Complications of cranial irradiation

INTRODUCTION — Cranial irradiation is used in the treatment of patients with primary or metastatic brain tumors and as prophylaxis for selected patients at high risk of neoplastic involvement of the nervous system. A full understanding of the potential consequences associated with cranial irradiation is needed both for management of potential complications and for proper counseling of the patient and/or family prior to treatment.

The primary factors influencing the likelihood of developing complications include the volume of normal brain tissue treated, the total radiation dose, and the fractionation schedule. Furthermore, the use of concurrent or sequential chemotherapy can significantly affect the incidence and severity of radiation-induced toxicity. In addition, the underlying tumor often can impair neurologic function, making it difficult to assess accurately the separate effect of radiation.

The complications of radiation therapy are commonly divided into acute effects that can occur during or immediately after a course of radiation, and late effects that can develop more than 90 days after the initiation of radiation therapy. Complications that might occur many months or even years following cranial irradiation are generally not as important as a consideration for patients with brain metastases or grade IV primary brain tumors, where a median survival in the range of 6 to 12 months is expected. Late effects are a more important consideration for patients with a much longer life expectancy (eg, low-grade glioma patients or pediatric patients with acute lymphocytic leukemia). The distinction between early and late complications is also important since early complications tend to be reversible while late reactions often are not.

Both the acute and late complications of fractionated cranial irradiation will be reviewed here. Complications of spinal cord and peripheral nerve irradiation and complications of single-fraction cranial radiosurgery are discussed elsewhere.

PATHOPHYSIOLOGY — The effects of radiation can be divided into the effects on the vasculature of the brain as well as the direct effects on glial cells and their precursors.

The vasculature appears to be affected first by radiation. As an example, the radiation injury to blood vessels has been studied in a rodent model, where single fraction doses to the whole brain in the range of 17.5-25 Gy caused endothelial cell death, blood vessel wall thickening, and vessel occlusion leading to ischemia and infarction. Vascular compromise in the brain can also indirectly lead to demyelination, which can occur within a few weeks after radiation and then resolve or continue to progress. If vascular compromise is sufficient enough, coagulative necrosis can ensue.

Radiation can also have direct cytotoxic effect upon glial cells and their progenitors, leading to cerebral atrophy. As an example, young mice that receive whole brain irradiation demonstrate impaired hippocampal neurogenesis and resulting cognitive deficits. Concern regarding a similar effect upon the developing human brain is an important clinical consideration when young children are treated with cranial irradiation for pediatric malignancies

ACUTE REACTIONS — Acute reactions of varying severity can occur during cranial radiation and can linger for several weeks following radiation. These reactions are almost always temporary. The Radiation Therapy Oncology Group (RTOG) has established grading criteria specifically for these acute toxicities.

Tumor progression can mimic the common side effects arising during or shortly after radiation therapy. Patients with unusually severe or worsening neurologic symptoms after cranial radiation should undergo CT or MRI to rule out tumor progression or other causes of neurologic deterioration.

The types of acute toxicity that may be seen include the following:

Acute encephalopathy — Acute encephalopathy may occur when radiation fractions >3 Gy are administered to a large brain volume, particularly in patients with elevated intracranial pressure (ICP). Symptoms include severe headache, nausea, drowsiness, and focal neurologic deficits; in extremely rare cases, death may occur due to cerebral herniation. This effect is thought to be due to disruption of the blood-brain barrier leading to cerebral edema. Acute encephalopathy is generally most severe following the first radiation dose and gradually lessens in severity thereafter.

With a better understanding of the relationship between dose, schedule, and toxicity, acute encephalopathy is rare with modern treatment techniques. However, in a 1970 study, patients received 10 Gy of whole brain radiation therapy (WBRT) in 1 fraction, and 7 percent of the patients died within 48 hours. To avoid this catastrophic complication, whole brain radiotherapy (WBRT) is generally administered in fractions of 3 Gy or less; with this schedule the risk of severe acute encephalopathy is negligible.

Cerebral edema — A transient worsening of pretreatment symptoms can occur early in the course of radiation, probably as a result of transient mild cerebral edema. Corticosteroids can lessen the radiation-induced blood-brain-barrier disruptions and improve the symptoms.

