Metabolite selection for machine learning in childhood brain tumour classification.

MRS can provide high accuracy in the diagnosis of childhood brain tumours when combined with machine learning. A feature selection method such as principal component analysis is commonly used to reduce the dimensionality of metabolite profiles prior to classification. However, an alternative approach of identifying the optimal set of metabolites has not been fully evaluated, possibly due to the challenges of defining this for a multi-class problem. This study aims to investigate metabolite selection from in vivo MRS for childhood brain tumour classification. Multi-site 1.5 T and 3 T cohorts of patients with a brain tumour and histological diagnosis of ependymoma, medulloblastoma and pilocytic astrocytoma were retrospectively evaluated. Dimensionality reduction was undertaken by selecting metabolite concentrations through multi-class receiver operating characteristics and compared with principal component analysis. Classification accuracy was determined through leave-one-out and k-fold cross-validation. Metabolites identified as crucial in tumour classification include myo-inositol (P < 0.05, AUC=0.81±0.01 ), total lipids and macromolecules at 0.9 ppm (P < 0.05, AUC=0.78±0.01 ) and total creatine (P < 0.05, AUC=0.77±0.01 ) for the 1.5 T cohort, and glycine (P < 0.05, AUC=0.79±0.01 ), total N-acetylaspartate (P < 0.05, AUC=0.79±0.01 ) and total choline (P < 0.05, AUC=0.75±0.01 ) for the 3 T cohort. Compared with the principal components, the selected metabolites were able to provide significantly improved discrimination between the tumours through most classifiers (P < 0.05). The highest balanced classification accuracy determined through leave-one-out cross-validation was 85% for 1.5 T 1 H-MRS through support vector machine and 75% for 3 T 1 H-MRS through linear discriminant analysis after oversampling the minority. The study suggests that a group of crucial metabolites helps to achieve better discrimination between childhood brain tumours.

Added value of magnetic resonance spectroscopy for diagnosing childhood cerebellar tumours.

1 H-magnetic resonance spectroscopy (MRS) provides noninvasive metabolite profiles with the potential to aid the diagnosis of brain tumours. Prospective studies of diagnostic accuracy and comparisons with conventional MRI are lacking. The aim of the current study was to evaluate, prospectively, the diagnostic accuracy of a previously established classifier for diagnosing the three major childhood cerebellar tumours, and to determine added value compared with standard reporting of conventional imaging. Single-voxel MRS (1.5 T, PRESS, TE 30 ms, TR 1500 ms, spectral resolution 1 Hz/point) was acquired prospectively on 39 consecutive cerebellar tumours with histopathological diagnoses of pilocytic astrocytoma, ependymoma or medulloblastoma. Spectra were analysed with LCModel and predefined quality control criteria were applied, leaving 33 cases in the analysis. The MRS diagnostic classifier was applied to this dataset. A retrospective analysis was subsequently undertaken by three radiologists, blind to histopathological diagnosis, to determine the change in diagnostic certainty when sequentially viewing conventional imaging, MRS and a decision support tool, based on the classifier. The overall classifier accuracy, evaluated prospectively, was 91%. Incorrectly classified cases, two anaplastic ependymomas, and a rare histological variant of medulloblastoma, were not well represented in the original training set. On retrospective review of conventional MRI, MRS and the classifier result, all radiologists showed a significant increase (Wilcoxon signed rank test, p < 0.001) in their certainty of the correct diagnosis, between viewing the conventional imaging and MRS with the decision support system. It was concluded that MRS can aid the noninvasive diagnosis of posterior fossa tumours in children, and that a decision support classifier helps in MRS interpretation.

Magnetic resonance spectroscopy assessment of brain injury after moderate hypothermia in neonatal encephalopathy: a prospective multicentre cohort study.

