Radiomic Phenotypes Distinguish Atypical Teratoid/Rhabdoid Tumors from Medulloblastoma.

BACKGROUND AND PURPOSE: Atypical teratoid/rhabdoid tumors and medulloblastomas have similar imaging and histologic features but distinctly different outcomes. We hypothesized that they could be distinguished by MR imaging-based radiomic phenotypes. MATERIALS AND METHODS: We retrospectively assembled T2-weighted and gadolinium-enhanced T1-weighted images of 48 posterior fossa atypical teratoid/rhabdoid tumors and 96 match-paired medulloblastomas from 7 institutions. Using a holdout test set, we measured the performance of 6 candidate classifier models using 6 imaging features derived by sparse regression of 900 T2WI and 900 T1WI Imaging Biomarker Standardization Initiative-based radiomics features. RESULTS: From the originally extracted 1800 total Imaging Biomarker Standardization Initiative-based features, sparse regression consistently reduced the feature set to 1 from T1WI and 5 from T2WI. Among classifier models, logistic regression performed with the highest AUC of 0.86, with sensitivity, specificity, accuracy, and F1 scores of 0.80, 0.82, 0.81, and 0.85, respectively. The top 3 important Imaging Biomarker Standardization Initiative features, by decreasing order of relative contribution, included voxel intensity at the 90th percentile, inverse difference moment normalized, and kurtosis-all from T2WI. CONCLUSIONS: Six quantitative signatures of image intensity, texture, and morphology distinguish atypical teratoid/rhabdoid tumors from medulloblastomas with high prediction performance across different machine learning strategies. Use of this technique for preoperative diagnosis of atypical teratoid/rhabdoid tumors could significantly inform therapeutic strategies and patient care discussions.

Deep Learning for Pediatric Posterior Fossa Tumor Detection and Classification: A Multi-Institutional Study.

BACKGROUND AND PURPOSE: Posterior fossa tumors are the most common pediatric brain tumors. MR imaging is key to tumor detection, diagnosis, and therapy guidance. We sought to develop an MR imaging-based deep learning model for posterior fossa tumor detection and tumor pathology classification. MATERIALS AND METHODS: The study cohort comprised 617 children (median age, 92 months; 56% males) from 5 pediatric institutions with posterior fossa tumors: diffuse midline glioma of the pons (n = 122), medulloblastoma (n = 272), pilocytic astrocytoma (n = 135), and ependymoma (n = 88). There were 199 controls. Tumor histology served as ground truth except for diffuse midline glioma of the pons, which was primarily diagnosed by MR imaging. A modified ResNeXt-50-32x4d architecture served as the backbone for a multitask classifier model, using T2-weighted MRIs as input to detect the presence of tumor and predict tumor class. Deep learning model performance was compared against that of 4 radiologists. RESULTS: Model tumor detection accuracy exceeded an AUROC of 0.99 and was similar to that of 4 radiologists. Model tumor classification accuracy was 92% with an F1 score of 0.80. The model was most accurate at predicting diffuse midline glioma of the pons, followed by pilocytic astrocytoma and medulloblastoma. Ependymoma prediction was the least accurate. Tumor type classification accuracy and F1 score were higher than those of 2 of the 4 radiologists. CONCLUSIONS: We present a multi-institutional deep learning model for pediatric posterior fossa tumor detection and classification with the potential to augment and improve the accuracy of radiologic diagnosis.

MR Imaging-Based Radiomic Signatures of Distinct Molecular Subgroups of Medulloblastoma.

