Skip to main content
NCBI home page
Neuro-Oncology logoLink to Neuro-Oncology
. 2012 Jun;14(Suppl 1):i140–i147. doi: 10.1093/neuonc/nos107

RADIOLOGY

PMCID: PMC3483353
Neuro Oncol. 2012 Jun;14(Suppl 1):i140–i147.

RA-01. ATYPICAL NEUROIMAGING CHARACTERISTICS IN CHILDREN WITH EMBRYONAL TUMOR WITH ABUNDANT NEUROPIL AND TRUE ROSETTES (ETANTR)

Jeffrey Murray 1, Emily Braly 1, Hayden Head 1, David Donahue 1

Abstract

INTRODUCTION: Pediatric CNS embryonal neoplasms represent a unique group of poorly differentiated malignant tumors with propensity to disseminate throughout the neuraxis and exhibit aggressive clinical behavior. They are typically classified as medulloblastoma, other CNS site PNET (e.g., pineoblastoma) and AT/RT. A novel PNET variant, ETANTR, has been reported in approximately 30 children worldwide. ETANTR appears to afflict young children, have a female predominance and occurs primarily in the cerebrum. The prognosis is dismal. We report atypical neuroimaging characteristics of two young children with ETANTR. REPORTS: A 5-year-old girl presented with headaches, ataxia and photophobia. MRI revealed a 5.1 x 7.4 x 6 cm mass seeming to arise from the right lateral ventricle. The mass showed T1 hypointensity, T2 hyperintensity and was isointense to gray matter on FLAIR sequences. DWI and ADC images showed patchy reduced diffusion. T1 post-contrast images showed enhancement of large vessels, but little parenchymal tumor enhancement. Gross total resection (GTR) was achieved, ETANTR was diagnosed and there was no evidence of neuraxis metastases. A 2-year-old boy presented with seizures. MRI revealed a 2.2 x 2.9 x 2.2 cm mass in the right frontal lobe parenchyma, showing the same characteristics as the first case, except for a thin, partial FLAIR-hyperintense rim, partial reduced diffusion, and post-contrast enhancement only of a few small vessels. GTR was achieved, ETANTR was diagnosed and there was no evidence of neuraxis metastases. CONCLUSIONS: ETANTR is now recognized as a distinct type of CNS embryonal tumor/PNET. It occurs in young children and carries a poor prognosis. The neuroimaging characteristics in the two reported cases here, namely FLAIR isointensity to gray matter and paucity of parenchymal tumor enhancement, are atypical of malignant embryonal tumors and should be considered as a factor in the neurodiagnostic differential diagnosis of children with newly discovered CNS tumors.

Neuro Oncol. 2012 Jun;14(Suppl 1):i140–i147.

RA-02. VOLUME OF VESTIBULAR SCHWANNOMA DOES NOT CORRELATE WITH HEARING LOSS IN PEDIATRIC PATIENTS WITH NEUROFIBROMATOSIS II

Sarah Rush 1, Nicholas Stence 1, Arthur Liu 1

Abstract

INTRODUCTION: Vestibular schwannomas are common tumors in individuals affected by Neurofibromatosis type II (NF2). NF2 is often diagnosed in the third and fourth decades of life and as such there is little known about pediatric patients with NF2. Hearing loss is a common result of vestibular schwannomas in adults, but much remains unknown about the correlation between the size of vestibular schwannomas and their effects on hearing in the pediatric population. METHODS: MRI data and hearing data were obtained for eight children with NF2. All patients were between the ages of 10 and 21 at evaluation. Vestibular schwannoma tumor volumes were measured, and hearing scores were calculated based on the CTCAE version 4.0. RESULTS: The mean age of the population assessed was 15 years. Six of eight children were found to have hearing loss in at least one ear. Four of eight were noted to have hearing loss of 20db at 3Hz and above. The analysis of the data found no correlation between the size of the vestibular schwannoma and the degree of hearing loss. It was noted that tumors as small as 92mm3 were associated with grade II hearing loss while tumors as large as 904mm3 were associated with no loss of hearing. CONCLUSIONS: We found no correlation between size of vestibular schwannomas and the degree of hearing loss associated with these tumors in children. Often children with small vestibular schwannomas are not screened for hearing loss. Our findings suggest that using tumor size on MRI as a criterion for hearing screening could result in missed diagnoses of potentially salvageable hearing loss. At our institution, we recommend that all children with NF2 should have regular hearing screens regardless of tumor size. These findings represent a single institution's experience and need to be replicated on a larger scale.

Neuro Oncol. 2012 Jun;14(Suppl 1):i140–i147.

RA-03. IMAGING FEATURES ON MRI IN THE DIFFERENT SUBTYPES OF MEDULLOBLASTOMAS

Julia Kleinhenz 1, Brigitte Bison 1, Torsten Pietsch 2, Katja von Hoff 3, Andre von Bueren 3, Stefan Rutkowski 3, Monika Warmuth-Metz 1

Abstract

PURPOSE AND PATIENTS: The latest WHO classification (2007) differentiates 5 subtypes of medulloblastomas (MB): classic, desmoplastic, extensively nodular (MBEN), anaplastic and large cell. We retrospectively evaluated the imaging data of children and adolescents at the time of diagnosis collected during central review of craniospinal magnetic resonance imaging (MRI) and if present computed tomography (CT) of the patients registered to the prospective multicenter trial HIT 2000 for characteristic demographic, localization and imaging features of these subtypes. The evaluation comprised age, gender, macroscopic metastasis-status, tumor localization and imaging features like contrast behavior, delimitation, edema, T1 and T2-signal intensity and density. RESULTS: We identified 460 patients with a centrally reviewed and histologically confirmed diagnosis of MB according to the 2007 WHO-criteria (356 classic, 75 desmoplastic, 15 MBEN, 7 anaplastic and 7 large cell MB). Striking differences were the fact that classic MBs were nearly exclusively localized in the vermis and 4. ventricle but 28% of desmoplastic tumors were found in the hemispheres. Classic MBs were accumulating contrast media least and in the smallest percentage of tumor volume. All children with MBEN were at or below the age of 3 years and frequently showed a typical pattern of contrast enhancement and only rarely a meningeal dissemination while the rare large cell subtype was accompanied by a very high frequency of meningeal dissemination (70%). CONCLUSION: The defined histological subtypes of MBs show characteristic differences in demographic and imaging features and a rough correlation can be attempted based on the results of MRI.

Neuro Oncol. 2012 Jun;14(Suppl 1):i140–i147.

