Classification of Neurological Abnormalities in Children With Congenital Melanocytic Naevus Syndrome Identifies Magnetic Resonance Imaging as the Best Predictor of Clinical Outcome

R. Waelchli; S.E. Aylett; D. Atherton; D.J. Thompson; W.K. Chong; V.A. Kinsler


The British Journal of Dermatology. 2015;173(3):739-750. 

In This Article

Patients and Methods


Patients in this cohort were seen sequentially in the paediatric dermatology department at Great Ormond Street Hospital between 1991 and 2013, and followed prospectively. Therefore, this study includes the majority of patients studied in two previous publications by our group,[3,20] and, owing to the larger patient cohort and systematic sequential collection of patients in recent years, supplants the previous studies from the point of view of recommendations. All patients with an abnormal MRI scan were also evaluated by a paediatric neurologist and followed prospectively, which included a detailed neurodevelopmental and neurological history and examination. In total, 636 new patients were seen (see Fig. 1), of whom 376 were deemed after a review of the notes to have had sufficient clinical data collected at the first visit for them to be included in this analysis. The total of 376 includes all new patients (n = 291) seen sequentially since 2006, after a standardized clinical data collection form was introduced. In total, 289 of the 376 patients fulfilled our criteria for MRI of the CNS, and 271 had successfully completed scans at the time of analysis. The MRI scans were undertaken almost exclusively under sedation only, rather than under general anaesthesia, as is standard practice in our hospital for children < 1 year of age. Where scans were not completed this was due to either failure of sedation of the patient or failure to attend for the appointment on two occasions. Images were analysed by a paediatric neuroradiologist with expertise in CMN, who was aware of the diagnosis of multiple CMN but who did not have details on clinical outcomes.

Figure 1.

Flow chart of numbers and timings of patients seen and included in this study. MRI, magnetic resonance imaging.

Inclusion Criteria

Before 2008 the criteria for MRI were a CMN > 2 cm overlying the spine or brain, or a CMN elsewhere of at least the size of the patient's hand. After results published in 2008,[21] the criteria were changed to include only those with multiple CMN (i.e. more than one at birth) independent of the site and size of the lesions, if presenting before the age of 2 years. If the patient is older than 2 years at first presentation and neurologically normal we do not perform a screening MRI as most congenital lesions requiring treatment should have declared themselves by this stage. Urgent MRI is performed at any age with any CMN if neurological symptoms occur. Only four children were neurologically symptomatic at the time of the referral to our department.


The phenotypic variables analysed were radiological (MRI result and subsequent classification) and clinical (largest CMN projected adult size, total number of naevi at birth and at enrolment). Adverse outcome measures were seizures, neurodevelopmental problems, whether neurosurgery was required, primary CNS malignant melanoma, and death from primary CNS melanoma or primary cutaneous melanoma. Neurodevelopmental problems were defined as those diagnosed by a paediatrician or a paediatric neurologist, or officially by school assessment and requiring an educational intervention.

Nonfatal melanoma was not included as an outcome measure as there were no cases in this cohort where melanoma was not fatal. Statistical modelling of outcome measures was performed by multiple logistic regression, considering the whole cohort together. For statistical analysis of the effect of guidelines introduced in 2008, the frequency of scans ordered and the percentage of abnormal scans was compared in the pre- and post-2008 cohorts by a two-tailed Fisher's exact test.