BACKGROUND AND PURPOSE: Distinct molecular subgroups of pediatric medulloblastoma confer important differences in prognosis and therapy. Currently, tissue sampling is the only method to obtain information for classification. Our goal was to develop and validate radiomic and machine learning approaches for predicting molecular subgroups of pediatric medulloblastoma. MATERIALS AND METHODS: In this multi-institutional retrospective study, we evaluated MR imaging datasets of 109 pediatric patients with medulloblastoma from 3 children's hospitals from January 2001 to January 2014. A computational framework was developed to extract MR imaging-based radiomic features from tumor segmentations, and we tested 2 predictive models: a double 10-fold cross-validation using a combined dataset consisting of all 3 patient cohorts and a 3-dataset cross-validation, in which training was performed on 2 cohorts and testing was performed on the third independent cohort. We used the Wilcoxon rank sum test for feature selection with assessment of area under the receiver operating characteristic curve to evaluate model performance. RESULTS: Of 590 MR imaging-derived radiomic features, including intensity-based histograms, tumor edge-sharpness, Gabor features, and local area integral invariant features, extracted from imaging-derived tumor segmentations, tumor edge-sharpness was most useful for predicting sonic hedgehog and group 4 tumors. Receiver operating characteristic analysis revealed superior performance of the double 10-fold cross-validation model for predicting sonic hedgehog, group 3, and group 4 tumors when using combined T1- and T2-weighted images (area under the curve = 0.79, 0.70, and 0.83, respectively). With the independent 3-dataset cross-validation strategy, select radiomic features were predictive of sonic hedgehog (area under the curve = 0.70-0.73) and group 4 (area under the curve = 0.76-0.80) medulloblastoma. CONCLUSIONS: This study provides proof-of-concept results for the application of radiomic and machine learning approaches to a multi-institutional dataset for the prediction of medulloblastoma subgroups.

MRI surrogates for molecular subgroups of medulloblastoma.

BACKGROUND AND PURPOSE: Recently identified molecular subgroups of medulloblastoma have shown potential for improved risk stratification. We hypothesized that distinct MR imaging features can predict these subgroups. MATERIALS AND METHODS: All patients with a diagnosis of medulloblastoma at one institution, with both pretherapy MR imaging and surgical tissue, served as the discovery cohort (n = 47). MR imaging features were assessed by 3 blinded neuroradiologists. NanoString-based assay of tumor tissues was conducted to classify the tumors into the 4 established molecular subgroups (wingless, sonic hedgehog, group 3, and group 4). A second pediatric medulloblastoma cohort (n = 52) from an independent institution was used for validation of the MR imaging features predictive of the molecular subtypes. RESULTS: Logistic regression analysis within the discovery cohort revealed tumor location (P < .001) and enhancement pattern (P = .001) to be significant predictors of medulloblastoma subgroups. Stereospecific computational analyses confirmed that group 3 and 4 tumors predominated within the midline fourth ventricle (100%, P = .007), wingless tumors were localized to the cerebellar peduncle/cerebellopontine angle cistern with a positive predictive value of 100% (95% CI, 30%-100%), and sonic hedgehog tumors arose in the cerebellar hemispheres with a positive predictive value of 100% (95% CI, 59%-100%). Midline group 4 tumors presented with minimal/no enhancement with a positive predictive value of 91% (95% CI, 59%-98%). When we used the MR imaging feature-based regression model, 66% of medulloblastomas were correctly predicted in the discovery cohort, and 65%, in the validation cohort. CONCLUSIONS: Tumor location and enhancement pattern were predictive of molecular subgroups of pediatric medulloblastoma and may potentially serve as a surrogate for genomic testing.

Arterial spin-labeled perfusion of pediatric brain tumors.