RA-04. GENERAL IMAGING PROTOCOL FOR THE SIOP-E BRAIN TUMOUR STUDIES

Tim Jaspan 2, Herve Brisse 3, Paolo Potepan 4, Monika Warmuth-Metz 1

Abstract

Response assessment based on imaging has been accepted as the gold standard for many brain tumour studies in children. The results of the recently reported SIOP PNET IV study showed a crucial role of a central response assessment. Between 20 and 25% of discrepancies are found in the HIT-studies between the local and the central evaluation and the quality of MR-imaging varies widely. For a postoperative evaluation of a residual tumour the correct time period of day 1 and 2 after surgery is crucial and non specific changes on MRI have to be known to avoid a false interpretation. Imaging and response evaluation is up the national groups of SIOP-E. To harmonize the response assessment in the SIOP-E groups to obtain comparable results across Europe and to facilitate the imaging prescription of the various SIOP studies the reference radiologists agreed on a general imaging protocol of the brain and spinal canal as a minimum requirement. In the SIOP-E Brain Imaging Group a discussion on the future development of the imaging prescriptions and the implementation of sophisticated imaging tool has been instituted. Details of this agreement contain standard imaging rules, examples of mistakes, agreement of the mode of tumour measurement and the way of response evaluation for a postoperative residual tumour or on follow-up.

Neuro Oncol. 2012 Jun;14(Suppl 1):i140–i147.

RA-05. MR-FEATURES OF EPENDYMOBLASTOMAS

Frank Berg 1, Brigitte Bison 1, Torsten Pietsch 2, Nicolas Gerber 3, Stefan Rutkowski 4, Monika Warmuth-Metz 1

Abstract

PURPOSE: Ependymoblastomas are rare tumors of childhood with a high malignancy and neuropathologically grouped together with stPNETs. Neuroradiological descriptions are rare. A larger patient group with this tumor is described and compared to ependymomas and stPNETs. PATIENTS AND METHODS: In the database of the German national reference center for neuroradiology 12 children with an ependymoblastoma confirmed by central reference pathology could be identified. Their MR-imaging results at presentation were compared to 12 contiguous children each with a verified ependymoma and stPNET out of the same database. RESULTS: Ten children had a supratentorial and 2 an infratentorial tumor. Mean age was 2.4y (9m – 4y) compared to children with a PNET (mean 5.05y; 0.33-14.25y) and an ependymoma (mean 9.13; 0.58-16.92y). The tumors were mainly very large (mean 123 ml; 3.4-403 ml)(PNET mean 69.9ml; ependymoma mean 95ml). 9 were located within or in close relation to the ventricles (PNET 8/12; ependymoma 7/12). The demarcation to the surrounding brain was mainly (10/12) sharp (PNET 9/12; ependymoma 11/12), perifocal edema minimal (mean 0.1 cm; 0-0.5 cm) (PNET mean 0.63cm; ependymoma mean 2.1cm) and contrast enhancement mainly scarce (11/12 minimal to moderate enhancement) (PNET 9/12; ependymoma 3/12). All these features are much more comparable to the findings in supratentorial hemispheric PNETs than in ependymomas. CONCLUSION: Ependymoblastomas show distinct MR-imaging features which are more similar to those in stPNETs than in ependymomas.

Neuro Oncol. 2012 Jun;14(Suppl 1):i140–i147.

RA-06. MAGNETIC RESONANCE SPECTROSCOPIC DETECTION OF LACTATE IS PREDICTIVE OF A POOR PROGNOSIS IN PATIENTS WITH DIFFUSE INTRINSIC PONTINE GLIOMA

Kazuhiko Sugiyama 1, Kaoru Kurisu 1, Yoshinori Kajiwara 1, Takashi Takayasu 1, Taiichi Saito 1, Ryosuke Hanaya 2, Fumiyuki Yamasaki 1

Abstract

Diffuse brainstem glioma has a poor prognosis, and there are few long-term survivors. We looked for clinical, conventional magnetic resonance (MR), and MR spectroscopic (MRS) findings predictive of the prognosis of patients with brainstem glioma. Our institutional review board approved this retrospective study of 23 patients with diffuse intrinsic pontine or diffuse medullary brainstem glioma treated during the period 2000-2009. To evaluate prognostic values, we performed a Kaplan-Meier survival analysis (log-rank test) that incorporated the patients' age and sex, symptom duration, the presence or absence of cranial nerve palsy, long tract sign, ataxia, and cysts, the chemotherapeutic regimen, Gd enhancement, longitudinal and cerebellar extension, basilar artery encasement, and MRS parameters. Of the 23 diffuse brainstem gliomas, 19 were located at the pons (ratio of male to female patients, 1.1:1). The mean age of the 23 patients was 15.9 years (range, 4-50 years); 16 were aged <20 years. The duration of overall survival was 19.7 months; in patients with diffuse intrinsic pontine glioma, it was 16.6 months, and in patients aged <20 years, it was 11.8 months. Clinical and conventional MR findings at presentation were not predictive of the prognosis in children with diffuse intrinsic pontine glioma. In addition, a patient age <20 years and the detection of lactate by MRS were poor prognostic factors. The MRS detection of lactate is a prognostic factor in patients with diffuse intrinsic pontine glioma. Additional studies of larger patient populations using other imaging modalities are needed.

Neuro Oncol. 2012 Jun;14(Suppl 1):i140–i147.

RA-07. ACCURATE CLASSIFICATION OF CHILDHOOD BRAIN TUMOURS BY IN VIVO 1H MRS - A MULTI-CENTRE STUDY

Javier Vicente 1, Elies Fuster-Garcia 1, Salvador Tortajada 1, Juan M García-Gómez 1, Nigel Davies 2, Kal Natarajan 2, Martin Wilson 2, Richard G Grundy 3, Pieter Wesseling 4, Daniel Monleón 5, Bernardo Celda 6, Montserrat Robles 1, Andrew C Peet 2