BACKGROUND: In neonatal encephalopathy, the clinical manifestations of injury can only be reliably assessed several years after an intervention, complicating early prognostication and rendering trials of promising neuroprotectants slow and expensive. We aimed to determine the accuracy of thalamic proton magnetic resonance (MR) spectroscopy (MRS) biomarkers as early predictors of the neurodevelopmental abnormalities observed years after neonatal encephalopathy. METHODS: We did a prospective multicentre cohort study across eight neonatal intensive care units in the UK and USA, recruiting term and near-term neonates who received therapeutic hypothermia for neonatal encephalopathy. We excluded infants with life-threatening congenital malformations, syndromic disorders, neurometabolic diseases, or any alternative diagnoses for encephalopathy that were apparent within 6 h of birth. We obtained T1-weighted, T2-weighted, and diffusion-weighted MRI and thalamic proton MRS 4-14 days after birth. Clinical neurodevelopmental tests were done 18-24 months later. The primary outcome was the association between MR biomarkers and an adverse neurodevelopmental outcome, defined as death or moderate or severe disability, measured using a multivariable prognostic model. We used receiver operating characteristic (ROC) curves to examine the prognostic accuracy of the individual biomarkers. This trial is registered with ClinicalTrials.gov, number NCT01309711. FINDINGS: Between Jan 29, 2013, and June 25, 2016, we recruited 223 infants who all underwent MRI and MRS at a median age of 7 days (IQR 5-10), with 190 (85%) followed up for neurological examination at a median age of 23 months (20-25). Of those followed up, 31 (16%) had moderate or severe disability, including one death. Multiple logistic regression analysis could not be done because thalamic N-acetylaspartate (NAA) concentration alone accurately predicted an adverse neurodevelopmental outcome (area under the curve [AUC] of 0·99 [95% CI 0·94-1·00]; sensitivity 100% [74-100]; specificity 97% [90-100]; n=82); the models would not converge when any additional variable was examined. The AUC (95% CI) of clinical examination at 6 h (n=190) and at discharge (n=167) were 0·72 (0·65-0·78) and 0·60 (0·53-0·68), respectively, and the AUC of abnormal amplitude integrated EEG at 6 h (n=169) was 0·73 (0·65-0·79). On conventional MRI (n=190), cortical injury had an AUC of 0·67 (0·60-0·73), basal ganglia or thalamic injury had an AUC of 0·81 (0·75-0·87), and abnormal signal in the posterior limb of internal capsule (PLIC) had an AUC of 0·82 (0·76-0·87). Fractional anisotropy of PLIC (n=65) had an AUC of 0·82 (0·76-0·87). MRS metabolite peak-area ratios (n=160) of NAA-creatine (<1·29) had an AUC of 0·79 (0·72-0·85), of NAA-choline had an AUC of 0·74 (0·66-0·80), and of lactate-NAA (>0·22) had an AUC of 0·94 (0·89-0·97). INTERPRETATION: Thalamic proton MRS measures acquired soon after birth in neonatal encephalopathy had the highest accuracy to predict neurdevelopment 2 years later. These methods could be applied to increase the power of neuroprotection trials while reducing their duration. FUNDING: National Institute for Health Research UK.

Application of pattern recognition techniques for classification of pediatric brain tumors by in vivo 3T 1 H-MR spectroscopy-A multi-center study.

PURPOSE: 3T magnetic resonance scanners have boosted clinical application of 1 H-MR spectroscopy (MRS) by offering an improved signal-to-noise ratio and increased spectral resolution, thereby identifying more metabolites and extending the range of metabolic information. Spectroscopic data from clinical 1.5T MR scanners has been shown to discriminate between pediatric brain tumors by applying machine learning techniques to further aid diagnosis. The purpose of this multi-center study was to investigate the discriminative potential of metabolite profiles obtained from 3T scanners in classifying pediatric brain tumors. METHODS: A total of 41 pediatric patients with brain tumors (17 medulloblastomas, 20 pilocytic astrocytomas, and 4 ependymomas) were scanned across four different hospitals. Raw spectroscopy data were processed using TARQUIN. Borderline synthetic minority oversampling technique was used to correct for the data skewness. Different classifiers were trained using linear discriminative analysis, support vector machine, and random forest techniques. RESULTS: Support vector machine had the highest balanced accuracy for discriminating the three tumor types. The balanced accuracy achieved was higher than the balanced accuracy previously reported for similar multi-center dataset from 1.5T magnets with echo time 20 to 32 ms alone. CONCLUSION: This study showed that 3T MRS can detect key differences in metabolite profiles for the main types of childhood tumors. Magn Reson Med 79:2359-2366, 2018. © 2017 International Society for Magnetic Resonance in Medicine.

Magnetic Resonance Biomarkers in Neonatal Encephalopathy (MARBLE): a prospective multicountry study.