Patients with recognized significant pretreatment cerebral edema should begin oral or parenteral corticosteroids (eg, dexamethasone 4 mg four times daily) prior to initiating radiation.  Maintaining the dose for the first two weeks of radiation therapy can prevent clinical deterioration due to transiently worsened peritumoral edema. However, a short-term increase in corticosteroid dose may be warranted if symptoms are severe.

Nausea and vomiting — Nausea and vomiting are common side effects of cranial radiation. Antiemetics or corticosteroids are used to prevent or mitigate symptoms.

Radiation dermatitis and alopecia — Total or partial alopecia is extremely common after cranial irradiation, and it may be permanent with higher total doses. Alopecia occurs only where the radiation beams traverse the scalp, so patients receiving partial brain irradiation may develop patchy alopecia.

Radiation dermatitis is usually mild and may be treated with soothing moisturizing ointments. Areas of skin receiving particularly high doses may demonstrate dry or moist desquamation that may be protected with requires more intensive bacteriostatic topical treatments. Otitis externa is common if the ear is within the radiation treatment field; serous otitis media may also occur.

Myelosuppression — Clinically significant myelosuppression is unlikely to occur as a result of cranial irradiation. However, myelosuppression may be seen in patients receiving simultaneous treatment to the spine (craniospinal irradiation), which involves treatment to a large volume of bone marrow. Myelosuppression may also be seen in patients receiving concomitant chemotherapy.

Mucositis — Pharyngeal mucositis, characterized by local irritation, erythema, swelling, and pain, is also uncommon unless patients are receiving craniospinal irradiation. During craniospinal irradiation, the spinal fields exit through the oropharynx and mediastinum causing oropharyngeal and esophageal mucositis, respectively. Such mucositis often occurs after about 25 to 30 Gy in conventional fractionation and usually resolves about two to four weeks after radiation is completed.

MRI contrast enhancement — In addition to well-defined clinical syndromes, the routine use of cranial magnetic resonance imaging (MRI) has led to the identification of posttreatment asymptomatic MRI contrast enhancement, a transient radiographic change that follows the use of cranial irradiation.

LATE REACTIONS — By definition late reactions are observed more than 90 days after the start of cranial radiation, though sometimes reactions occurring in the range of 60 to 90 days after radiation are labeled "early delayed" reactions. Late radiation side effects are sometimes, but not always, reversible. The Radiation Therapy Oncology Group (RTOG) has established grading criteria specifically for these delayed toxicities.

Common late reactions to cranial radiation include the following:

Somnolence syndrome — The somnolence syndrome is characterized by extreme sleepiness in conjunction with signs of increased intracranial pressure (eg, headache, nausea, vomiting, anorexia, and irritability). The somnolence syndrome is more common in children than in adults, with the onset typically occurring one to six months after prophylactic cranial irradiation for acute lymphoblastic leukemia. Although moderate fatigue is common after cranial irradiation in adults, an overt somnolence syndrome is rare. The syndrome is occasionally accompanied by transient recrudescence of a previously resolved neurologic deficit. Brain MRI may show non-specific white matter hyperintensity.

The somnolence syndrome usually resolves spontaneously within two to three weeks. Corticosteroid treatment may provide symptomatic relief and reduce the incidence of this syndrome. However, the routine use of corticosteroids for treatment of somnolence syndrome is not widely recommended. Patients with this syndrome are not at increased risk for other late effects.

Transient focal neurologic symptoms — Some patients undergo a delayed transient worsening of symptoms that can suggest tumor progression, with headache and/or recurrence of pretreatment neurologic symptoms and signs. The likely etiology is a combination of tumor response and peritumoral edema or demyelination; corresponding imaging changes can include focally increased enhancement as a result of blood-brain barrier disruption.

Differentiating treatment-related imaging changes from tumor progression can be a major diagnostic challenge. Fluorodeoxyglucose (FDG) PET scanning can be helpful, as can single-photon emission computerized tomography (SPECT) scanning and magnetic resonance spectroscopy. However, a repeat biopsy or surgical decompression may be needed to determine definitively whether symptoms are due to recurrence or are a post-treatment reaction.

