Posttherapy Neurologic Sequelae
Greater recognition of adverse sequelae from multimodal therapy has accompanied the increase in pediatric brain tumor survival over the past three decades. This chapter summarizes the current knowledge about late neurologic sequelae, including neurologic and neurosensory impairment, neurovascular morbidity, and neuroendocrine dysfunction. Cognitive impairment and secondary malignancies are addressed in Chapters 48 and 49, respectively. Our current understanding of the incidence, developmental aspects, risk factors including chemotherapy and radiation therapy (RT) dosimetry, and management of neurologic sequelae will be reviewed. Long-term neurologic impairment has been mostly attributed to post-RT effects and less so to chemotherapy, tumor location, and surgery. However, research on late effects is mostly available from survivors who were treated with less chemotherapy than is used today, and future studies are needed to examine current treatment regimens. Nevertheless, an increasingly significant body of knowledge about adverse neurologic sequelae has already had an impact on treatment regimens for pediatric brain tumors (e.g., delaying or eliminating RT in very young children), and additional research on patients treated with modern therapy will build on these paradigms. It is becoming increasingly clear that clinicians and clinical researchers must carefully characterize, document and study late effects to improve treatment for brain tumor survivors and to inform future decisions about treatment that will have impact on long-term quality of life.
A large part of the data reported in this chapter comes from the Childhood Cancer Survivor Study (CCSS), a national longitudinal study of over 14,000 five-year survivors of childhood cancer diagnosed and treated between 1970 and 1986.1 The advantages of this study include its large size, diverse patient and treatment characteristics, chart extraction for baseline data, and prolonged follow-up period, but the disadvantages are the reliance primarily on self-report and the fact that data from a later cohort has not yet matured for analysis. When available, we provide representative data from other studies as well.
Neurologic and Neurosensory Impairments
Neurologic sequelae from pediatric brain tumor treatment may affect both quality of life and long-term survival. Long-term neurologic effects include focal motor or sensory deficits, headaches, seizures, vascular abnormalities and stroke, as well as secondary neoplasms ( Table 47.1 ).
Focal Neurologic Deficits
Although the incidence of motor or coordination problems is highest during initial diagnosis and treatment, one or more motor problems developed at least 5 years after treatment in 4.6% of long-term survivors in the 2003 CCSS study, and the relative risk was 12.5 times that of an unaffected sibling. Children with at least 50 Gy RT to frontal brain regions had double the risk for motor problems compared with those who received less than 30 Gy.2 It is not known to what extent these late focal deficits may be attributed to radiation necrosis or are secondary to other later neurologic effects discussed below. Late developing cranial neuropathies are also reported, with the optic nerve being the most affected followed by the hypoglossal nerve.3 In addition to RT effects, chemotherapeutic agents such as high-dose arabinosylcytosine (Ara-C) and methotrexate are known to cause a subacute neurotoxicity with focal neurologic deficits characterized by paresis, anesthesia, pseudobulbar palsy, and visual disturbances. Furthermore, late neurotoxicity from chemotherapy may cause a rare progressive demyelinating leukoencephalopathy ( Fig. 47.1 ) associated with limb spasticity, dementia, or coma.4 Iphosphamide has been associated with short-term cerebellar dysfunction, weakness, and cranial nerve dysfunction that is transient with no apparent long-term neurologic sequelae, although it is has not been studied recently.5 Patients receiving RT are usually more sensitive to neurotoxicity from intrathecal chemotherapy, and this must be taken into account when planning treatments. Thus, there are many etiologies for focal neurologic dysfunction in survivors of pediatric brain tumors, and these factors must be considered, in addition to tumor recurrence or secondary malignancy.
With regard to long-term postsurgical neurologic morbidity, increasing attention has been paid to the cerebellar mutism/posterior fossa syndrome. Characterized by paucity of speech, severe hypotonia, ataxia, and emotional lability in the postoperative periods, the mutism typically resolves, but it is becoming increasingly clear that children with this syndrome, occurring in up to 30% of patients with medulloblastoma, are at increased risk for long-term neurologic sequelae and cognitive deficits.6 Although early research attempted to identify specific risk factors, larger studies have shown that the development of the syndrome cannot be predicted by preoperative patient characteristics including tumor size or the presence of hydrocephalus, or by findings on magnetic resonance imaging (MRI) immediately postoperatively. Recently diffusion tensor imaging studies have supported the view that the syndrome occurs as a result of damage to dentatethalamocortical outflow tracks.7 Although some institutions have found alteration in surgical technique may help reduce the incidence, this is yet to be proven. Future studies may address medical treatment and targeted rehabilitation for executive dysfunction.
