Cerebellar mutism syndrome: the importance of preoperative language assessment.

Children undergoing surgical removal of tumors in the posterior cranial fossa can encounter a varied and complex constellation of neurological symptoms, called cerebellar mutism, defined as a disturbance in the planning and programming of motor language with preserved understanding, behavioral disorders such as inattention, visual-spatial disorganization, personality change, as well as ataxia and dysmetria. In the last years, several groups have been trying to establish risk factors or even predictive scores in order to be able at least in part to predict the appearance of speech disorders before surgery. We report on a child with pilocytic astrocytoma of the cerebellar vermis who had already been diagnosed with developmental linguistic delay two years earlier. This disorder initially worsened after surgery and later improved in the following 12 months. The aim of this paper is to emphasize the importance of preoperative neuropsychological evaluation. The present case, along with those reported in the literature, suggests that the risk of long-term cerebellar mutism is higher in children with preoperative speech disorders. In these patients a thorough assessment of cognitive and linguistic functions is therefore necessary to better evaluate the risk of cerebellar mutism after surgery.

Deep learning-based automatic tumor burden assessment of pediatric high-grade gliomas, medulloblastomas, and other leptomeningeal seeding tumors.

BACKGROUND: Longitudinal measurement of tumor burden with magnetic resonance imaging (MRI) is an essential component of response assessment in pediatric brain tumors. We developed a fully automated pipeline for the segmentation of tumors in pediatric high-grade gliomas, medulloblastomas, and leptomeningeal seeding tumors. We further developed an algorithm for automatic 2D and volumetric size measurement of tumors. METHODS: The preoperative and postoperative cohorts were randomly split into training and testing sets in a 4:1 ratio. A 3D U-Net neural network was trained to automatically segment the tumor on T1 contrast-enhanced and T2/FLAIR images. The product of the maximum bidimensional diameters according to the RAPNO (Response Assessment in Pediatric Neuro-Oncology) criteria (AutoRAPNO) was determined. Performance was compared to that of 2 expert human raters who performed assessments independently. Volumetric measurements of predicted and expert segmentations were computationally derived and compared. RESULTS: A total of 794 preoperative MRIs from 794 patients and 1003 postoperative MRIs from 122 patients were included. There was excellent agreement of volumes between preoperative and postoperative predicted and manual segmentations, with intraclass correlation coefficients (ICCs) of 0.912 and 0.960 for the 2 preoperative and 0.947 and 0.896 for the 2 postoperative models. There was high agreement between AutoRAPNO scores on predicted segmentations and manually calculated scores based on manual segmentations (Rater 2 ICC = 0.909; Rater 3 ICC = 0.851). Lastly, the performance of AutoRAPNO was superior in repeatability to that of human raters for MRIs with multiple lesions. CONCLUSIONS: Our automated deep learning pipeline demonstrates potential utility for response assessment in pediatric brain tumors. The tool should be further validated in prospective studies.

The clinical and financial impact of a pediatric surgical neuro-oncology clinical trial.

Pediatric surgical trials are rare and the impact of such trials on the institutions in which they are conducted is unknown. The purpose of this study was to analyze the clinical and financial impact of The Re-MATCH trial, a Phase I clinical trial requiring the biopsy or resection of recurrent medulloblastoma or PNET for enrollment. Inpatient financial and clinical volume information was collected during the 3 years of trial enrollment and the years preceding and following it. The primary endpoints were the difference in direct contribution margin (DCM), or net gain, of study and non-study patients and the difference in surgical volume during the study and non-study periods. The trial enrolled 18 patients; 15 had surgery at the sponsor institution and three had surgery at their home institution, then transferred tumor material to the sponsor institution. There were no differences between the two groups for potentially confounding variables such as neurosurgical procedure work relative value units (P = 0.13) or insurance provider (P = 0.26). There was no difference between the inpatient DCM per case for the institution for non-study patients (mean ± SD, $9039 ± $28,549) and study patients ($14,332 ± $20,231) (P = 0.4819). During the non-study period, there were a mean of 2.78 ± 1.65 pediatric brain tumor resections per month compared to 3.34 ± 1.66 cases per month during the study period, a 17% increase. When the 15 study patients were excluded, there were 2.97 ± 1.64 cases per month, a 7% increase. However, this increase in total case volume including study and non-study patients was not significant (P = 0.121). Phase I investigator-initiated surgically-based clinical trials may increase institutional surgical volume without imposing a financial burden. Finances are unlikely to be a barrier for researchers negotiating for resources to conduct such trials.