Radiation necrosis — Radiation necrosis is a serious complication that develops one to three years after radiation. Clinically, symptoms produced by localized brain necrosis depend upon on the location of the lesion and can include focal neurologic deficits or more generalized signs and symptoms of increased intracranial pressure.

Focal radiation necrosis is caused by vascular endothelial cell damage, resulting in fibrinoid necrosis of small arterial vessels. Occlusion of these vessels results in focal coagulative necrosis and demyelination of the overlying brain parenchyma.

Radiation necrosis typically develops at or adjacent to the original site of tumor, the location that received the highest radiation dose. Rarely, radiation necrosis may develop in normal brain parenchyma that was included in the treatment field of a tumor outside the brain, such as a nasopharyngeal cancer.. In this situation, necrosis is typically manifested by new focal neurologic signs, and imaging studies such as CT or MRI may show an enhancing mass with edema.

Focal radiation necrosis is rarely seen with doses of 60 Gy or less using conventional fractionation. The dose that would cause a 5 percent risk of focal radionecrosis using conventional fractionation is usually estimated to be in the range of 55 to 60 Gy. Radionecrosis is more likely to occur when high doses per fraction are administered. Also, concomitant chemotherapy (especially with carmustine) may increase the risk for necrosis.

  Diagnosis — Differentiating recurrent tumor from radiation necrosis may be very difficult. MRI typically shows a contrast-enhancing mass with white matter changes and edema within or immediately adjacent to the site of the original tumor, which received the highest dose of radiation dose.

Other imaging modalities, including FDG-PET, SPECT, and magnetic resonance spectroscopy may occasionally be helpful. However, biopsy of the suspicious lesion is often needed for a definitive diagnosis.

  Treatment — Surgical decompression of any radionecrosis-related mass effect can provide helpful palliative effect. Corticosteroids usually produce prompt symptomatic improvement, and symptoms of cerebral edema can resolve with conservative management. Therapeutic anticoagulation and hyperbaric oxygen therapy have also been reported to provide possible benefit.

Diffuse white matter injury — Non-specific diffuse white matter changes can be seen on MRI in nearly all patients receiving doses over 55 Gy. Most patients are asymptomatic, at least initially, but the incidence and severity of symptoms appears to correlate with the severity of radiographic changes.

Leukoencephalopathy, the clinical syndrome associated with these radiographic changes, was first described in children receiving prophylactic cranial irradiation and systemic methotrexate. Symptoms include lethargy, seizures, and dysarthria and are usually temporally related to postirradiation systemic methotrexate. In severe cases, symptoms can progress to ataxia, confusion, memory loss, dementia, and, rarely, death.

A similar syndrome has been described in adults, also generally related to chemotherapy delivered around the same time as radiation. Symptoms are similar to those found in children but do not usually occur until one to two years following therapy. Later in the follow-up period, cerebral atrophy may become evident.

Neurocognitive effects — Differentiating adverse effects of cranial irradiation upon neurocognitive function (NCF) from the effects of the underlying malignancy can be very difficult. Within the last decade, prospective studies have been performed to characterize accurately the effects of cranial irradiation upon NCF . It is important for such studies to include long follow-up whenever feasible, because impaired NCF may become evident as a late effect of treatment. Important data on the impact of RT on NCF has been derived from studies in adults with brain metastases and low-grade gliomas, as well as in pediatric patients.

  Brain metastases — Several important observations were made in a large trial of patients receiving whole brain radiation therapy (WBRT) for brain metastases. First, over 90 percent of patients had some evidence of NCF impairment at baseline, especially in the realms of fine motor control, executive function, and memory. The extent of impairment at baseline was correlated with the indicator lesion volume prior to treatment. Secondly, higher baseline NCF was associated with longer overall survival duration. Finally, changes in NCF two months after treatment correlated closely with changes in the indicator lesion. Patients with radiographic evidence of progressive disease had a significant decrease in NCF across all categories evaluated, whereas patients with partial response demonstrated more modest changes and even improvement in some areas.