Abstract

AIMS: To evaluate the automatic classification of pediatric brain tumours by the non-invasive MRI based technique of 1H-MRS using a single-voxel at two different echo times (TE). Our hypotheses are (1) that the classification provides an accurate non-invasive diagnosis in multi-centre datasets and (2) the combination of TE improves the accuracy of classification compared with the use of one TE alone. METHODS: 78 patients under 16 years old with histologically proven brain tumours from 10 international centres were investigated. Discrimination of 29 medulloblastomas, 11 ependymomas and 38 pilocitic astrocytomas was evaluated. Single voxel MRS was undertaken prior to diagnosis (1.5Tesla PRESS, PROBE or STEAM, TE 20-32 ms, and PRESS, 135-136 ms). MRS data was processed using two strategies, determination of metabolite concentrations using TARQUIN software and automatic feature extraction with Peak Integration. Linear Discriminant Analysis was applied to this data to produce diagnostic classifiers. A stringent evaluation of the diagnostic accuracy was performed based on resampling of the available cases to measure the Balanced Accuracy Rate (BAR) where a BAR of 1.0 is 100%. RESULTS: The accuracy of the diagnostic classifiers for discriminating the three tumour types was found to be high (BAR 0.98) when a combination of TE was used. The combination of both TE significantly improved the classification performance (p < 0.01, Tukey's test) compared with the use of one TE alone. Other tumour types were classified accurately as glial or primitive neuroectodermal (BAR 1.00). CONCLUSIONS: 1H-MRS with combined Short and Long TE has very good accuracy for the non-invasive diagnosis of common childhood brain tumours. The method is widely available and robust when applied in a multicentre environment. It should become part of routine clinical assessment for these children.

Neuro Oncol. 2012 Jun;14(Suppl 1):i140–i147.

RA-08. INCIDENTAL FINDINGS OF MASS LESIONS ON NEUROIMAGING OF CHILDREN

Corinne Perret 1, Eugen Boltshauser 1, Ianina Scheer 1, Christian Kellenberger 1, Michael Grotzer 1

Abstract

Increasing use of neuroimaging in children has led to more incidental findings of central nervous system (CNS) mass lesions, the management of which is uncertain. The aims of this study are to describe these mass lesions and their evolution, as well as to discuss management options, and to determine the prevalence of incidental CNS mass lesions in our pediatric clinic, A retrospective study was undertaken on children under 18 years old with primary CNS tumors, admitted to the University Children's Hospital of Zurich, Switzerland, from January 1995 to December 2010. In 19 (5.7%) of 335 patients with newly diagnosed CNS tumors, the diagnosis of CNS mass lesion was an incidental finding. Reasons for neuroimaging in these 19 patients were: head trauma (n = 6), neurological evaluation (n = 4), research protocols (n = 3), malformations (n = 2), seizures (n = 1), endocrinological evaluation (n = 1), psychiatric evaluation (n = 1), and orbital lymphangioma (n = 1). 7 patients underwent immediate surgery: low-grade glioma (n = 4); craniopharyngioma (n = 1); ependymoma (n = 1); choroid plexus papilloma (n = 1) and 12 were treated conservatively/observation. 10/12 conservatively treated patients remained stable (median follow-up time: 1.8 years) and 2/12 underwent delayed surgery because of tumor progression (medulloblastoma (n = 1); fibrillary astrocytoma (n = 1)). Clinicians are increasingly challenged by the discovery of incidental CNS mass lesions. A subgroup of such lesions (with typical imaging pattern such as tectal glioma and dysembryoplastic neuroepithelial tumor) can be followed conservatively, clinically and radiographically. Future prospective studies are needed to define optimum management strategies based on larger collections of natural histories, as well as to assess the true prevalence of incidental CNS mass lesions.

Neuro Oncol. 2012 Jun;14(Suppl 1):i140–i147.

RA-09. PONTINE MEASUREMENTS PREDICT SURVIVAL IN CHILDREN WITH DIFFUSE INTRINSIC PONTINE GLIOMA (DIPG)

Emilie Steffen-Smith 1, David Venzon 2, Robyn Bent 1, Eva Baker 3, Shaefali Shandilya 1, Katherine Warren 1

Abstract

PURPOSE: Diagnosis and management of DIPG relies upon assessment of changes on magnetic resonance imaging (MRI). However, heterogeneous radiographic appearance and interobserver variability in tumor measurement complicate interpretation of tumor response and progression. This study evaluated the prognostic value of pontine measurements rather than tumor measurements in patients with DIPG. METHODS: Patients (<21 years) with newly diagnosed, refractory or progressive DIPG underwent MRI evaluation according to their treatment protocol or as clinically indicated. Two experienced readers evaluated scans independently using axial pre- and post-contrast fluid attenuated inversion recovery (FLAIR) images and consistent anatomical landmarks to determine maximum diameter (1D) and maximal bidimensional measurements (2D) of the pons. Variability between readers was estimated using the coefficient of variation. The prognostic value of averaged 1D- and 2D-measurements from the two readers was evaluated using a univariate Cox proportional hazards model. RESULTS: Seventy-five patients (median = 6 years, range = 2 - 17 years) were evaluated; 63 died during the follow-up period. Median survival from the first scan was 38 weeks (range = 2 - 502+ weeks). Readers evaluated 386 scans. Variability between readers for 1D and 2D-measurements was 3.1% and 1.2%, respectively. 2D-measurements obtained at the first scan were predictive of survival (relative hazard, RH = 2.62, p = 0.007). Increases over time in both 1D- and 2D-measurements predicted shorter overall survival (RH = 1.85, p = 0.0085 and RH = 3.29, p = 0.0047, respectively). CONCLUSIONS: Pontine measurements were obtained with low variability between readers compared to previous studies of tumor measurements. The size of the pons, rather than tumor measurement only, was predictive of outcome. As expected, patients with increases in either 1D or 2D pontine measurements over time had an increased risk of death. Incorporation of pontine measurements over the course of treatment is warranted to decrease interobserver variability in measurement and to further evaluate the prognostic value of this technique.

Neuro Oncol. 2012 Jun;14(Suppl 1):i140–i147.

