Medulloblastoma, a cerebellar primitive neuroectodermal tumor (PNET) is the most prevalent malignant tumor of the central nervous system (CNS) in children. In the United States, medulloblastoma comprises 20% of all pediatric intracranial tumors, with approximately 540 cases diagnosed yearly. Found more commonly in young children than adults, it has a bimodal age distribution with peak incidence at 3 to 4 and 8 to 9 years old.1 Currently classified as grade IV by the World Health Organization, medulloblastoma is a highly aggressive tumor with frequent metastasis and early local invasion. Arising primarily in the cerebellar vermis, tumor cells typically expand throughout the cerebellar cortex, nuclei, and deep white matter and may ultimately invade the brainstem.2 Many children initially present with symptoms of increased intracranial pressure due to tumor growth into the subarachnoid space and fourth ventricle causing obstructive hydrocephalus.3 A highly metastatic tumor, more than 30% of children have evidence of CNS metastasis at the time of diagnosis.2 Evidence demonstrating both radio- and chemosensitivity in PNETs supports multimodality therapy, including surgical excision, external beam radiation therapy, and multiagent chemotherapy.1, 3, 4
The ototoxic effects of chemotherapeutic agents, specifically the platinum compounds, as well as radiation therapy, have been well documented and remain a challenging consequence of medulloblastoma treatment. Although current multimodality therapy has shown cure rates of over 80%, up to 50% of these patients will experience significant ototoxicity requiring amplification.5, 6 In patients who progress to profound sensorineural hearing loss (SNHL), auditory rehabilitation in the form of cochlear implantation (CI) should be considered. To date, however, issues unique to cochlear implantation in this population have not been examined in the literature.
Ototoxicity is not the only otologic complication that can result from comprehensive, multimodality treatment of pediatric medulloblastoma. Effects of temporal bone chemoradiation vary from otitis externa to chronic serous otitis media and conductive hearing loss due to osteoradionecrosis.7–9 Pathologic changes in the external auditory canal, tympanic membrane, middle ear, mastoid bone, and inner ear may all impact future cochlear implantation and must be incorporated into treatment planning. Additional possible treatment sequelae, including impaired wound healing, long-term endocrinopathies, declining neurocognitive function, intracranial ischemic events, and secondary malignancies, may have global or specific effects on post-treatment medulloblastoma patients.3, 10, 11 Individually or in sum, each of these sequelae may impact the eligibility, surgical execution and utilization of cochlear implantation in this population and should be identified and addressed.
This study aims to identify and examine issues of cochlear implantation in patients with a history of pediatric medulloblastoma. Details of neuro-oncologic treatment, specifically surgical excision, chemoradiation, and associated complications, including middle ear and temporal bone pathology, are highlighted and reviewed. Specifics related to cochlear implantation, including eligibility, treatment of preoperative middle ear disease, operative and postoperative course, and performance data are presented and analyzed. Concerns related to long-term tumor surveillance, recurrence, and future malignancy are also discussed.
MATERIALS AND METHODS
A retrospective review was conducted of cochlear implant candidates from November 1984 to June 2008, and three patients with a history of childhood medulloblastoma were identified. Each medical record was reviewed for details of neuro-oncologic treatment and associated sequelae, middle ear pathology, perioperative cochlear implant course, and postimplantation performance data. Factors included in the analysis were age of diagnosis of medulloblastoma; surgical treatment of medulloblastoma; total radiation dose; chemotherapeutic agents used, duration and dose; intracranial complications and sequelae of medulloblastoma treatment; otologic manifestations after medulloblastoma treatment; surgical treatment for otologic disease; age at candidacy for CI; preimplantation hearing loss; intraoperative management at time of CI; postimplantaton complications; and performance data.
All patients had histologically confirmed cerebellar medulloblastoma and underwent gross total resection via suboccipital craniectomy by a single surgeon. Hyperfractionated radiotherapy to the craniospinal axis, including posterior fossa boost, totaling 72 Gy was administered to each patient within 4 weeks of surgery. Monthly, multiagent chemotherapy was administered intravenously in the three consecutive courses: 1) cisplatin (20 mg/m2) daily with etoposide (50 mg/m2) for 5 days; 2) cyclophosphamide (900 mg/m2) daily for 2 days plus vincristine; 3) carboplatin (150 mg/m2) weekly for 4 weeks with vincristine (1.5 mg/m2) every other week ×2. Taken together, these three courses constituted one cycle. The first cycle was begun 1 month after completion of radiation therapy, and a total of three cycles were administered over 9 months.4 Total cisplatin dose in this regimen is 300 mg. Monitoring of side effects, including body weight, mean white blood count, platelet count, and physical exam with otoscopic evaluation were performed throughout chemo- and radiation therapy. Magnetic resonance imaging (MRI) of the brain and spinal cord were performed at the conclusion of treatment, every 6 months for 5 years, and yearly thereafter.