BACKGROUND AND PURPOSE: Pediatric brain tumors have diverse pathologic features, which poses diagnostic challenges. Although perfusion evaluation of adult tumors is well established, hemodynamic properties are not well characterized in children. Our goal was to apply arterial spin-labeling perfusion for various pathologic types of pediatric brain tumors and evaluate the role of arterial spin-labeling in the prediction of tumor grade. MATERIALS AND METHODS: Arterial spin-labeling perfusion of 54 children (mean age, 7.5 years; 33 boys and 21 girls) with treatment-naive brain tumors was retrospectively evaluated. The 3D pseudocontinuous spin-echo arterial spin-labeling technique was acquired at 3T MR imaging. Maximal relative tumor blood flow was obtained by use of the ROI method and was compared with tumor histologic features and grade. RESULTS: Tumors consisted of astrocytic (20), embryonal (11), ependymal (3), mixed neuronal-glial (8), choroid plexus (5), craniopharyngioma (4), and other pathologic types (3). The maximal relative tumor blood flow of high-grade tumors (grades III and IV) was significantly higher than that of low-grade tumors (grades I and II) (P < .001). There was a wider relative tumor blood flow range among high-grade tumors (2.14 ± 1.78) compared with low-grade tumors (0.60 ± 0.29) (P < .001). Across the cohort, relative tumor blood flow did not distinguish individual histology; however, among posterior fossa tumors, relative tumor blood flow was significantly higher for medulloblastoma compared with pilocytic astrocytoma (P = .014). CONCLUSIONS: Characteristic arterial spin-labeling perfusion patterns were seen among diverse pathologic types of brain tumors in children. Arterial spin-labeling perfusion can be used to distinguish high-grade and low-grade tumors.

Prognostic role for diffusion-weighted imaging of pediatric optic pathway glioma.

Optic pathway glioma (OPG) has an unpredictable course, with poor correlation between conventional imaging features and tumor progression. We investigated whether diffusion-weighted MRI (DWI) predicts the clinical behavior of these tumors. Twelve children with OPG (median age 2.7 years; range 0.4-6.2 years) were followed for a median 4.4 years with DWI. Progression-free survival (time to requiring therapy) was compared between tumors stratified by apparent diffusion coefficient (ADC) from initial pre-treatment scans. Tumors with baseline ADC greater than 1,400 × 10(-6) mm(2)/s required treatment earlier than those with lower ADC (log-rank p = 0.002). In some cases, ADC increased leading up to treatment, and declined following treatment with surgery, chemotherapy, or radiation. Baseline ADC was higher in tumors that eventually required treatment (1,562 ± 192 × 10(-6) mm(2)/s), compared with those conservatively managed (1,123 ± 114 × 10(-6) mm(2)/s) (Kruskal-Wallis test p = 0.013). Higher ADC predicted earlier tumor progression in this cohort and in some cases declined after therapy. Evaluation of OPG with DWI may therefore be useful for predicting tumor behavior and assessing treatment response.

Cerebrovascular disease in childhood cancer survivors: A Children's Oncology Group Report.

BACKGROUND: Curative therapy for childhood cancer has dramatically improved over past decades. Therapeutic radiation has been instrumental in this success. Unfortunately, irradiation is associated with untoward effects, including stroke and other cerebrovascular disease (CVD). The Children's Oncology Group (COG) has developed guidelines for screening survivors at risk for persistent or late sequelae of cancer therapy. OBJECTIVES: This review summarizes the pathophysiology and relevant manifestations of radiation-induced CVD and outlines the specific patient groups at risk for early-onset stroke. The reader will be alerted to the availability of the COG recommendations for monitoring, and, when applicable, specific screening and treatment recommendations will be highlighted. METHODS: A multidisciplinary task force critically reviewed the existing literature and scored the evidence to establish the current COG guidelines for monitoring health of survivors treated with head and neck irradiation. RESULTS: Previous head and neck exposure to therapeutic radiation is associated with latent CVD and increased risk for stroke in some patient groups. Common manifestations of radiation-induced CVD includes steno-occlusive disease, moyamoya, aneurysm, mineralizing microangiopathy, vascular malformations, and strokelike migraines. CONCLUSION: Risk for stroke is increased in survivors of pediatric CNS tumors, Hodgkin lymphoma, and acute lymphoblastic leukemia who received radiation to the brain and/or neck. As the population of survivors ages, vigilance for stroke and cerebrovascular disease needs to continue based on specific exposures during curative cancer therapy.