INTRODUCTION: Despite cooling, adverse outcomes are seen in up to half of the surviving infants after neonatal encephalopathy. A number of novel adjunct drug therapies with cooling have been shown to be highly neuroprotective in animal studies, and are currently awaiting clinical translation. Rigorous evaluation of these therapies in phase II trials using surrogate MR biomarkers may speed up their bench to bedside translation. A recent systematic review of single-centre studies has suggested that MR spectroscopy biomarkers offer the best promise; however, the prognostic accuracy of these biomarkers in cooled encephalopathic babies in a multicentre setting using different MR scan makers is not known. METHODS AND ANALYSIS: The MR scanners (3 T; Philips, Siemens, GE) in all the participating sites will be harmonised using phantom experiments and healthy adult volunteers before the start of the study. We will then recruit 180 encephalopathic infants treated with whole body cooling from the participating centres. MRI and spectroscopy will be performed within 2 weeks of birth. Neurodevelopmental outcomes will be assessed at 18-24 months of age. Agreement between MR cerebral biomarkers and neurodevelopmental outcome will be reported. The sample size is calculated using the 'rule of 10', generally used to calculate the sample size requirements for developing prognostic models. Considering 9 parameters, we require 9×10 adverse events, which suggest that a total sample size of 180 is required. ETHICS AND DISSEMINATION: Human Research Ethics Committee approvals have been received from Brent Research Ethics Committee (London), and from Imperial College London (Sponsor). We will submit the results of the study to relevant journals and offer national and international presentations. TRIAL REGISTRATION NUMBER: Clinical Trials.gov Number: NCT01309711.

Open versus needle biopsy in diagnosing neuroblastoma.

BACKGROUND: Open surgical biopsy is traditionally advocated prior to initiating therapy in UKCCLG neuroblastoma protocols. We report a single centre experience comparing the utility of open biopsy vs image guided needle biopsy in aiding the definitive diagnosis and risk stratification of neuroblastoma - (Shimada classification, MYCN expression, cytogenetics - 1p 11q, 17 q). METHODS: Medical records of all new cases of neuroblastoma presenting to a single UKCCLG centre during January 2002-July 2013 were examined. RESULTS: Thirty nine patients underwent a biopsy of primary tumour for neuroblastoma during the study. Twenty one children had open biopsy and eighteen cases had a needle biopsy. Staging of neuroblastoma revealed - stage 4 (n=26), stage 3 (n=7), stage 2 (n=3) and stage 4S (n=3). Sites of primary tumour were adrenal gland (n=20), abdomen (n=12), thoracic (n=4), abdomino-thoracic (n=2) and abdomino pelvic regions (n=1). All patients (open vs needle) had adequate tissue retrieved for histological diagnosis of neuroblastoma. One needle and one open biopsy case did not have MYCN status determined despite adequate tissue sampling. Seventeen patients (7 open and 10 needle biopsies) had 1p and 17q status reported in MLPA testing (Multiplex Ligation-dependent Probe Amplification). No single patient required a repeat tumour biopsy. Morbidity in the series was minimal with only one child - open biopsy group, requiring emergent laparotomy to control bleeding from an abdominal primary tumour. No complications were recorded with needle biopsy. CONCLUSIONS: Open and image guided needle biopsy appear to yield adequate tissue sampling for diagnosis, risk classification and staging of neuroblastoma. Further larger co-operative studies may usefully guide national and international protocols.

Rapid skeletal turnover in a radiographic mimic of osteopetrosis.

Among the high bone mass disorders, the osteopetroses reflect osteoclast failure that prevents skeletal resorption and turnover, leading to reduced bone growth and modeling and characteristic histopathological and radiographic findings. We report an 11-year-old boy with a new syndrome that radiographically mimics osteopetrosis (OPT), but features rapid skeletal turnover. He presented at age 21 months with a parasellar, osteoclast-rich giant cell granuloma. Radiographs showed a dense skull, generalized osteosclerosis and cortical thickening, medullary cavity narrowing, and diminished modeling of tubular bones. His serum alkaline phosphatase was >5000 IU/L (normal <850 IU/L). After partial resection, the granuloma re-grew but then regressed and stabilized during 3 years of uncomplicated pamidronate treatment. His hyperphosphatasemia transiently diminished, but all bone turnover markers, especially those of apposition, remained elevated. Two years after pamidronate therapy stopped, bone mineral density (BMD) Z-scores reached +9.1 and +5.8 in the lumbar spine and hip, respectively, and iliac crest histopathology confirmed rapid bone remodeling. Serum multiplex biomarker profiling was striking for low sclerostin. Mutation analysis was negative for activation of lipoprotein receptor-related protein 4 (LRP4), LRP5, or TGFß1, and for defective sclerostin (SOST), osteoprotegerin (OPG), RANKL, RANK, SQSTM1, or sFRP1. Microarray showed no notable copy number variation. Studies of his nonconsanguineous parents were unremarkable. The etiology and pathogenesis of this unique syndrome are unknown.