RA-10. DEVELOPMENT OF MRI DIFFUSION TENSOR IMAGING OF OPTIC NERVE GLIOMAS IN CHILDREN WITH NEUROFIBROMATOSIS TYPE 1

Chie-Schin Shih 1, John West 1, Chang Ho 1, David Porter 2, Yang Wang 1, Andrew Saykin 1, Brenna McDonald 1, Konstantinos Arfanakis 3

Abstract

Diffusion Tensor imaging (DTI) is a novel MRI technique which shows promise in qualitatively and quantitatively assessing the integrity of white matter tracts and nerves within the brain. Children with neurofibromatosis type 1 have up to a 40% incidence of developing optic nerve glioma with associated morbidity of visual loss. DTI may increase the sensitivity of MRI and allow early detection of optic nerve damage in this high-risk patient population. However, susceptibility artifacts that are common in typical DTI single-shot echo planar imaging (EPI) introduce severe artifacts in areas of interest such as the optic tracts, particularly through the bony optic canal. In this work, we used a multi-shot DTI sequence named RESOLVE to greatly reduce typical image artifacts and allow assessment of the microstructural integrity of the optic nerves. This study was done in sequence with 5 healthy adult volunteers, 3 children with NF1 without tumor, and 4 children with NF1 with optic glioma. Based on this pilot study, we have solidified a protocol that allows us to image the optic nerves, chiasm, bony canal, and radiation at a quality not available with typical DTI imaging. A final axial protocol, consisting of 7 shots, 12 diffusion directions, 3 b-values, 1.2 x 1.2 x 2.2 mm3 voxel resolution, and 16 slices was chosen. Positioning of the slices was done so as to cover the optic tracts, optic chiasm, and as much of the optic radiations as possible. The data set consisting of subjects with NF1 with and without tumors will be utilized to develop a clinical analysis protocol. A clinical trial is under development to assess the sensitivity of this protocol in the early detection of optic nerve gliomas and correlation with functional compromise with patients with neurofibromatosis type 1.

Neuro Oncol. 2012 Jun;14(Suppl 1):i140–i147.

RA-11. RESPONSE ASSESSMENT IN PEDIATRIC NEURO-ONCOLOGY (RAPNO): CURRENT PRACTICE AND ISSUES

Katherine Warren 1, Gilbert Vezina 2, Darren Hargrave 3, Tina Young Poussaint 4, Stewart Goldman 5, Roger Packer 2, Patrick Wen 6, Ian Pollack 7, David Zurakowski 4, Larry Kun 8, Michael Prados 9, Mark Kieran 4

Abstract

Criteria for FDA approval of a new drug include demonstration of safety and effectiveness. Efficacy is typically estimated in the context of Phase II investigational trials. Patients with brain tumors are followed by MRI to determine the effects of therapy, with response being defined using various iterations of the MacDonald criteria. However, the paradigm of clinical drug investigation was developed in the era of cytotoxic chemotherapeutic agents and assumes a relationship between dose and toxicity or response. Limitations of this model have been identified, coinciding with the initiation of clinical trials involving cytostatic, molecularly targeted and antiangiogenic agents. The RANO (Response Assessment in Neuro-Oncology) working group was established to address the limitations in defining endpoints for clinical trials in adult neuro-oncology and to develop standardized response criteria for clinical trials in brain tumor patients. RAPNO was created as a counterpart to the RANO working group as a number of issues unique to pediatric neuro-oncology have been identified. In addition to pseudoresponse, pseudoprogression, and interobserver variability observed in both adult and pediatric patients, issues pertaining more specifically to pediatric neuro-oncology include histologic heterogeneity, relatively small numbers of patients available for clinical trials, and the lack of standard definitions of response or progression. Tumor enhancement is often variable, nonspecific, and not representative of tumor burden; decreased tumor size does not always correlate with improved patient survival. The primary goal of RAPNO is to develop a consensus on the radiology assessment to be used as endpoints in clinical trials of children with brain tumors. We will discuss response criteria utilized in current clinical trials, the challenges with their use, and the work of RAPNO in moving toward development of appropriate study endpoints in pediatric neuro-oncology.

Neuro Oncol. 2012 Jun;14(Suppl 1):i140–i147.

RA-12. DESMOPLASTIC INFANTILE GANGLIOGLIOMAS - IMAGING APPEARANCE DISCRETELY DIFFERENT FROM PREVIOUSLY DESCRIBED

Laurence Eckel 1, Gesina Keating 1, Caterina Giannini 1, Nicholas Wetjen 1, Alice Patton 1

Abstract

PURPOSE: To describe a varied appearance of desmoplastic infantile gangliogliomas. METHODS: This retrospective study was approved by our Institutional Review Board. We screened our pathology database for all pathologically proven desmoplastic infantile gangliogliomas in patients 17 years of age or younger at the time of diagnosis. All cases since 1997 were included, and only those patients with preoperative electronic imaging were scrutinized. We evaluated the following parameters: age at presentation, gender, dominant clinical features, pathologic diagnosis, location, enhancement characteristics, size, and cystic component. RESULTS: 5 patients (3 males, 2 females, mean age of 4 months, ranging from 2 months to 6 months of age) were identified. MRI was performed in all these patients and only two (40%) of these demonstrated the classic description of a massive, predominately cystic, frontoparietal tumor, with peripheral solid and deep cystic portions. Instead, 3 (60%) shared a common suprasellar location and appearance, being relatively large and homogenously enhancing, without cystic change. CONCLUSIONS: Desmoplastic infantile gangliogliomas may not be either cystic or frontoparietal, as classically described in the literature. Indeed, in an infant with a relatively large homogenously enhancing, suprasellar, solid tumor, desmoplastic infantile ganglioglioma should be a radiographic differential consideration.

Neuro Oncol. 2012 Jun;14(Suppl 1):i140–i147.

RA-13. RELATIONSHIP BETWEEN DIFFUSION TENSOR IMAGING (DTI) AND MAGNETIC RESONANCE SPECTROSCOPIC IMAGING (MRSI) IN DIFFUSE INTRINSIC PONTINE GLIOMA (DIPG)

Emilie Steffen-Smith 1, Joelle Sarlls 2, Carlo Pierpaoli 3, Lindsay Walker 3, David Venzon 4, Robyn Bent 1, Katherine Warren 1

Abstract

BACKGROUND: MRSI and DTI are noninvasive techniques used to investigate the characteristics of brain tumors. MRSI evaluates tissue metabolism, typically elevated in active tumors, while DTI reflects tissue microstructure with lower mean diffusivity (MD) observed in regions of higher cellularity, such as high-grade tumor. We explored the relationship between MRSI and DTI in children with DIPG. METHODS: Patients with DIPG underwent longitudinal MRI evaluations including MRSI and DTI. Pre-contrast axial fluid attenuated inversion recovery (FLAIR) images were used to co-register MD and MRSI maps. Regions of interest included the entire tumor identified on FLAIR. Voxels with the maximum choline:N-acetylaspartate (Cho:NAA) and minimum MD value were identified, presumably corresponding to areas of highest metabolic activity and highest cellularity, respectively. MRSI and DTI were considered co-localized if max Cho:NAA and minimum MD were from the same or adjacent voxels. The relationship between MRSI and DTI was evaluated using percent co-localization and repeated measures analysis of variance. RESULTS: Thirty patients (median = 5.4 yrs, range = 1.8-14.1 yrs) were evaluated, with a total of 64 scans. Max Cho:NAA ranged from 0.7 to 7.9 (median = 2.0). Minimum MD ranged from 0.8 to 1.4 x 10− 6mm2/s (median = 1.0 x 10− 6mm2/s). Co-localization between max Cho:NAA and minimum MD voxels was 53%. Correlation between Max Cho:NAA and minimum MD was not significant (p = 0.27). CONCLUSIONS: Maximum Cho:NAA and minimum MD values did not co-localize well. Tumor areas with greatest metabolic activity determined by MRSI typically did not correspond to regions of minimum MD determined by DTI. In addition, all minimum MD values were greater than that of normal tissue, indicating a broad influence of edema that may mask areas of increased cellularity. Max Cho:NAA and minimum MD do not seem to provide complementary information about DIPG. Further investigation is needed to understand the relationship of results from these modalities in patients with DIPG.