Other important results were noted in the only randomized study comparing surgical resection versus surgery plus WBRT in patients with a solitary brain metastasis. The rate of recurrence within the brain was 70 percent for patients who had surgery alone versus only 18 percent in patients who received surgery plus WBRT (p<0.001). Consequently, patients who received the combination were less likely to die of neurologic causes than patients who were treated with surgery alone (p = 0.003).

Thus, it seems likely that a major determinant of NCF in patients with brain metastases is the burden of disease in the brain, perhaps even more than the effect of WBRT. However, it is difficult to estimate the separate quantitative impact from either of these contributing effects in this population.

  Low-grade glioma — The effects of cranial irradiation upon NCF in patients with low-grade glioma are particularly important since these patients often enjoy long survivorship. There has been one randomized study comparing immediate postoperative radiotherapy (RT) versus delayed RT given at the time of tumor progression. The time to neurological progression was delayed significantly by the use of immediate postoperative RT. Unfortunately, NCF outcomes were not studied in detail.

In another study, 195 patients with low-grade glioma, including 104 who had received RT between 1 and 22 years previously, were compared with a control group of healthy individuals. The low-grade glioma patients had lower ability in all cognitive domains compared with controls. However, the use of RT was not associated with poorer cognitive function, except among patients who received a daily radiation dose greater than 2 Gy, where compromised memory was observed. In the same study, anticonvulsant medication was associated with a sixfold increase in risk of compromised capacity for perceptual tasks and in attention and executive function. Overall, as with the observations from patients with brain metastases, the presence of a low-grade glioma was the major cause of compromised NCF, not the RT.

In a prospective study of 203 patients with low-grade glioma randomized to receive either a lower dose (50.4 Gy in 28 fractions) or a higher dose (64.8 Gy in 36 fractions) of post-operative RT, 5 percent of those without tumor progression were noted to have had deterioration of NCF five years after treatment, when assessed by the Folstein Mini-Mental State Examination (MMSE). However, the significant limitations of this particular instrument for assessing NCF were discussed in an accompanying editorial.

In addition, in any analysis of the late effects of cranial radiation upon NCF, the potential contributing effects of other anti-cancer treatments must be carefully weighed. For example, systemic chemotherapy for breast cancer or lymphoma is associated with measurable deficits in NCF.

  Pediatric patients — Cranial irradiation given to pediatric patients warrants special attention in view of the potential for impairment of growth and development of the central nervous system. As an example, in one small study of children diagnosed with a posterior fossa tumor at an age less than 3 years, the seven children who received postoperative radiotherapy manifested lower intelligence quotient (IQ) and motor domain scores than the 14 who had not received radiotherapy. There is likely a dose-response relationship in terms of the degree of effect of radiation upon subsequent cognitive development.

Prophylactic cranial irradiation (PCI) is often administered to children with high-risk leukemia to reduce the risk of relapse within the central nervous system. Furthermore, high-dose methotrexate, when given in combination with cranial PCI, appears to have a marked impact upon the risk of subsequent NCF impairment. Modern regimens typically are limited to a total dose of 18 Gy, based upon observations from older studies that higher doses of PCI were associated with unacceptable late toxicity.

Follow-up analyses from a study of 201 patients treated in this manner indicate that after a median interval of 9.1 years, the children were noted to have normal IQ and memory but compromised capacity in a complex figure drawing task. Children who were treated prior to 3 years of age had average overall IQ but compromised verbal skills.

Efforts to reduce further the risk of late toxicity have included the use of twice-daily radiotherapy given in smaller doses per fraction. However, it has been observed that this strategy seems counter-productive: no significant advantage has been observed in terms of overall late cognitive toxicity, and there is a strong suggestion of impaired tumor control with twice-daily treatment, compared to conventional once-daily treatment.

The late effects of RT on NCF in children treated for low-grade glioma have also been studied within the context of a prospective trial, but results are not yet available. In a trial of 22 children who had been treated for medulloblastoma on a Pediatric Oncology Group (POG) study that randomly assigned them to either standard dose (36 Gy) or reduced dose (23.4 Gy) craniospinal irradiation after surgical resection, the children receiving the reduced dose had significantly less severe neuropsychologic sequelae than the higher dose children.