Effect of thyroxine on brain microstructure in extremely premature babies: magnetic resonance imaging findings in the TIPIT study.

BACKGROUND: In order to assess relationships between thyroid hormone status and findings on brain MRI, a subset of babies was recruited to a multi-centre randomised, placebo-controlled trial of levothyroxine (LT4) supplementation for babies born before 28 weeks' gestation (known as the TIPIT study, for Thyroxine supplementation In Preterm InfanTs). These infants were imaged at term-equivalence. MATERIALS AND METHODS: Forty-five TIPIT participants had brain MRI using diffusion tensor imaging (DTI) to estimate white matter development by apparent diffusion coefficient (ADC), fractional anisotropy (FA) and tractography metrics of number and length of streamlines. We made comparisons between babies with the lowest and highest plasma FT4 concentrations during the initial 4 weeks after birth. RESULTS: There were no differences in DTI metrics between babies who had received LT4 supplementation and those who had received a placebo. Among recipients of a placebo, babies in the lowest quartile of plasma-free thyroxine (FT4) concentrations had significantly higher apparent diffusion coefficient measurements in the posterior corpus callosum and streamlines that were shorter and less numerous in the right internal capsule. Among LT4-supplemented babies, those who had plasma FT4 concentrations in the highest quartile had significantly lower apparent diffusion coefficient values in the left occipital lobe, higher fractional anisotropy in the anterior corpus callosum and longer and more numerous streamlines in the anterior corpus callosum. CONCLUSION: DTI variables were not associated with allocation of placebo or thyroid supplementation. Markers of poorly organised brain microstructure were associated with low plasma FT4 concentrations after birth. The findings suggest that plasma FT4 concentrations affect brain development in very immature infants and that the effect of LT4 supplementation for immature babies with low FT4 plasma concentrations warrants further study.

Importance of intraoperative magnetic resonance imaging for pediatric brain tumor surgery.

BACKGROUND: High-field intraoperative MRI (IoMRI) is gaining increasing recognition as an invaluable tool in pediatric brain tumor surgery where the extent of tumor resection is a major prognostic factor. We report the initial experience of a dedicated pediatric 3-T intraoperative MRI (IoMRI) unit with integrated neuronavigation in the management of pediatric brain tumors. METHODS: Seventy-three children (mean age 9.5 years; range 0.2-19 years) underwent IoMRI between October 2009 and January 2012, during 79 brain tumor resections using a 3-T MR scanner located adjacent to the neurosurgical operating theater that is equipped with neuronavigation facility. IoMRI was performed either to assess the extent of tumor resection after surgical impression of complete/intended tumor resection or to update neuronavigation. The surgical aims, IoMRI findings, extent of tumor resection, and follow-up data were reviewed. RESULTS: Complete resection was intended in 47/79 (59%) operations. IoMRI confirmed complete resection in 27/47 (57%). IoMRI findings led to further resection in 12/47 (26%). In 7/47 (15%), IoMRI was equivocal for residual tumor and no evidence of residual tumor was found on re-inspection. In 32/79 (41%) operations, the surgical aim was partial tumor resection. In this subset, surgical resection was extended following IoMRI in 13/32 (41%) operations. None of the patients required early second look procedure for residual disease. CONCLUSIONS: At our institution, IoMRI has led to increased rate of tumor resection and a change in surgical strategy with further tumor resection in 32% of patients. While interpreting IoMRI, it is important to be aware of the known pitfalls.

Intraoperative 3-Tesla MRI in the management of paediatric cranial tumours--initial experience.

