Cerebellar mutism syndrome in children with brain tumours of the posterior fossa

Central nervous system (CNS) tumours account for 25% of all cancers in children and over half of these are located in the posterior fossa [1]. For most of these patients, treatment includes surgery. Posterior fossa tumours in children are associated with high risk of chronic neurological and neurocognitive disability [2‐6]. The cerebellar mutism syndrome (CMS) refers to the constellation of transient mutism, ataxia, hypotonia and irritability following surgery for cerebellar or fourth ventricle tumours in children and adolescents [7]. Although some patients may recover completely, recovery may be prolonged, and many are left with permanent disabling sequelae in the form of e.g. dysarthria, dysfluency, slowed speech rate and ataxia. Many may in addition be burdened by emotional problems and lower IQ [8‐12].

The spectrum of CMS definitions varies greatly [7, 12‐14], leading to differences in reported incidence and uncertainties about recovery. Incidence figures thus range from 8% [15, 16] to 32% [17] in children with any kind of cerebellar tumour when a variety of definitions are used, compared to 24% [12] to 39% [18] in patients with medulloblastomas using a more precise CMS definition. A recent study of 148 children with cerebellar tumours found that the overall incidence of the broader Posterior Fossa Syndrome was 28%, subdivided by tumour pathology into 40% for medulloblastoma, 16% for astrocytoma and 20% for ependymoma [19]. The CMS definition created by the Neurology Committee of the Children’s Cancer Group in USA in 1993 is currently the only one associated with a specific scoring scale [12] and is used in this project.

Cerebellar mutism is thought to be caused by bilateral disturbance of the dentate nuclei and/or their efferents [16, 20‐24]. Ataxia and irritability together with other cognitive, affective and motor symptoms that are frequently observed in CMS patients are caused by damage to various parts of the cerebellum and cerebello-cerebral pathways passing through the brainstem [25‐28]. This can result in secondary diaschisis of supratentorial brain areas due to lack of excitatory input from cerebellum [29‐31].

Known risk factors are brainstem involvement by the tumour, midline location and tumour type; thus the incidence in children with medulloblastoma is two to three times higher than for astrocytoma or ependymoma but the biological mechanisms behind these associations are uncertain [12, 19, 22, 32‐35]. Recently proposed risk factors include brainstem compression by the tumour, pre-operative language impairment, low socioeconomic level of the families and left-handedness [24, 36‐38]; however, these remain to be verified on large patient cohorts. Tumour size, neurosurgical techniques and approaches, radical resection and younger age at diagnosis are uncertain risk factors, as previous studies have been inconclusive [12, 23, 32‐35, 39, 40]. Gender, hydrocephalus, post-operative central nervous system infections, type of neurosurgeon (adult/paediatric), and oedema/swelling of the cerebellum have not been significantly correlated to CMS and are considered unlikely risk factors [12, 16, 34, 35, 41, 42].

In traumatic brain injury common host genomic variants are related to the severity of symptoms and degree of recovery [43‐45]. Similar associations are likely for surgical brain injury, but such studies have not been performed for CMS.

Supportive speech and rehabilitation therapy is often offered to patients with CMS, but the benefit hereof has not been demonstrated. No publications exist on systematic approaches to pharmacological neuroprotection, and pharmacological interventions are only sporadically reported in the literature [46‐50]. Glucocorticosteroids are routinely given to most patients pre-, intra- and post-operatively to reduce inflammation and edema [51] [52], but there is no consensus recommendation and the impact on the clinical course of CMS is undetermined.

The study focuses on the risk factors for development and severity of CMS including surgery (approaches, techniques and tissue and vascular damage, re-operation) and host genome variants. The aims of this study are thus to describe differences in incidence, severity and clinical course of CMS related to:

This open observational study is registered at Clinicaltrials.​gov (file NCT02300766) and EANS (http://​www.​sbns.​org.​uk/​index.​php/​research/​eu-multi-centre-trials). All children younger than 18 years with a tumour in the posterior fossa requiring surgery or open biopsy at one of the participating centres will be included following informed consent. Patients who have received surgery, chemotherapy and/or radiotherapy previously are also eligible. The study will run for five years with a targeted sample size of 500 patients. It opened late 2014 with participation from 20 Nordic and Baltic centres. From 2016, five British centres and four Dutch centres will join leading to an expected annual accrual of 130 patients. Three other major European centres are invited to join from 2016/17. The target of 500 patients is expected to be reached in 2018. Patients will be followed for 12 months after inclusion of the last patient, and the study will thus be completed during 2019.The participating centres provide surgery and supportive care according to local practice and register all study information in an online database developed specifically for this study. Consensus concerning the study aims, study design and data registration was achieved at three international planning meetings among the initiating centres during 2013. The annual enrolment from each country will be compared to the number of registered patients in the national cancer registries to document the inclusion rate and representativeness.

The primary endpoint is incidence and severity of CMS. Symptoms and severity are scored according to the CMS survey published by Robertsons et al. [12]. Our main focus is the impact of different surgical tumour approaches. We hypothesize that 1) minimally traumatic techniques and 2) sparing the dentate nuclei and their efferents will be associated with a 50% reduced risk of CMS when compared to more invasive tumour removal approaches. Furthermore, we hypothesize that the risk of developing CMS is higher after re-operation(s) compared to primary surgery.

The secondary endpoint is incidence of “reduced speech output” defined as “severely reduced speech production limited to single words or short sentences which can only be elicited after vigorous stimulation” [19]. The risk of reduced speech output will be related to different surgical approaches with the underlying hypothesis that damage to the dentate nuclei and/or their efferents increases the risk.

Furthermore, we want to explore the following:

The following data will be registered at five time points by online standard registration forms:

In case of emergency surgery (e.g. due to risk of incarceration or coma) information about the study and invitation to participate can be given within seven days postoperatively. These patients will be included in all parts of the study except for the recording and analysis of preoperative speech and language status.

In cases with repeat tumour surgery during the twelve months follow-up, the patient can re-enter the study and start a new follow-up programme (Fig. 1, Repeated Surgeries). A new pre-operative registration is then performed corresponding to the re-operation. Post-operative registrations will be performed again, and used in the analysis of risk related to first versus further surgeries. If surgery is performed again after the twelve months follow-up period, the patient will be re-invited to participate in the study.

In accordance with our surgical hypothesis of 50% risk reduction by less traumatic techniques, and assuming that 35% of the patients are operated using an approach with a low risk of CMS (assumed to be 10%) and the remaining 65% of patients are operated using other approaches (assumed carrying a 20% risk), a total of 450 patients have to be included to identify a 5% significance level and 80% power. Based on a projected overall risk of CMS of 20%, an estimated frequency of a specific SNP of 30%, and a projected doubled risk of CMS with this particular SNP, we will need to include a total of 343 patients to identify such a genetic predisposition at a 5% significance level with 90% power.