The especially high risk of late radiotherapy effects in very young children with brain tumors prompted research into the use of chemotherapy to delay the use of radiotherapy in children diagnosed with brain tumors before the age of 3 years. However, this strategy is not advisable for all tumor types.

Cerebrovascular effects — It is unusual to irradiate a large volume of the carotid arteries during the treatment of intracranial tumors, but irradiation to these vessels is common in the management of head and neck cancers. However, radiation effects upon the carotid arteries, and the possible consequential effects in causing cerebrovascular accidents, are probably an under-studied clinical complication. As an example, it has been suggested that RT might contribute to an increased risk of stroke in women with breast cancer.

Children are probably much more susceptible to radiation-induced vasculopathy than adults. The supraclinoid region of the internal carotid seems especially vulnerable. Occlusive vascular disease, similar in angiographic appearance to Moyamoya disease, presenting with ischemic events and stroke, has been reported after radiotherapy for optic glioma; neurofibromatosis type 1 (NF1) seems to increase the risk markedly. Intracerebral cavernous malformations have also been reported in children after cranial irradiation.

Effects on the eyes and optic pathways — RT may have a series of acute and delayed effects on the eyes and optic pathways. As an example, during treatment, radiation may stimulate the retinal photoreceptors, causing patients to perceive light sensations transiently immediately after radiation treatment to fields that include segments of the optic pathways.

  Cataracts — Radiation-induced cataracts may result after low doses of radiation to the lens of the eye. Patients typically present with painless visual impairment two to eight years following radiation therapy. In a retrospective study of over 1000 patients who received total body irradiation for bone marrow transplant, either 10 Gy in one fraction or 12 Gy in conventional fractionation, 60 percent of the patients receiving one dose and 43 percent of the patients receiving fractionated radiation developed cataracts. Cataract development is strongly correlated with the chronic use of steroids in these patients. The treatment is the same as in non-radiation-induced cataracts, namely cataract removal and prosthetic lens placement.

  Optic neuropathy — Optic neuropathy typically presents with painless monocular or bilateral visual impairment developing over one to several weeks and beginning around 6 to 24 months after radiation. The incidence of mild, transient radiation effects on the optic nerves is not well reported. Fortunately, severe optic neuropathy after cranial irradiation is uncommon with doses less than 54 Gy to the optic chiasm and less than 59 Gy to an optic nerve. In one study, no injuries were seen in 106 optic nerves receiving less than 59 Gy.

Fraction size greatly impacts the incidence of optic neuropathy, with larger fraction sizes associated with higher rates of neuropathy. In a study of 55 patients receiving 45 to 55 Gy for pituitary adenomas or craniopharyngiomas, 18 percent of patients receiving fraction sizes larger than 2.5 Gy developed optic neuropathy, while no patients receiving fraction sizes less than 2.5 Gy developed optic neuropathy In the more severe cases of optic neuropathy, steroids are unlikely to provide benefit. Hyperbaric oxygen has provided limited success.   Anticoagulation is sometimes tried, with uncertain results.

  Retinopathy — Retinal injury due to retinal ischemia can also occur after radiation. Radiation retinopathy is often asymptomatic and is found incidentally on fundoscopic exam. If symptomatic, it presents as painless loss of vision months to years after radiation. Risk factors for radiation retinopathy include radiation dose, prior chemotherapy, and diabetes. In one study of 64 patients receiving radiation therapy for head and neck tumors, 27 eyes in 26 patients developed symptomatic radiation retinopathy resulting in visual acuity of 20/200 or worse. Fourteen of the injured eyes developed rubeosis iridis and/or neovascular glaucoma. No patient who received less than 45 Gy developed retinal complications, but the incidence of retinopathy increased with higher doses.

  Cortical blindness — Cortical blindness has been reported after a combination of surgical resection, radiation therapy, and interleukin-2 for a patient with a brain metastasis of renal cell carcinoma in the left occipital cortex. MR imaging disclosed signal abnormalities without mass effect in the white matter of the parietal and occipital lobes bilaterally, including the optic radiations.

Effects upon hearing — Tinnitus and high frequency hearing loss are acute effects occasionally occurring during cranial irradiation. Often, these symptoms are due to radiation-induced otitis media, causing mucosal vasodilatation and eustachian tube edema. Symptoms often resolve spontaneously, but occasional patients may need myringotomy for symptomatic relief.