Surveillance neuroimaging to detect relapse in childhood brain tumors: a Pediatric Oncology Group study.

PURPOSE: To investigate the prognostic significance of surveillance neuroimaging for detection of relapse among children with malignant brain tumors. PATIENTS AND METHODS: A historical cohort study examined all children who experienced relapse from 1985 to 1999 on one of 10 Pediatric Oncology Group trials for malignant glioma, medulloblastoma, or ependymoma. RESULTS: For all 291 patients (median age at diagnosis, 8.2 years), median time to first relapse was 8.8 months (range, 0.6 to 115.6 months). Ninety-nine relapses were radiographic, and 192, clinical; median time to relapse was 15.7 versus 6.6 months, respectively (P = .0001). When stratified by pathology, radiographic and clinical groups showed differences in median time to relapse for malignant glioma (7.8 v 4.3 months, respectively; P = .041) and medulloblastoma (23.6 v 8.9 months, respectively; P = .0006) but not ependymoma (19.5 v 13.3 months, respectively; P = .19). When stratified by early (< 8.8 months) or late (> or = 8.8 months) time to relapse, 115 early relapses were clinical, and 32, radiographic; for late relapses, 77 were clinical, and 67, radiographic (P = .001). Overall survival (OS) from relapse was significantly longer for radiographic compared with clinical detection (median, 10.8 months; 1-year OS, 46% v median, 5.5 months; 1-year OS, 33%; P = .002), but this trend did not retain significance when analyzed by pathology subgroups. CONCLUSION: Surveillance neuroimaging detects a proportion of asymptomatic relapses, particularly late relapses, and may provide lead time for other therapies on investigational trials. During the first year after diagnosis, radiographic detection of asymptomatic relapse was infrequent. A prospective study is needed to formulate a rational surveillance schedule based on the biologic behavior of these tumors.

Daily low-dose carboplatin as a radiation sensitizer for newly diagnosed malignant glioma.

Surgical resection followed by local field radiotherapy is currently our most effective approach to treatment for most patients with malignant glioma. Carboplatin chemotherapy has direct cytotoxic effects on glioma cells and acts as a radiation sensitizer to enhance cell killing. Its demonstrated efficacy as a sensitizer in other solid tumors led to this clinical trial of carboplatin as a radiation sensitizer in the treatment of newly diagnosed glioblastoma multiforme (GBM) and anaplastic astrocytoma (AA). Fourteen patients (nine GBM and five AA) were treated with daily low-dose carboplatin 25 mg/m2 intravenously within 2 h of their fractionated radiotherapy to a total dose of 600 mg/m2. No significant toxicities attributable to this combined therapy were observed. All patients have progressed, with median time to progression of 16 weeks. Eleven patients have died, with median survival of 38 weeks for the entire cohort. Although this regimen appeared safe, there was no benefit in survival time compared to historical patients treated with radiotherapy. The limitations and future potential for the strategy of radiation sensitization are discussed.

Weekly dosing of carboplatin increases risk of allergy in children.

