy in 10 fractions found divergent results with one study finding improved cognitive preservation based on delayed free recall,[81] and the other finding no improvement in cognitive preservation based on verbal learning compared to standard WBRT.[82] As there was no difference in brain failure, HA-IMRT should be considered for this group of favorable patients without clinically apparent brain metastases.

The use of the N-methyl-D-aspartate (NMDA) receptor antagonist memantine during and for 6 months after WBRT improved preservation of cognitive function, executive function, processing speed, and delayed recognition although its primary endpoint of delayed recall did not reach statistical significance.[83] Memantine was started at 5 mg AM for week 1, 5 mg BID for week 2, 10 mg AM and 5 mg PM for week 3, and 10 mg BID for week 4 and maintenance. Due to its favorable side effect profile, memantine should be initiated in most patients receiving WBRT. A randomized trial failed to show benefit of prophylactic methylphenidate on fatigue scores in patients undergoing radiotherapy for primary or metastatic brain cancers and methylphenidate should not be administered prophylactically for this indication.[84]

Data regarding treatments for patients who have already developed significant neurocognitive deficits are scarce. A randomized trial comparing a structured multidisciplinary intervention to standard care improved overall patient QOL but failed to show improvement in fatigue in patients receiving radiotherapy.[85,86] Psychostimulants such as methylphenidate and modafinil have been successfully used for cancer-related fatigue.[87–91] A small, randomized study of methylphenidate (immediate release, 10 mg BID) and modafinil (200 mg qAM) given for 4 weeks in patients with primary brain tumors found improvements in processing speed, executive function, and patient-reported fatigue, mood, and QOL.[92] The acetylcholinesterase inhibitor donepezil was studied in a phase III randomized trial among patients ≥6 months after partial or whole brain radiation, which failed to find a difference in its primary composite endpoint, but did result in improved memory, motor speed, and dexterity.[93] The dose of donepezil was 5 mg daily for 6 weeks followed by 10 mg daily for 18 weeks. The benefit of drugs such as methylphenidate, modafinil, and donepezil appear to be greatest in patients with worse baseline functioning. These medications can be considered in patients with significant neurocognitive decline following radiation for brain metastases. The choice of a specific agent should be dependent on the side effect profile and tolerability.

Optic Neuropathy

Radiation-induced optic neuropathy (RION) is one of the most feared complications of intracranial radiation due to its devastating consequences. RION is characterized by progressive partial to complete monocular or binocular vision loss with corresponding contrast enhancement and thickening of the affected anterior visual pathway on MRI with a history of radiation exposure to that anatomic distribution.[94,95] RION is rare below conventionally fractionated radiation doses of 50 Gy or less. The risk of RION is greater when SRS is used to treat brain metastases in close proximity to the optic nerves with an estimated risk of 1% for single fraction SRS doses of >8 Gy and 10% for >12 Gy.[96]

Proven treatment options for RION are limited. Corticosteroids and therapeutic anticoagulation have been studied but do not appear to be effective when given alone.[97–100] HBOT is controversial for this indication, as some studies have reported improvement in vision with HBOT,[101–103] whereas others have found no benefit.[100,104,105] When given, treatment should ideally be initiated within 72 hours of symptom onset.[102] HBOT delivery regimens are variable but range from 14–30 daily sessions lasting 90–120 minutes each at 2.4–2.8 atmospheres. There are case reports of bevacizumab improving RION in conjunction with or following a failed trial of corticosteroids.[106–108] Farooq et al. gave bevacizumab at 7.5 mg/kg q3 weeks for 3 doses with dexamethasone and pentoxifylline while Dutta et al. gave bevacizumab 5 mg/kg alone initially followed by 10 mg/kg q2 weeks for up to 6 doses. Animal models have demonstrated efficacy of the ACE inhibitor ramipril in mitigating RION after SRS, but this therapy remains experimental in humans.[109,110]

When monocular RION occurs, the contralateral eye should be carefully observed with serial clinical exams and imaging as patients are at higher risk of developing subsequent contralateral RION. Unfortunately, there is no evidence that prophylactic treatment is effective in preventing RION, but early HBOT can be considered if imaging changes develop, even in the absence of clinical symptoms.[103]

Neuroendocrine Dysfunction

Historically, patients with brain metastases rarely survived long enough to develop clinically significant hypothalamic-pituitary (HP) axis dysfunction following treatment of brain metastases with radiotherapy and HP dysfunction in this population is therefore not well-described in literature. HP dysfunction is related to dose to the pituitary gland and is rare with doses of <20 Gy but occurs frequently with doses of >50 Gy.[111] Reported latency times range widely from 1–26 years after radiation treatment, but most cases are thought to occur between 1–5 years after radiation.[112] Extrapolating from data in patients with primary brain tumors, it would be reasonable to screen patients with brain metastases surviving longer than 1 year after WBRT annually for endocrine deficiencies with hormone replacement as clinically indicated.[113–115]