Neuro Oncol. 2012 Jun;14(Suppl 1):i140–i147.

RA-14. SURVEILLANCE SPINE MRIS TO DETECT TUMOR RELAPSE IN PATIENTS WITH MEDULLOBLASTOMA

Sebastien Perreault 1, Robert Lober 1, Kristen Yeom 1, Anne-Sophie Carret 2, Hannes Vogel 1, Sonia Partap 1, Paul Fisher 1

Abstract

BACKGROUND: Surveillance magnetic resonance imaging (MRI) has become standard practice to detect tumor relapse. Whether spine MRI is efficacious to detect asymptomatic relapse of medulloblastoma (MB) is controversial. METHODS: We conducted an historical cohort study of children diagnosed with MB from January 2000 to December 2011 to analyze relapse presentation. Patients with at least one surveillance MRI were included. Relapse was defined as the appearance of a new lesion, significant increase (more than 25% in size) of a known previous lesion, or leptomeningeal disease. RESULTS: 44 patients (median age 7.3 yr [range = 0.9-21 yr], 31 males) were assessed. During the observation period (median 39.5 months [range 3-135 months] 478 surveillance brain MRIs and 382 surveillance spine MRIs were conducted. Relapses were identified on 23 brain MRIs, 10 brain and spine MRIs, and 4 spine MRIs for a total of 37 relapses in 17 patients. Relapse was significantly more frequent in patients with anaplastic/large cell MB (8 out of 11) when compared to classic MB (8 out of 26) (p = 0.03) and in patients classified as M3 (8 out of 12) when compared to M0 (8/27) (p = 0.041). Out of 37 relapses, only 11 relapses were associated with symptoms and only 2 MRIs were prompted because new acute onset symptoms. Among the 4 patients who presented with isolated spine relapse, 3 had high-risk MB (M3 staging at diagnosis) and one patient had average-risk MB. CONCLUSION: Our study demonstrates that most relapses are asymptomatic. Routine surveillance spine MRI does detect relapses, often with synchronous brain relapse. Isolated spinal relapses remain rare and mainly observed in patients at high risk. Frequency of routine spine MRIs could possibly be tailored to medulloblastoma subtype.

Neuro Oncol. 2012 Jun;14(Suppl 1):i140–i147.

RA-15. NON-INVASIVE MOLECULAR CHARACTERISATION OF MEDULLOBLASTOMAS IN CHILDREN

Simrandip K Gill 1, Martin Wilson 1, Nigel P Davies 1, Lesley MacPherson 2, Theodoros N Arvanitis 1, Andrew C Peet 1

Abstract

INTRODUCTION: Medulloblastomas (MBs) are highly aggressive tumours and not all patients with localised disease survive. Molecular markers from tumour tissue can define different prognostic groups and there is evidence that metabolism particularly related to glutamate could be important. This finding gives the opportunity to identify MB tumour subgroups non-invasively using magnetic resonance spectroscopy (MRS). In this study we investigate whether prognostic subgroups of M0 MB can be characterised by their MRS metabolite profiles. PATIENTS AND METHODS: Single voxel MRS (1.5T, TE 30ms, TR 1500ms) was performed on 17 children with Chang stage M0 MB prior to treatment; patients were followed-up for a median of 3 years, 6 had died at the end of the study period. MRS data was analysed using TARQUIN software to provide metabolite concentrations and a combined principal component and linear discriminant analysis (PCA-LDA) used to determine metabolite profiles for good and poor prognostic groups. Kaplan Meier curves were constructed comparing these groups. Pair-wise correlation analysis (Pearson's) was performed on the metabolite values for the two groups. RESULTS: Metabolite profiles were constructed, which discriminated well between survivors and non-survivors with 9/9 and 2/8 being alive in the good and poor prognosis groups respectively. There was a significant difference in Kaplan Meier survival between the two groups (log rank test, p = 0.03). The poor prognosis group had a higher tumour lipid/choline ratio (t-test, p = 0.03). Citrate levels were correlated with glutamate + glutamine levels (Pearson's correlation analysis p = 0.001) in the group that demonstrated a better prognosis but not in the poor prognosis group (p = 0.5). CONCLUSIONS: Non-invasive MRS metabolite profiles can be constructed that correspond to good and poor prognosis groups for M0 MB. The biomarkers identified concur with previous findings from in vitro molecular profiling of MB and MRS studies.

Neuro Oncol. 2012 Jun;14(Suppl 1):i140–i147.

RA-16. MRS SHOWS METABOLITE DIFFERENCES WITH AGE IN CHILDREN AND YOUNG PEOPLE WITH BRAIN TUMOURS

Nigel Davies 1, Simrandip Gill 2, Martin Wilson 2, Lesley MacPherson 3, Theodoros Arvanitis 2, Andrew Peet 2