Audiometric evaluation was performed at the conclusion of chemotherapy and upon development of otologic pathology thereafter. Pure tone and speech audiometry, and speech perception testing using word and sentence recognition were performed in an Industrial Acoustics Company (New York, NY) soundproof suite using recorded material presented once in an auditory-alone condition. GSI high-performance sound field loud speakers mounted 1 m in front, right, and left of the patient were used in each evaluation. Prior to implantation, data are reported for the aided condition. After implantation, subjects were evaluated at the device settings used by the patient.
Pre- and postimplantation speech recognition testing included the Consonant-Nucleus-Consonant (CNC) word test and the Hearing-In-Noise sentence test (HINT) in quiet and in noise.12, 13 The CNC tests scores of open-set speech recognition with words (CNC-W) or phonemes (CNC-P) and includes 10 lists with 50 words each, each approximating the phonemic distribution of the English language. Scoring is reported as percent words or phonemes correct. The HINT test consists of 250 three- to seven-word sentences divided into 25 lists of 10 sentences each. The test can be administered in quiet (HINT-Q) or noise (HINT-N) with a signal/noise ratio of +10 dB. Scoring is based on percentage of words correctly identified.
All patients underwent complete evaluation for cochlear implant candidacy. In addition to audiometric and speech perception testing, preoperative computed tomography scan and MRI were obtained in all patients. All patients remained free of intracranial disease as assessed by last MRI.
Cochlear implantation was performed by a single surgeon at the primary institution and included intraoperative implant functional evaluation (impedance and neural response telemetry,) plain film confirmation of placement, and perioperative antibiotics (intraoperative cefazolin and postoperative amoxicillin clavulanate acid for 7 days). Device stimulation and external processor programming occurred 3 to 4 weeks after surgery.
Medulloblastoma Treatment and Nonotologic Sequelae
Each patient was diagnosed with medulloblastoma in childhood at ages 8, 4, and 14, respectively (Table I). As defined above, each underwent gross subtotal resection via suboccipital craniectomy with postoperative hyperfractioned craniospinal radiotherapy and multiagent adjuvant chemotherapy. Patients 1 and 3 tolerated treatment without dose manipulation, however patient 2 required reduction in cisplatin dose in the third cycle due to high-frequency hearing loss. Only patient 3 experienced an immediate post-treatment complication of hydrocephalus requiring a ventriculoperitoneal shunt.
Table I. Patient Demographics and Treatment Complications
|1||8 (1989)||CVA||7||Radical tympanomastoidectomy with blind sac closure||25||Nucleus CI22M||Wound dehiscence treated conservatively|
|2||4 (1999)||Growth and thyroid hormone deficiency||7||Tympanoplasty with OCR, radical tympanomastoidectomy with blind sac||13||Nucleus CI22M||None|
|3||14 (1992)||VPS, growth, and thyroid hormone deficiency||12||Tympanoplasty with OCR, canalplasty, modified radical tympanomastoidectomy||29||HiRes 90K||Mastoid-cutaneous fistula requiring closure|
Long-term post-treatment sequelae were experienced by all patients (Table II). Approximately 12 years postchemoradiation, patient 1 developed left-sided hemiparesis due to a right cerebrovascular accident believed to be caused by CNS radiation necrosis. Neurologically, each patient experienced nonspecific issues of neurocognitive dysfunction, diagnosed as attention deficit and learning difficulties. Patients 2 and 3 also demonstrated multiple endocrinopathies and both required oral supplementation with growth and thyroid hormone.
Table II. Cochlear Implantation Performance Data
Otologic Sequelae: Chronic Ear Disease
All patients demonstrated chronic ear disease between 4 and 16 years after completion of chemoradiation. Each required multiple surgical procedures, including tympanoplasty with ossicular chain reconstruction (patients 2 and 3), transcanal canalplasty for medial canal stenosis (patient 3) and tympanomastoidectomy (all patients) (Table I). Between 3 and 12 months prior to cochlear implantation, patients 1 and 2 underwent a canal wall down, radical mastoidectomy, and ear obliteration with external auditory canal blind sac creation in preparation for future implantation.