Focal neurologic impairments
Cerebellar mutism syndrome
Cerebrovascular disease (stroke, moyamoya, cavernous malformation)
Headaches in survivors of pediatric brain tumors have multiple possible etiologies and are therefore a frequent source of neurologic consultation. Headache frequency may vary with tumor type, with an incidence as high as 39% in survivors of pediatric craniopharyngioma.8 Early headaches from chemotherapy such as cyclophosphamide and mafosfamide may be diminished with slower infusions.9 Headaches associated with RT frequently occur during treatment; however, the incidence of late headaches is unknown. Complicated, or “stroke-like” migraines with episodic neurologic dysfunction have been described, with risk factors including posterior tumor location, 50 Gy or higher cranial radiation therapy (CRT) dose, and latency longer than 1 year from irradiation.10 Radiation-induced endovascular changes may cause migraine-like headaches or, alternatively, aggravate a preexisting tendency to migraine.10 Similar to headaches in the general population, the incidence of migraines and other severe headaches in survivors of pediatric brain tumors treated with RT is likely influenced by genetics, and a family history obtained from pediatric brain tumor survivors should include history of headaches. Headaches should be assessed at every follow-up visit, not only to screen for a pathological cause, but also because effective treatment is available and may improve quality of life.
In the 2003 CCSS report, seizure disorders were reported in 25% of pediatric brain tumor survivors, with 6.5% of the population having their first seizure more than 5 years after treatment. As mentioned above, seizures may be seen in patients with leukoencephalopathy from chemotherapy, although this condition is rare. RT doses of at least 30 Gy to any cortical segment have been associated with a twofold elevated risk for a late seizure disorder.2 Although seizures are most often associated with cortical lesions, they may also occur in patients who received RT for posterior fossa tumors.11 A new-onset seizure is typically assessed with a routine electroencephalogram and an MRI scan to rule out recurrence or other structural abnormality including stroke. Unlike in the general population, most practitioners treat a first-time seizure in a brain tumor survivor with an antiepileptic medication to reduce the likelihood of recurring seizures, which may cause injury or limit independence, particularly in patients who drive. The length of time for treatment varies, but as with other causes of seizures, neurologists usually continue medication for late seizures until the patient is seizure free for at least 2 years before considering a trial off the medication (unlike with seizures that occur pretreatment, in which case anti-epileptic medications are discontinued earlier).
Hearing loss is common in survivors of pediatric brain tumors. A Children′s Oncology Group (COG) review of hearing loss determined that the major risk factors are younger age, higher cumulative dose of chemotherapy, central nervous system (CNS) tumors, and concomitant CNS radiation.12 The CCSS found that the incidence of hearing loss more than 5 years after brain tumor treatment was 5.9%, and the relative risk compared with unaffected siblings was 17.3%.2 Higher rates of hearing loss were associated with surviving a primitive neuroectodermal tumor (PNET) compared with an astroglial tumor or ependymoma. RT can lead to multiple forms of ototoxicity, conductive hearing loss, sensorineural hearing loss, and tinnitus.12 Radiation exposure > 50 Gy to the posterior fossa was associated with a 3.7 higher likelihood of developing hearing impairment, compared with a dose < 30 Gy.2 In children with low-grade gliomas treated with 54 Gy CRT, the maximum cumulative incidence was 5.7% at 5 and 10 years.13 A prospective study of 78 pediatric patients with localized brain tumors treated in 1997 to 2001 who had not received platinum-based chemo found hearing loss in 14% of patients, with onset most often 3 to 5 years post-CRT and increased incidence at greater than 40 Gy.14 Based on Merchant et al′s15 2004 finding that high-frequency sensorineural hearing loss is uncommon at cumulative doses of radiation < 35 Gy given over 6 weeks, it has been recommended that that the average cochlear dose should not exceed 30 to 35 Gy delivered over 6 weeks.12,15 Although the mean total dose to the cochlea during fractionated RT is a significant risk factor, the effect of dose per fraction is not clear16 ( Table 47.2 ).
High-cumulative chemotherapy dose
Concomitant radiation therapy
Synergistic toxicity from chemotherapy combined with RT has been shown in prospective and retrospective studies. Platinum-based chemotherapy including mainly cisplatin but also carboplatin and oxaliplatin has resulted in sensorineural hearing loss due to cochlear damage, specifically the outer hair cells of the organ of Corti.17 Rates of hearing loss are as high as 65%, and long-term follow-up reveals continued worsening of hearing loss. A retrospective study of 123 pediatric patients found overall ototoxicity from cisplatin or carboplatin worsened over time, with ototoxicity in 29% (grade 2), 14% (grade 3), and 1% (grade 4) at 2 years, compared with 4% (grade 2), 1% (grade 3), and 0% (grade 4) during treatment.18 Hearing loss may cause more long-term sequelae in younger children who are in a developmental window when phonations of speech at higher frequencies are acquired.4 Hearing loss is considered irreversible, although hearing aids can help. Based on these findings, new trials are requiring more frequent hearing tests.