Mapping radiation dose distribution on the fractional anisotropy map: applications in the assessment of treatment-induced white matter injury.

We describe a method to map whole brain radiation dose distribution on to diffusion tensor MR (DT-MR) fractional anisotropy (FA) images and illustrate its applications for studying dose-effect relationships and regional susceptibility in two childhood medulloblastoma survivors. To determine the FA changes voxel-by-voxel in white matter, the post-treatment follow-up FA maps were coregistered to baseline pre-treatment FA maps and automatic segmentation for white matter was carried out. DeltaFA maps representing relative FA change in white matter were hence generated for visual inspection and quantitative analysis. The radiation dose distribution, calculated from radiotherapy plan and exported as images, was coregistered to baseline FA images. DT-MR imaging and processing noise was small with root mean square value of 1.49% for mean DeltaFA. We evaluated the mean DeltaFA changes of regions-of-interest according to radiation dose regions to provide an estimate of the dose-response and found increasing reduction in mean DeltaFA with increasing radiation dose up to 45 Gy after which there was a reversal in the mean FA trend and mean FA approached baseline value. We also found more severe mean FA reduction in the frontal lobes compared to the parietal lobes despite the same radiation dose, suggesting regional susceptibility in the frontal lobe, and mean FA increase in the brainstem after radiation in both patients. We conclude that the method described may be useful in estimating dose-effect relationships and studying regional susceptibility of the brain to radiation in medulloblastoma survivors.

Myelography in the assessment of children with medulloblastoma.

Forty-seven out of 49 patients with a histologically-proven diagnosis of medulloblastoma admitted to our institution between 1 January 1984 and 13 December 1990 were examined post-operatively using either myelography or gadolinium-enhanced magnetic resonance imaging in an attempt to detect clinically occult spinal metastatic disease. Spinal spread of the tumour was identified in nine (19%) patients. Of the 16 patients for whom a 5-year follow-up has been available 12 had no myelographic evidence of spinal disease and of these 10 (83%) remain disease-free. Of the six who have died during this period four had evidence of spinal metastases at presentation. Of all the children in the series with spinal spread identified at the time of their initial hospitalization, only one survives to date (15 months after diagnosis). The remainder had an average life expectancy of 18 months.

Brain tumours under the age of three. The price of survival. A retrospective study of 20 long-term survivors.

Between 1975 and 1989, 98 children with brain tumours under the age of three at time of diagnosis were entered into a retrospective study. Twenty of them are alive and free of tumour more than five years after treatment and were evaluated in this study. Thirteen tumour localizations were infratentorial and 7 were supratentorial. A histological examination was performed in 15 patients: 5 ependymomas, 6 medulloblastomas and 4 astrocytomas were identified. Fifteen patients underwent surgical removal of tumour, all but one received radiotherapy and 8 were given chemotherapy. Only two children have not late effects. Analysis of long-term sequelae in survivors showed central endocrinopathies in 14 (70%), a neurological handicap in 13 (65%) and impaired cognitive functions in 17 (85%). Irradiation was clearly responsible for mental sequelae in 7 patients and endocrinopathies in 6 patients. The other possible causes are tumour injury, hydrocephalus or surgery. The risks incurred with radiotherapy and advances in infant brain tumour therapy are discussed.