Otologic Sequelae: Hearing Loss
All patients demonstrated post-treatment hearing loss and began amplification 10, 1, and 3 years, respectively, following medulloblastoma treatment. Length of preimplantation amplification was 7 years (patients 1 and 2) and 12 years (patient 3.) Prior to cochlear implantation, pure-tone averages were 105 to 100 decibel hearing level (dBHL) in the implanted ear and 65 to 80 dBHL in the contralateral ear. Speech detection thresholds in the implanted ear were 90 to105 dBHL and 50 to 85 in the opposite ear. Preoperative speech recognition scores were obtainable for patients 1 and 3. Patient 1 scored 8% on CNC-W, 24% on CNC-P, and 19% on HINT-Q. Patient 3 had a score of 0% for all preoperative tests, including CNC-W, CNC-P, HINT-Q, and HINT-N. Patient 2 had difficulty repeating the words back in an open set and could not be evaluated using CNC or HINT testing.
Patients underwent cochlear implantation at 17, 8, and 15 years following medulloblastoma treatment at ages 25, 13, and 29, respectively. All patients were implanted in their right ear. Patients 1 and 2 received the Nucleus Freedom Contour Advance (Cochlear, Lane Cove, NSW, Australia) and patient 3 chose the Advanced Bionics HighRes 90K (Advanced Bionics, Valencia, CA). All had full insertion of the implant electrode array and no patient experienced intraoperative complications.
Postoperative course was uneventful in patient 2, but complicated by wound healing in patients 1 and 3. Fourteen days after CI, patient 1 demonstrated minor dehiscence in blind sac closure and was treated with topical (bacitracin) and oral (ciprofloxacin) antibiotics for 7 days with complete resolution. Patient 3 developed a small postauricular wound dehiscence with mastoid-cutaneous fistula requiring closure using local anesthesia. Intraoperatively, the fistula tract was excised and closed with a cartilage and perichondrial rotation flap without disruption of the implant. Complete healing was documented after 4 weeks.
Postimplant performance data demonstrates improvement in speech recognition in all 3 patients (Table II). Length of device usage ranges from 6 months (patients 2 and 3) to 2 years (patient 1). For patient 1, CNC-W and CNC-P improved from 8% to 72% and 24% to 87%, respectively. Patient 3 demonstrated similar benefit with improvement from 0% to 36% and 0% to 64%, respectively. Patient 2, unable to complete preoperative CNC or HINT testing, identified 94% and 92% of words correctly on HINT-Q and HINT-N, respectively.
Treatment of pediatric medulloblastoma, the most common malignant tumor of the CNS in children, requires a combination of surgical excision, radiation therapy, and chemotherapy. Prior research has documented ototoxicity and middle ear disease in 25% to 90% of patients treated with multimodality therapy.5, 7, 11 At this institution, pediatric medulloblastoma patients receive a cumulative cisplatin dose of 300 mg, decreased from the standard of 600 mg. Early analysis suggests that this reduction may minimize ototoxic sequelae, however long-term data on SNHL is lacking. Further, the percentage of patients in this population that will ultimately qualify for cochlear implantation has not yet been documented.
Damage to the auditory system from chemotherapeutic agents, specifically platinum-based compounds, as well as radiation therapy is well documented. Cisplatinum ototoxicity is typically characterized by irreversible, bilateral, symmetric, high-frequency SNHL.5, 7 Ototoxicity can be related to cumulative dose as well as dose intensity, reinforcing the need for audiologic monitoring in these patients. Histopathologic studies of temporal bones treated with cisplatinum demonstrate damage to the inner and outer hair cells, atrophy of the stria vascularis, as well as some degeneration of the spiral ganglion cells and cochlear nerve.7
Similar histologic findings are seen in temporal bones subjected to radiation therapy. Acute effects, such as serous effusion, mucosal inflammation, and intimal hypertrophy of arterioles can lead to serous effusion and temporary conductive hearing loss.7 In contrast, SNHL is typically progressive, irreversible, and dose dependent, and can manifest long after completion of radiation.8, 14–16 Although the mechanism of sensorineural damage is not fully understood, ischemic effects on cochlear vessels leading to hair cell and supporting cell death and intracochlear fibrosis have all been implicated.7, 9, 11, 17 Long-term vascular changes are seen throughout the temporal bone and involve extensive soft tissue fibrosis, atherosclerosis, obliterative endarteritis, and thrombosis.7, 11 Bony histologic changes, including loss of osteocytes, and lack of new bone formation can predispose to localized infection, such as otitis externa, and can also lead to osteoradionecrosis.7, 9
Importantly, effects of radiation on the retrocochlear auditory pathways remain incompletely understood and functionality of this pathway is a prerequisite for successful auditory rehabilitation. Although evidence supports a cochlear pathology for SNHL following radiation therapy, brainstem and intracranial auditory pathways are subjected to radiation in medulloblastoma patients. One study by Low et al. examined the retrocochlear pathways of nasopharyngeal cancer patients treated with radiation therapy and suggested that these auditory pathways remained intact at 2 years post-treatment.18 A follow-up study documented successful cochlear implantation in four postradiated nasopharyngeal cancer patients with postimplant performance scores similar to nonirradiated controls.19 In addition, the postimplant performance of the patients in this study support functionality of the retrocochlear auditory pathway after radiation.