Abstract

INTRODUCTION: Magnetic resonance spectroscopy (MRS) provides non-invasive metabolite profiles, useful for the clinical management of brain tumours. Studies demonstrate that the biology and behaviour of brain tumours differs between children and adults; however it is not known how this affects the metabolite profiles measured by MRS. We investigate whether MRS metabolite profiles vary with age in paediatric brain tumours. PATIENTS AND METHODS: Good quality short-TE MRS was acquired prior to treatment in 185 cases with subsequent diagnosis of a brain tumour confirmed by histopathology (N = 150) or imaging and follow-up (N = 35). Age at diagnosis ranged from 2 months to 16.5 years. LCModel and TARQUIN were used to analyse the MRS and estimate metabolite levels. The effect of age at diagnosis on metabolite and lipid / macromolecular (LMM) levels was investigated for the whole cohort, and for Medulloblastomas (MB), Pilocytic Astrocytomas (PA) and unbiopsied Optic Pathway Gliomas(OPG) separately, using Pearson's correlation analysis and ANOVA between age groups defined as 0-3 yrs, 4-11 yrs and 12-17 yrs. RESULTS: No significant correlations with age or differences between age groups were found for the whole cohort or for PAs (N = 39). In MBs (N = 32), phosphocholine was positively correlated with age (R = 0.5, P < 0.01). In OPGs (N = 11), creatine (Cr) (R = 0.93, P < 0.0001), glutamate (Glu) (R = 0.78, P < 0.01) and N-acetyl-aspartate (NAA) (R = 0.77, P < 0.01) were positively correlated with age and LMM at 0.9ppm was negatively correlated with age (R = -0.64, P < 0.05). Cr, Glu and NAA were significantly higher in older (10-17 yrs, N = 6) versus younger (0-3 yrs, N= 5) OPGs, with the opposite finding for LMM at 0.9ppm. CONCLUSIONS: Specific metabolic subtypes of brain tumours emerge at different ages through the childhood and teenage years. These effects should be considered when interpreting metabolite biomarkers for childhood brain tumours. Further investigations need to be performed together with tumour biological studies to determine their origin.

Neuro Oncol. 2012 Jun;14(Suppl 1):i140–i147.

RA-17. DROP METASTASES REVEALED WITH DIFFUSION-WEIGHTED IMAGING OF THE SPINE

Laura Hayes 1, Richard Jones 1, Claire Mazewski 2, Dolly Aguilera 2, Susan Palasis 1

Abstract

BACKGROUND: Historically, diffusion-weighted imaging (DWI) of the spine has been limited by excessive artifact. CSF and arterial pulsations, respiration, and surrounding bony structures produce this artifact thus limiting the evaluation of the contents of the thecal sac. Our study evaluates a new diffusion technique, readout-segmented echo-planar imaging (RS-EPI), for detecting drop metastases from hypercellular pediatric central nervous system (CNS) tumors. METHOD: DWI of the spine using RS-EPI was performed on 1.5T or 3T MR scanners. This sequence was added to standard conventional imaging protocols performed for evaluation for drop metastases in patients with hypercellular pediatric CNS tumors. 44 surveillance studies were. Scans were reviewed independently by two pediatric neuroradiologists. RESULTS: RS-EPI produced excellent image quality, nearly free from the artifacts plaguing other types of diffusion imaging of the spine. Traditional artifacts were drastically reduced, thereby allowing identification of hypercellular drop metastases. Drop metastases were identified in 5 patients with this technique where no lesions were identified on conventional T1- and T2-weighted imaging sequences on the same scan. The presence of drop metastases was subsequently proven on conventional sequences on follow-up scans and CSF analysis. Furthermore, RS- EPI was helpful in confirming absence of metastatic disease when conventional sequences were equivocal. CONCLUSIONS: (RS) EPI is a new DWI technique that appears to be helpful in identifying drop metastases from hypercellular CNS tumors in children. Our findings suggest that this technique warrants further evaluation in an effort to identify drop metastases as early as possible in an effort to improve these patient's quality of life.

Neuro Oncol. 2012 Jun;14(Suppl 1):i140–i147.

RA-18. DIFFUSION-WEIGHTED IMAGING AS A TOOL TO DIFFERENTIATE PEDIATRIC POSTERIOR FOSSA TUMORS

Anne Bendel 1, Richard Patterson 1, Joseph Petronio 1

Abstract

Several small studies have described diffusion-weighted imaging (DWI) as a reliable tool in differentiating posterior fossa tumors. Medulloblastoma typically restricts diffusion, whereas ependymoma and juvenile pilocytic astrocytoma (JPA) typically do not. This study was undertaken to further define the reliability of DWI in diagnosing pediatric posterior fossa tumors and the ability to determine restricted diffusion without the aid of Apparent Diffusion Coefficient mapping (ADC). METHODS: DWI of all posterior fossa tumors with confirmed histology diagnosed at our institution from 2004-present were retrospectively reviewed. ADC mapping was not available for the majority of cases, therefore restricted diffusion was recorded if visually present in > 50% of the solid portion of the tumor. Restricted diffusion was defined as signal seen on DWI (B = 1000 s/mm2) that is greater than the signal of the cerebellar cortex. A single neuroradiologist blinded to the tumor pathology reviewed the DWI of each case. RESULTS: A total of 99 non-brainstem posterior fossa tumors were reviewed (47 JPA, 35 medulloblastoma, 8 ependymoma, 4 grade II astrocytoma, 2 ATRT, 1 GBM, 1 anaplastic glioneuronal tumor, 1 atypical choroid plexus papilloma (ACPP)). DWI showed restricted diffusion in 1/47 JPA (2%), 34/35 medulloblastoma (97%), 1/8 ependymoma (12.5%), 0/4 grade II astrocytoma (0%), 2/2 ATRT (100%), 0/1 GBM (0%), 1/1 anaplastic glioneuronal tumor (100%), 0/1 ACPP (0%). The JPA and ependymoma cases with restricted diffusion showed less conspicuous signal intensity than the signal seen in most medulloblastomas and ATRTs. Five medulloblastoma patients with metastatic nodules demonstrated restricted diffusion within the metastatic foci. CONCLUSION: Diffusion-weighted imaging is very useful in differentiating pediatric posterior fossa tumors, with restricted diffusion being rare in JPA and ependymoma and common in medulloblastoma and ATRT. DWI should be used in conjunction with other MRI features to aid in the diagnosis of pediatric posterior fossa tumors.

Neuro Oncol. 2012 Jun;14(Suppl 1):i140–i147.