BACKGROUND: Carboplatin (CBDCA) has been used increasingly to treat pediatric low-grade gliomas. Allergic reactions to CBDCA have been reported in 2% to 30% of children. The reason for this high incidence of allergy is unclear. METHODS: To determine the risk factors for CBDCA allergy, an historic cohort study was conducted for all children who received the drug during a 6-year period at the Lucile Salter Packard Children's Hospital at Stanford. The patients' medical records were reviewed for data on age, tumor type, CBDCA dose schedule, total number of doses, cumulative dosage, dose per treatment, other chemotherapy administered, and allergic reaction. RESULTS: Fifty-four children (mean age 7.2 years, 35 boys) were identified. Six children (11.1%) had an allergic reaction to CBDCA. All reactors had low-grade gliomas treated with weekly CBDCA and vincristine, with a dosage per treatment <500 mg/m2. Overall, six (75%) of eight children administered weekly CBDCA, 6 (46.2%) of 13 children with brain tumors, and 6 (40%) of 15 administered CBDCA dosage <500 mg/m2 manifested allergic reactions. Patients receiving more than five doses had significant risk for CBDCA allergy (relative risk [RR] = 11.8; 95% confidence interval [CI]: 1.5-94.1). Using logistic regression with multiple variables, weekly dose schedule was the most predictive covariate for allergic reaction (P < 0.000 1), and other factors were unrelated or redundant. CONCLUSIONS: Children with low-grade gliomas receiving CBDCA weekly are at significantly increased risk for CBDCA allergy. The repetitive, weekly dosing schedule of CBDCA appears to be a key risk factor for allergic reaction in brain tumor patients. The high frequency of allergy with weekly CBDCA warrants further consideration when planning future trials.

A guide to children with acute and chronic headaches.

Children with acute and chronic headaches are often seen by primary care providers. A complete, elaborate history, obtained from both the parents and child, is key in diagnosing and managing the child who presents with a headache. A thorough social and educational history may reveal significant school or family stresses. Historic features of concern must be explored immediately. A thorough physical examination with a focused neurologic examination must be done Focal neurologic findings may indicate serious organic problems. A comprehensive approach to the management of headaches in children consisting of reassurance, education, pharmacologic interventions, and nonpharmacologic interventions is presented. A two-tiered management plan is used in conjunction with medications. The aim of this article is to provide the novice or experienced practitioner with a comprehensive review of acute and chronic headache pathogenesis, assessment, and management. This review includes migraines and other nonmigraine types of headaches.

Visual loss caused by pseudotumor cerebri in an infant on peritoneal dialysis.

Infants with chronic renal insufficiency have multiple risk factors for developing pseudotumor cerebri (PTC) and are at particular risk for being diagnosed with PTC late, because of their inability to express symptoms. We describe a 13-month-old infant dependent on peritoneal dialysis, without evidence of central nervous system infection or inflammation, who developed acute vision loss secondary to PTC. Signs of PTC in infants include torticollis, inattentiveness, inability to track, facial paresis, or new-onset strabismus. Physicians responsible for the care of children with renal failure should be aware of the potential for PTC, as the diagnosis should be made as early as possible to prevent permanent visual loss.

Prognostic implications for gadolinium enhancement of the meninges in low-grade astrocytomas of childhood.

BACKGROUND: Persistent gadolinium enhancement on MRI of the meninges in some children with low-grade astrocytomas (LGA) is a widely recognized phenomenon. The relationship of this finding with the clinical course is unclear. METHODS: From a consecutive cohort of 282 children with pathologically confirmed LGA we identified all patients with asymptomatic gadolinium enhancement of the meninges found on surveillance MRI. A nested case-control study was performed, comparing patients with meningeal enhancement to controls without enhancement. RESULTS: Twenty-one children were identified with meningeal enhancement. The median follow-up was 5.2 years with enhancement noted for a median of 2.2 years. The 5-year overall survival for this cohort was 91.2% (Greenwood SE 8.0%), and the 5-year progression-free survival was 20.9% (SE 11.9%). Five patients are now free of disease, while 15 continue to have stable disease. The overall and progression-free survival was not significantly different compared to controls. CONCLUSIONS: Gadolinium enhancement of the meninges on MRI may occur in a significant number of children with LGA, particularly juvenile pilocytic astrocytoma, but does not appear to affect progression-free or overall survival. Change in management based on this finding alone is unwarranted.

A clinicopathologic reappraisal of brain stem tumor classification. Identification of pilocystic astrocytoma and fibrillary astrocytoma as distinct entities.

