Third ventriculostomy effectively controls obstructive triventricular hydrocephalus in more than 70% of children and should be preferred to placement of extracranial cerebrospinal shunts in this group of patients. When performed using ventriculographic guidance, the technique has a higher mortality rate and a higher failure rate in children younger than 6 months of age and is, therefore, no longer preferred. When third ventriculostomy is performed using endoscopic guidance, the same long-term results are achieved in children younger than 6 months of age as in older children and, thus, patient age should no longer be considered as a contraindication to using the technique. Delayed failures are usually secondary to obstruction of the stoma and often can be managed by repeating the procedure. Midline sagittal T2-weighted MR imaging sequences combined with cine PC MR imaging flow measurements provide a reliable tool for diagnosis of aqueductal stenosis and for ascertaining the patency of the stoma during follow-up evaluation.

BackgroundThe use of chemotherapy to manage newly diagnosed low grade glioma (LGG) was first introduced in the 1980s. One randomised trial has studied two- versus four-drug regimens with a duration of 12 months of treatment after resection.MethodsWithin the European comprehensive treatment strategy for childhood LGG, the International Society of Paediatric Oncology–Low Grade Glioma (SIOP LGG) Committee launched a randomised trial involving 118 institutions and 11 countries to investigate the addition of etoposide (100 mg/m2, days 1, 2 & 3) to a four-course induction of vincristine (1.5 mg/m2 × 10 wkly) and carboplatin (550 mg/m2 q 3 weekly) as part of 18-month continuing treatment programme. Patients were recruited after imaging diagnosis, resection or biopsy with progressive disease/symptoms. Some 497 newly diagnosed patients (M/F 231/266; median age 4.26 years (interquartile range (IQR) 2.02–7.06)) were randomised to receive vincristine carboplatin (VC) (n = 249) or VC plus etoposide (VCE) during induction (n = 248), stratified by age and tumour site.FindingsNo differences between the two arms were found in term of survival and radiological response. Response and non-progression rates at 24 weeks for VC and VCE, were 46% versus 41%, and 93% versus 91% respectively; 5-year Progression-Free Survival (PFS) and Overall Survival (OS) were 46% (StDev 3.5) versus 45% (StDev 3.5) and 89% (StDev 2.1) versus 89% (StDev 2.1) respectively. Age and diencephalic syndrome are adverse clinical risk factors for PFS and OS. 5-year OS for patients in early progression at week 24 were 46% (StDev 13.8) and 49% (StDev 16.5) in the two arms, respectively.InterpretationThe addition of etoposide to VC did not improve PFS or OS. High non-progression rates at 24 weeks justify retaining VC as standard first-line therapy. Infants with diencephalic syndrome and early progression need new treatments to be tested. Future trials should use neurological/visual and toxicity outcomes and be designed to discriminate between the impact on disease outcomes of ‘duration of therapy’ and ‘age at stopping therapy’.

Third ventriculostomy is feasible even in the presence of posterior fossa tumors (including brainstem tumors). When performed prior to posterior fossa surgery, it significantly reduces the incidence of postoperative hydrocephalus. The procedure provides a valid alternative to placement of a permanent shunt in cases in which hydrocephalus develops following posterior fossa surgery, and it may negate the need for the shunt in cases in which the shunt malfunctions. Furthermore, in patients in whom CSF has caused spread of the tumor at presentation, third ventriculostomy allows chemotherapy to be undertaken prior to tumor excision by controlling hydrocephalus. Although the authors acknowledge that the routine application of third ventriculostomy in selected patients results in a proportion of patients undergoing an "unnecessary" procedure, they believe that because patients' postoperative courses are less complicated and because the incidence of morbidity is low and the success rate is high in those patients with severe hydrocephalus that further investigation of this protocol is warranted.

Abnormal intracranial CSF hydrodynamics was found in 8.1% of the patients (3.4% of whom had received shunts and 4.5% of whom had not). Three types of CSF hydrodynamic disturbance were observed: progressive hydrocephalus with ventricular dilation, nonprogressive ventriculomegaly, and dilation of the subarachnoid spaces. Hydrocephalus occurred much more frequently in patients with syndromic craniosynostosis (12.1%) than in those with isolated craniosynostosis (0.3%). In fact, patients with kleeblattschädel exhibited hydrocephalus as a constant feature and patients with Crouzon's syndrome were far more likely to have hydrocephalus than those with other syndromes. In Apert's syndrome, ventricular dilation occurred very frequently, but it was almost always nonprogressive in nature. In most cases of syndromic craniosynostosis, venous sinus obstruction and/or chronic tonsillar herniation were found. Their role in the pathophysiology of hydrocephalus in craniosynostosis is discussed.

Although the development of tonsillar herniation (acquired Chiari malformation) in association with lumboperitoneal (LP) shunting is well recognized, it has previously been considered rare. In order to ascertain the incidence of this complication after LP shunting, the authors undertook a retrospective study of all patients in whom this form of shunt had been inserted between 1974 and 1991 at The Hospital for Sick Children, Toronto. In the 143 patients, the mean age at insertion was 3.3 years and the indications for shunt placement were hydrocephalus (81%), pseudotumor cerebri (7%), cerebrospinal fluid fistula (6%), and posterior fossa pseudomeningocele (6%). The mean follow-up period was 5.7 years, during which time there was one shunt-related death due to unsuspected tonsillar herniation. Five other patients developed symptomatic tonsillar herniation treated by suboccipital decompression. Review of all computerized tomography (CT) scans not degraded by artifact showed evidence of excess soft tissue at the level of the foramen magnum in 38 (70%) of 54 patients so studied. In order to confirm that this CT finding represented hindbrain herniation, sagittal and axial magnetic resonance (MR) images were obtained for 17 asymptomatic patients and revealed tonsillar herniation (range 2 to 21 mm) in 12 (70.6%). In addition, some of these asymptomatic patients had evidence of uncal herniation and mesencephalic distortion. Similarities and distinctions are drawn between the morphological changes occurring after LP shunting and those seen in association with the Chiari I and II malformations. Although less than 5% of this study population required treatment for tonsillar herniation, the incidence of this complication was high in asymptomatic patients; MR imaging surveillance for patients with LP shunts is therefore recommended.