RA-19. A NOVEL RADIOLOGICAL RESPONSE CRITERIA FOR EVALUATING LEPTOMENINGEAL METASTASIS

Lisethe Meijer 1, Tim Jaspan 2, Richard GG Grundy 1, David A Walker 1

Abstract

INTRODUCTION: Current tumour response criteria lack a standardised method for assessing leptomeningeal metastasis (LMM). We propose an anatomy-based system for classification of LMM response combining qualitative and quantitative criteria. Anatomical regional surfaces: 5 are specified: cerebral hemispheric with suprasellar cistern, ventricular cavities, brainstem, cerebellar, spinal cord and theca. Quantitative criteria: maximum lesional diameter per region. Qualitative criteria: nodularity, linearity and confluence. Responses are defined as: complete response (CR): disappearance of all evidence of LMM; partial response (PR): reduction in regions involved, >25% reduction in diameter or qualitative regression; stable disease (SD): no change in regions involved, < 25% dimensional change, no qualitative change; mixed response (MR): combination of two or more of other criteria; progressive disease (PD): increase in regions involved, >25% dimensional progression, qualitative progression. PATIENTS AND METHODS: These criteria were applied to MRI scans of 11/12 patients receiving intraventricular/intrathecal (InV/IT) chemotherapy for LMM, 1 patient could not be assessed for response due to initial scan being a MR myelogram. The scans were interpreted by a single neuro-radiologist (TJ). In the absence of alternative criteria, we compared LMM response criteria with MRI and clinical reports. RESULTS: CR: 1; PR: 1; SD: 0; MR: 5; PD: 4. Response assessment was compared with MRI and clinical reports. CR and PR cases responded likewise on MRI and clinically. MR cases: all 5 progressed on MRI and clinically. PD cases: 2 had transient clinical improvement, 1 had also MR report of improvement. Both had progressive disease on MRI and clinically very shortly thereafter. 11/12 patient died; 10 died from progressive disease, and 1 of toxicity post stem-cell transplant. CONCLUSION: LMM criteria demonstrated face validity, were applicable and were descriptive of subtle imaging changes of response and progression. This classification will be applied in an anticipated study of InV/IT therapy in LMM disease.

Neuro Oncol. 2012 Jun;14(Suppl 1):i140–i147.

RA-20. 18F-FLT FOR PET IMAGING OF BRAIN TUMORS IN CHILDREN

Nathan Robison 1, Frederick Grant 2, S Ted Treves 2, Pratiti Bandopadhayay 1, Peter Manley 1, Susan Chi 1, Mary Ann Zimmerman 1, Christine Chordas 1, Liliana Goumnerova 2, Edward Smith 2, Michael Scott 2, Nicole J Ullrich 2, Tina Poussaint 2, Mark Kieran 1

Abstract

BACKGROUND: Molecular imaging using positron emission tomography (PET) is a potentially valuable tool in the evaluation of pediatric brain tumors. However, utility of standard [(18)F]fluorodeoxyglucose (FDG)-PET for imaging of brain tumors is limited by high glucose utilization and thus high tracer uptake by normal brain. The radiolabeled nucleoside 3'-deoxy-3'-[(18)F]fluorothymidine (FLT), in contrast, is taken up primarily by cells undergoing active DNA synthesis. FLT-PET may thus differentiate neoplastic from normal tissue and high-grade from low-grade tumors with greater specificity than FDG-PET. METHODS: In two sequential prospective IRB-approved studies, seven pediatric patients with suspected brain tumors underwent preoperative imaging with FLT-PET. Imaging findings were correlated with tumor histopathology. RESULTS: No uptake was seen in two patients, both with low grade glioma: histologically, one was indolent-appearing without distinguishing features, and the other was a diffuse astrocytoma. Mild uptake was seen in two additional patients, both with juvenile pilocytic astrocytoma. One patient with an FLT-avid lesion was found to have non-Hodgkins B-cell lymphoma of the central nervous system. On another subject, whose tumor showed mild uptake overall but had an area of more intense FLT avidity, histology showed an unusual anaplastic embryonal neoplasm arising in a background of low grade astrocytoma. Finally a patient with two separate lesions, one with mild and the other intense FLT avidity, was found to have a non-neoplastic autoimmune lymphoproliferative disorder. In all cases, no uptake was seen in normal brain. All patients tolerated the isotope administration well, without adverse event. CONCLUSION: FLT-PET is a promising modality for molecular imaging evaluation of brain tumors in children.

Neuro Oncol. 2012 Jun;14(Suppl 1):i140–i147.

RA-21. PHANTOSMIA DURING RADIATION THERAPY: A REPORT OF 2 CASES

Joanna C Yang 1, Donita D Lightner 1, Yasmin Khakoo 1, Suzanne L Wolden 1

Abstract

INTRODUCTION: Phantosmia is an infrequently reported and poorly understood qualitative olfactory disorder characterized by the perception of a frequently unpleasant odor in the absence of an odorant stimulus. To the best of our knowledge, these are the first descriptions of phantosmia during focal or whole-brain radiation therapy (RT). CASE DESCRIPTIONS: A 6-year-old with diffuse intrinsic pontine glioma receiving focal intensity-modulated radiotherapy to the pons reported an intermittent odor resembling halitosis at day 22 of RT. His phantosmia made it difficult for him to comply with daily RT and resolved completely after RT completion. A 15-year-old with gliomatosis cerebri and WHO grade II astrocytoma in the bifrontal lobes undergoing RT with extensive fields to encompass his diffuse disease reported a similar odor at day 2 of RT. He noted a flash of "blue light" accompanied the odor. His phantosmia caused him significant distress as it was intermittently present between RT sessions, as well, and decreased his quality of life during treatment. Since RT completion, his phantosmia has slowly resolved and he no longer experiences symptoms. DISCUSSION: Two patients with incomparable tumors and radiation fields both experienced phantosmia featuring a halitosis-like odor during their courses of RT. Peripheral phantosmia is hypothesized to involve abnormally active olfactory receptor neurons while central phantosmia is theorized to be due to hyperactive neurons of the cortex. It is possible the phantosmia in the 6-year-old was central, caused by a small amount of dose to a portion of the olfactory cortex during pontine irradiation. In the 15-year-old, a mix of central and peripheral phantosmia is likely. The extensive RT fields necessary to treat his disease encompassed the olfactory cortex, but the "blue light" likely indicated activation of his retinal photoreceptors, thus making concurrent activation of his peripheral olfactory receptors probable, as well.

Neuro Oncol. 2012 Jun;14(Suppl 1):i140–i147.

RA-22. PAEDIATRIC BRAIN STEM GLIOMAS (BSG) - THE IMPORTANCE OF RADIOTHERAPY

Robert Smee 1, Cathy Zhao 1

Abstract

AIM: To evaluate the outcome of a group of children / adolescents treated by radiotherapy in recent years. MATERIALS AND METHODS: For this Ethics approved review all children / adolescents (age <18 years) were evaluated, the defining feature being clinical and imaging features consistent with a BSG, and commenced radiotherapy. Characteristics were grouped into patient, disease, and treatment factors with the primary end point being overall survival, plus secondary end point of progression-free survival. Progression-free survival was defined as stable or improved clinical / radiological status, with no episode of deterioration. Kaplan-Meier analysis was used to describe time to event data. Two and 5-year survival rates are reported. RESULTS: Between 1990-2007 (thus all patients had MR assessment) 23 patients were treated, 3 palliatively (with 1 patient receiving only 1 fraction) and 20 curatively (median dose 54Gy x 30 fractions, 17 stereotactically.) There were 13 females and 10 males with 18 having symptoms greater than 1 month. No tumour was focal in nature, with extension superiorly to the thalamus in 2 patients, and inferiorly into the cervical cord in 4 patients. Overall survival was 43.5% (medial survival 11.86 months), with patients aged over 8 years having a 5-year survival of 57%, for those 8 or less it was 38%. Progression-free survival for all was 61%. CONCLUSION: Whilst having a widespread poor outlook, there is some optimism from this series at least to suggest that reasonable outcomes are still possible.