BACKGROUND: Brain stem tumors in children have been classified pathologically as low grade or high grade gliomas and descriptively as diffuse gliomas, intrinsic gliomas, midbrain tumors, tectal gliomas, pencil gliomas, dorsal exophytic brain stem tumors, pontine gliomas, focal medullary tumors, cervicomedullary tumors, focal gliomas, or cystic gliomas. METHODS: To search for a simplified and prognostic clinicopathologic scheme for brain stem tumors, the authors reviewed a consecutive cohort of patients younger than age 21 years with tumors diagnosed from 1980 through 1997. Pathology specimens and neuroimaging were classified by masked review. Statistical and survival analysis along with Cox proportional hazards regression was performed. RESULTS: Seventy-six patients were identified, with initial diagnostic magnetic resonance imaging available for 51 and pathology specimens for 48 patients. Twenty cases were classified histologically as pilocytic astrocytoma (PA), 14 as fibrillary astrocytoma (FA), and 14 as other tumors or indeterminate pathology. For all tumors, characteristics significantly associated with a worse survival rate were: symptom duration less than 6 months before diagnosis (P = 0.004); abducens palsy at presentation (P < 0.0001); pontine location (P = 0.0002); and engulfment of the basilar artery (P = 0.006). Pilocytic astrocytoma was associated with location outside the ventral pons (P = 0.001) and dorsal exophytic growth (P = 0.013); Fibrillary astrocytoma was associated with symptoms less than 6 months (P = 0. 006), abducens palsy (P < 0.001), and engulfment of the basilar artery (P = 0.002). Pilocytic astrocytoma showed 5-year overall survival (OS) of 95% (standard error [SE], 5%) compared with FA 1-year OS of 23% (SE, 11%;P < 0.0001). CONCLUSIONS: Brain stem tumors can be succinctly and better biologically classified as diffusely infiltrative brain stem gliomas-generally FA located in the ventral pons that present with abducens palsy, often engulf the basilar artery, and carry a grim prognosis-and focal brain stem gliomas-frequently PA arising outside the ventral pons, often with dorsal exophytic growth, a long clinical prodrome, and outstanding prognosis for survival. Our findings emphasize the individuality of PA as a distinct clinicopathologic entity with an exceptional prognosis.

Case study: suprasellar germinoma presenting with psychotic and obsessive-compulsive symptoms.

This case describes a 13-year-old boy who had a suprasellar germinoma involving the bilateral basal ganglia. His presenting symptoms included left-sided weakness, diabetes insipidus, a decline in academic functioning as well as psychotic and obsessive-compulsive symptoms. His neuroradiological findings and clinical symptoms lend support to the potential role of the basal ganglia in psychotic and obsessive-compulsive symptomatology.

Pediatric astrocytomas with monomorphous pilomyxoid features and a less favorable outcome.

Among tumors classified as pilocytic astrocytoma (PA) in the Johns Hopkins Hospital Department of Pathology files, we identified 18 cases with a distinctive monomorphous pilomyxoid histological pattern and a higher recurrence rate than that of PA with classical histological features (classical PA). The majority of the tumors occurred in infants and young children and involved the hypothalamic/chiasmatic region. The tumors were histologically similar to PA, but they were more monomorphous and more myxoid. Rosenthal fibers were not seen and only 1 of 18 tumors had eosinophilic granular bodies. At the end of the follow-up period, 6 patients were dead and 12 were alive with evidence of disease. Progression free survival (PFS) at 1 year was 38.7%. In comparison, we identified a control group of 13 classical PAs in the same age range and location as the study group. In this group, PFS at 1 year was 69.2%, which was significantly better than that for pilomyxoid tumors (p = 0.04). There was no CSF dissemination or death due to tumor progression among patients with classical PA. Eight of these patients are alive with recurrent disease, and 4 have no evidence of disease. While the monomorphous pilomyxoid tumors have some resemblance to classical PA, our results suggest that the former is a more aggressive variant or a separate entity that needs to be recognized for prognostic purposes.