BACKGROUND: Intraoperative MRI (ioMRI) has been gaining recognition because of its value in the neurosurgical management of cranial tumours. There is limited documentation of its value in children. OBJECTIVES: To review the initial experience of a paediatric 3-Tesla ioMRI unit in the management of cranial tumours. MATERIALS AND METHODS: Thirty-eight children underwent ioMRI during 40 cranial tumour resections using a 3-Tesla MR scanner co-located with the neurosurgical operating theatre. IoMRI was performed to assess the extent of tumour resection and/or to update neuronavigation. The intraoperative and follow-up scans, and the clinical records were reviewed. RESULTS: In 27/40 operations, complete resection was intended. IoMRI confirmed complete resection in 15/27 (56%). As a consequence, surgical resection was extended in 5/27 (19%). In 6/27 (22%), ioMRI was equivocal for residual tumour. In 13/40 (33%) operations, the surgical aim was to partially resect the tumour. In 7 of the 13 (54%), surgical resection was extended following ioMRI. CONCLUSION: In our initial experience, ioMRI has increased the rate of complete resection, with intraoperative surgical strategy being modified in 30% of procedures. Collaborative analysis of ioMRI by the radiologist and neurosurgeon is vital to avoid errors in interpretation.

Caudate and hippocampal volumes, intelligence, and motor impairment in 7-year-old children who were born preterm.

Children who survive very preterm birth without major disability have a high prevalence of learning difficulty, attention deficit, and minor motor impairment (MMI). To determine whether these difficulties are associated with structural brain abnormalities, we studied 105 preterm children (<32 wk) at 7 y with tests of IQ and MMI (Movement ABC) and detailed magnetic resonance brain scans. Scans were assessed qualitatively for visible cerebral lesions. Volume measurements of the caudate nuclei and hippocampal formations were made. Total brain volume (TBV) was estimated from the head circumference. Qualitative assessment of the scans showed evidence of cerebral lesions in 20 (19%), which were associated with lower IQ and more frequent MMI. IQ correlated with right and left caudate volume (Spearman's rho 0.304 and 0.349; p < 0.01). This association persisted (except for verbal IQ) when caudate volume was expressed as a percentage of estimated TBV to allow for overall brain size. No significant correlations with hippocampal volumes were observed. These differences persisted when only scans from children without visible lesions on scan were considered. MMI was significantly associated only with TBV and was more common in children with evidence of thinning of the posterior corpus callosum, although most children with MMI have a normal corpus callosum. Lower IQs in children who were born preterm are related to poorer development of the caudate relative to the rest of the brain, independent of other lesions. These findings suggest abnormal brain development after perinatal injury or postnatal nutritional deficits is responsible for cognitive deficits in preterm children.

Meconium periorchitis--a rare cause of a paratesticular mass.

We report a 9-month-old boy who presented with a hard paratesticular mass. Clinical and ultrasound findings were highly suggestive of malignancy and orchidectomy was performed. Histology confirmed the unexpected diagnosis of meconium periorchitis, a rare, benign condition resulting from intrauterine bowel perforation.

Classification and imaging of vascular malformations in children.

Accurate diagnosis and classification of haemangiomas and vascular malformations is essential for prediction of outcome and rational therapeutic intervention. The purpose of this review is to demonstrate how radiological techniques, including projectional radiography, ultrasound and colour flow imaging, magnetic resonance imaging, and conventional venography and arteriography can aid in the diagnosis and classification of vascular malformations in children, and to guide the radiologist in the use of modern nomenclature.

Magnetic resonance imaging and T2 relaxometry of cerebral white matter and hippocampus in children born preterm.

The objective of this study was to determine whether intelligence and minor motor impairments in children who are born preterm without major disability are associated with cerebral white matter (CWM) and hippocampal abnormalities on magnetic resonance imaging (MRI). A total of 103 preterm children were studied at age 7 y with detailed magnetic resonance brain scans, including a T2-mapping sequence from which T2 relaxation times of the CWM and hippocampal formations were calculated. All of the children had no major motor disability, attended normal school, and had undergone assessment of IQ and a test for minor motor impairment (MMI). Twenty children had visible lesions on MRI, which were associated with lower IQ and more frequent MMI. Mean (SD) IQ was 90 (14.1). Twenty-five children were shown to have MMI (Movement ABC at below the fifth centile). This group was shown to have significantly longer T2 relaxation times for CWM (mean difference 2.1 ms right, 3.1 ms left) but not the hippocampus than the children without MMI. These differences persisted when only children without visible lesions on scans were considered (mean difference 1.5 ms bilaterally). There was no significant correlation between IQ and T2 relaxation times. Children who are born preterm without subsequent major neurodisability may, in addition to visible lesions on MRI scans, have a diffuse abnormality of CWM, manifest as an increase in T2 relaxation time. This abnormality shows a close correlation with minor motor impairment but not with full-scale IQ.