Further studies are needed to determine whether hearing loss will decrease with more conformational radiation, although findings will be complicated by the increased use of cisplatin chemotherapy. COG guidelines recommend a complete audiological evaluation for survivors at risk for hearing loss on entry into long-term follow-up, and more frequently if any change is noted, with longer follow-up in patients who received CRT.12
Although ocular late effects are relatively rare, they may be devastating to brain tumor survivors, particularly when there are concordant cognitive and neurosensory impairments. The CCSS found that patients surviving pediatric brain tumors are at elevated risk for legal blindness in one or both eyes, double vision, and cataracts more than 5 years after their diagnosis.2 Using more recent data from the full CCSS cohort of 12,500 long-term pediatric cancer survivors, Whelan and colleagues (2010) found that primary CNS malignancy survivors had an increased incidence of double vision (5.7%), dry eyes requiring treatment (5.3%), cataracts (2.1%), and legal blindness (1.3%).19 Legal blindness at 20 years postdiagnosis in one or both eyes occurred with higher rates in all patients who received RT doses greater than 5 Gy to the eye, greater than 30 Gy to the temporal lobe, and any dose to the posterior fossa. The cumulative incidence for legal blindness in survivors of CNS tumors was 1.3%. Moreover, these late effects increased with time, with cumulative incidence rates increasing up to 25 years after diagnosis. An older study demonstrated that both total dose and fractionated dose contribute to the risk for radiation-induced retinal injury.20 In addition to ocular damage, radiation-induced optic neuropathy is dose dependent, with higher rates associated with total doses more than 50 to 55 Gy, single doses more than 10 Gy, and radiation fraction size greater than 2 Gy.20,21
Chemotherapy-induced peripheral neuropathy results from damage and dysfunction of peripheral nerves and occurs with vinca alkaloids (vincristine), platinum drugs (cisplatin, oxaliplatin, carboplatin), taxanes, high-dose Ara-C, and methotrexate ( Table 47.3 ). It has also recently been reported with biologic agents like the antiinflammatory, antiangiogenic drug thalidomide and the monoclonal antibody proteosome inhibitor bortezomib.22 Manifestations range from uncomfortable to debilitating and may be sensory (tingling, pain, numbness), motor (foot drop, impaired walking, dropping things), or autonomic (constipation, orthostatic hypotension, cardiac arrhythmias). Specific deficits vary depending on which nerves are involved. Hyporeflexia is commonly seen as a result of damage to unmyelinated small nerve fibers. Sensory impairment occurs in a stocking-glove distribution, progressing toward the trunk, because the damage is length dependent and affects most distal nerve regions first. Cisplatin damages large sensory fibers resulting in diminished proprioception and vibration, whereas vincristine tends to cause damage to the spinothalamic tract affecting pain and temperature. There may also be damage to cranial nerves resulting in vocal cord paralysis, diplopia, and facial diplegia, which may be difficult to distinguish from the effects of the tumor itself. The pathophysiology is unclear but may include damage to neuronal cell bodies in the dorsal root ganglion as well as axonal damage via transport defects, energy failure, or ion channel dysfunction.22 Metabolic disturbances, nutritional deficiencies, and opportunistic infections increase the risk in patients with cancer.
Estimates of the incidence of chemotherapy-induced peripheral neuropathy vary widely but may be as high as 50% in patients treated with vincristine. The incidence is cumulative dose-dependent, with symptoms commonly seen in patients who receive vincristine > 6 mg/m2 and cisplatin > 400 mg/m2.23 Adolescents who receive vincristine and RT are at a higher risk for severe neuropathy than are younger children.24 Severity is quite variable even in children of the same age who receive similar RT doses, and other than having a preexisting or inherited neuropathy such as Charcot-Marie-Tooth, there are no known predictive risk factors. It is becoming clear that genetic risk factors and pharmacogenetics play a role, as evidenced by the recent finding that acute lymphoblastic leukemia (ALL) patients who express the cytochrome P-450 (CYP) 3A5 genotype experience less vincristine-induced neuropathy than do those who express CYP3A4.25
Neuropathic symptoms during brain tumor treatment may result in treatment delays, dose reduction, or discontinuation of therapy, which may impact survival. Although chemotherapy-induced peripheral neuropathy is considered reversible, effects may last months to years following treatment.26 Vitamin B6 and B12 and other neuroprotective agents have been ineffective at preventing neuropathy in this population. Although antiepileptics such as gabapentin, carbamazepine, and topiramate as well as amitriptyline are often used for symptom management, efficacy has not been proven.27