Hearing loss months to years after radiation is usually due to sensorineural damage that occurs in patients treated for head and neck cancer. In one study of 294 patients with nasopharyngeal carcinoma, deterioration of bone conduction threshold and pure tone average were noted in 31 and 14 percent, respectively. These measurements normalized in less than half at two years. Fractionated stereotactic radiotherapy or radiosurgery for acoustic neuromas is associated with a risk of hearing loss.

Hearing loss is dose related. In one study, no patients receiving less than 59 CGE (cobalt-gray-equivalent) to the normal cochlea or cranial nerve VIII developed hearing loss. However, almost 66 percent of those receiving more than 60 CGE to either structure developed significant loss.

Cisplatin, which itself may cause hearing loss, seems to heighten the radiation effect. Some reports suggest that up to 50 percent of long term survivors of medulloblastoma who have received cisplatin and radiation therapy require hearing aids, although the incidence may be lower with modern radiation techniques that spare key structures. Cochlear implantation can be considered in patients with complete deafness after radiotherapy.

Endocrinopathies — Hypothalamic and pituitary endocrinopathies occur commonly in both children and adults following radiation which includes these structures. Patients often manifest serum hormone irregularities long before clinical symptoms develop; therefore, it can be very helpful to monitor patients who are at risk for radiation-induced hormone deficiencies. If symptoms occur, they depend on the specific hormone deficiency. The time course of dysfunction is variable, but may occur after only 20 Gy.

In one study of 32 adult and pediatric patients who received between 39.6 Gy and 70.2 Gy to the hypothalamic/pituitary region for brain tumors, 91 percent had post-treatment endocrine abnormalities. Sixty-one percent of postpubertal patients had hypogonadism, 32 percent had cortisol deficiencies, and 28 percent had symptomatic thyroid deficiencies. In another study, 10 of 13 adult long-term survivors of low grade glioma had one or more asymptomatic serum hormone abnormalities after treatment with 45 to 61.2 Gy.

Growth hormone deficiency with growth failure is an important problem in children treated for nonpituitary-related primary brain tumors. . In one study of 34 prepubertal children with brain tumors receiving radiation, 14 (41 percent) developed growth hormone deficiency, and 12 percent developed primary hypothyroidism.

Radiation-induced brain tumors — There is an increased risk of meningiomas, nerve sheath tumors, and malignant gliomas with cranial or craniospinal radiation therapy. (See "Risk factors for brain tumors"). Meningeal cells appear to be particularly susceptible to effects of ionizing radiation. The median induction period from cranial radiation to the development of the new tumor may be lengthy, even more than 10 years.

Children undergoing cranial irradiation are at risk for radiation-induced brain tumors. Among a cohort of 8831 children diagnosed with ALL and enrolled on Children's Cancer Group therapeutic protocols between 1983 and 1995, 19 patients had developed a second brain tumor. The relative risk for brain tumors among children who had received 18-24 Gy of cranial irradiation was significantly increased to 2.4 (95% confidence interval, 1.1-5.2). In an analysis of one of the St. Jude Children's Research Hospital leukemia protocols, the combination of prophylactic cranial radiotherapy and antimetabolite therapy resulted in an unexpectedly high frequency of brain tumors (6 of 52, 13 percent), possibly magnified by a genetic defects in thiopurine methyltransferase in the affected individuals

Among the 198 children treated within the Pediatric Oncology Group of prolonged postoperative chemotherapy and delayed irradiation for children diagnosed with a brain tumor at less than 3 years of age, five developed second malignancies: choroid plexus carcinoma (2 children), ependymoma (1 child), desmoplastic infantile ganglioglioma (2 children), and medulloblastoma (1 child). The interval from diagnosis of initial tumor to second malignancy ranged from 33 to 92 months.

CONCLUSIONS — Cranial radiation is an integral component of the management of patients with primary and metastatic brain tumors and certain types of leukemia. However, cranial radiation is associated with risks of various acute and late toxicities, and these risks should be acknowledged and discussed during the pretreatment counseling session with any patient for whom cranial radiation is recommended.