In the long-term, extensive potential sequelae of chemoradiation necessitate rigorous and vigilant follow-up. In addition to audiologic monitoring, patients subjected to temporal bone radiation require both concurrent and long-term otologic exams to identify and treat sequelae. Chronic ear disease following radiation therapy is a prevalent side effect and was seen in each of the three patients discussed in this article.19 Careful assessment and management of middle ear and mastoid manifestations, including aggressive treatment of infection, is essential preoperative preparation for implantation. Further, operative treatment of chronic ear disease should incorporate the possibility of future cochlear implantation and may include subtotal petrosectomy, obliteration, and the creation of a blind sac of the external auditory canal.
Next, multiple intraoperative considerations are crucial to successful implantation in these patients. Previously mentioned histologic bony changes lead to softer bone than is typically encountered in nonradiated patients. Caution around important structures, such as the otic capsule and facial nerve, should be exercised to avoid unintentional damage.9, 19 Middle ear mucosal disease, including adhesions, may complicate identification of the round window niche.19 Importantly, postradiation obliteration of the cochlear lumen has been documented and may compromise insertion of the electrode.7, 9, 19 In this series, each patient had full insertion of the electrode array.
Poor wound healing and soft tissue changes following chemoradiation can have greater impact in the setting of foreign body implantation and concern for extrusion. Careful planning of skin incisions and flaps is essential, and areas of atrophic erythematous skin should be avoided if possible. Meticulous handling of skin flaps and soft tissue throughout the procedure, avoidance of flap thinning, and careful use of electrocautery may minimize post operative complications, such as device extrusion.9, 19 In this study, two patients demonstrated poor wound healing with one requiring operative closure. Keys to effective postoperative management include close follow-up, vigilant examination of the postauricular incision and magnet, as well as immediate aggressive treatment of wound complications if discovered.
Additional perioperative considerations related to general anesthesia include various endocrinopathies common in this population, notably those effecting the hypothalamic-pituitary axis. Multidisciplinary preoperative evaluation and appropriate perioperative hormonal replacement is critical in avoiding potential anesthetic complications.
Finally, issues of tumor surveillance, recurrence, and future radiation therapy must be addressed. MRI is typically used in evaluation of tumor recurrence in medulloblastoma patients up to 5 years post-treatment. Implantation prior to 5 years should be balanced by the need for effective tumor surveillance. In cases where MRI is anticipated or becomes necessary, options include implantation with a non–magnet-containing device or removal of the magnet after implantation.
In the event of tumor recurrence or secondary malignancy requiring surgical biopsy or excision, only bipolar electrocautery could be used for intraoperative hemostasis. Should radiotherapy be required, little data exists regarding the safety of radiation therapy in cochlear patients. One study by Klenzner et al., however, suggests that a large single dose of radiation poses a low risk of implant failure.20 In terms of radiation applicability and efficacy, issues of shadow and dose concentration from a metallic implanted device must be considered in treatment planning.
Current research into reduction of auditory side effects from chemotherapy and therapeutic radiation is promising. Recent studies examining cochlear protective medication suggest that amifostine may significantly reduce the ototoxicity in medulloblastoma patients treated with cisplatinum. With respect to radiation, Huang et al. found that patients treated with intensity-modulated radiation therapy instead of conventional radiation therapy received 68% of radiation to the auditory apparatus and experienced significantly lower levels of ototoxicity.21 Additional research on auditory and otologic side effects, otoprotective measures, and post-treatment of cochlear implantation in this population is warranted.