Neuro Oncol. 2012 Jun;14(Suppl 1):i140–i147.

RA-23. UTILITY OF ROTATIONAL ARC RADIOTHERAPY IN WHOLE VENTRICULAR RADIATION WITH BRAIN SPARING FOR CNS GERMINOMA PATIENTS

Brooke Spencer-Trotter 1, Abhirami Hallock 1, Andre Konski 1, Kanta Bhambani 2

Abstract

PURPOSE: The radiation treatment for pediatric CNS germinomas has changed from craniospinal radiation to whole ventricular radiation given the concerns of late neurocognitve sequelae. There are no comparative trials looking at dosimetric differences of helical tomotherapy (HT) to fixed field linear accelerator based therapy inversely planned IMRT (fIMRT) for whole ventricular radiation. We hypothesized that HT would provide superior organ at risk (OAR) sparing and target conformality, given the rotational delivery and increased degrees of freedom for beam modulation. MATERIALS/METHODS: Three previously treated pediatric germinoma cases were replanned using HT and inversely planned 5 field IMRT to deliver a D95PTV of 24 Gy as per the COG/ACNS 0232. Dmean, Dmin, Dmax, to PTV, Dmean to the OARs, (brain, cochlea, chiasm, optic nerves, hippocampi). Confomality Index (prescribed dose/target volume), Heterogeneity Index (max dose/prescribed dose), and Integral Dose (ID) or mean dose x volume, were used for dosimetric analysis. RESULTS: The mean ventricular target volume (PTV) was 491 cm3. HT plan yielded an average Dmean= 24.3 Gy +/−0.37, Dmin= 15.3 Gy, Dmax = 26.3 Gy to PTV; dose to Brain - PTV Dmean = 18 Gy, hippocampi Dmean = 24.2 Gy cochlea Dmean = 19.5 Gy, chiasm Dmean = 24.23 Gy, optic nerve Dmean = 9.15 Gy. CONCLUSION: The fIMRT plan yielded an average Dmean = 24.5 Gy +/−0.33, Dmin = 23.2 Gy, Dmax = 25.8 Gy to PTV, dose to Brain - PTV Dmean = 17.6 Gy, hippocampi Dmean = 24.5 Gy, cochlea Dmean = 16.9 Gy, chiasm Dmean= 24.5 Gy, optic nerve Dmean= 10.96 Gy. The CI, HI and ID for HT and fIMRT are 1.34 and 1.49, 1.09 and 1.1, 14.71 and 14.37 respectively. Our analysis suggests HT does not confer significant advantages in terms of PTV coverage, OAR sparing, dose homogeneity or integral dose for whole ventricular radiation compared to fIMRT.

Neuro Oncol. 2012 Jun;14(Suppl 1):i140–i147.

RA-24. PSEUDOPROGRESSION AFTER RADIOTHERAPY FOR LOW GRADE GLIOMA IN PEDIATRIC PATIENTS

Anita Mahajan 1, Jeremy Jones 2, Leena Ketonen 1, Arnold Paulino 3, Joann Ater 1, David Grosshans 1, Robert Dauser 2, Jeffrey Weinberg 1, Murali Chintagumpala 2

Abstract

PURPOSE: In this study we reviewed the MRI changes in pediatric patients with low-grade glioma (LGG) treated with proton radiotherapy (PRT) or x-ray based IMRT (IMXT). METHODS: All pediatric patients with LGG treated with RT with available pre and post-RT MRI imaging were identified. The medical and RT records reviewed. Two neuroradiologists, blinded to RT modality, evaluated the tumor size pre and post-RT. RESULTS: 66 patients (27 girls, 39 boys) with LGG diagnosed between 2001 and 2011 were identified. RT (31 IMXT, 35 PRT) was delivered for either progressive or symptomatic disease at a median age of 8.7 y (range 3-17.2 y). 44, 15, and 6 patients had pilocytic astrocytoma/optic nerve glioma, grade II astrocytoma, and other LGG's, respectively. 55 tumors were midline. 25 patients had pre-RT chemotherapy. The median RT dose was 50.4 Gy/GyE (range 45-54 Gy/GyE). 22/66 (33%) patients had tumor enlargement between 1.5-6 mo post-RT with a 16% median peak increase (range 5-259%) when treated with either PRT or IMXT. Of these: 18 tumors eventually decreased in size and stabilized, 2 were resected at 5 mo, 1 had progressive enlargement, and 1 was lost to follow up. At last follow up of 66 patients, the median tumor size decreased to 51% (range 0-196%) of the pre-RT size. 4/66 (6%) patients have had definite progression on follow up. 1 patient died. No significant difference has been noted in studied outcomes between the IMXT and PRT patients. CONCLUSIONS: Early tumor enlargement was noted in 33% of patients with LGG within 6 mo post-RT with decrease and stabilization over the following year. The RT modality does not appear to affect the incidence of this finding. Pseudoprogression should be recognized and intervention should be delayed unless the patient becomes symptomatic.

Neuro Oncol. 2012 Jun;14(Suppl 1):i140–i147.

RA-25. STEROID USAGE AMONG CHILDREN RECEIVING CRANIOSPINAL IRRADIATION (CSI) MAY MANDATE THE CASTING OF MULTIPLE IMMOBILIZATION MASKS AND RE-SIMULATION

Rina Dvir 1, Ronit Elhasid 1, Benjamin Corn 2, Haim Tempelhoff 1, Diana Matceyevsky 1, Veronica Makrin 1, Natan Shtraus 1, Dalia Yavetz 1, Shlomi Constantini 2, Eliahu Gez 1

Abstract

INTRODUCTION: Delivery of radiotherapy to children with malignancy is complex and mandates cooperation between pediatric oncologists, anesthesiologists and radiation oncologists. Steroids use may engender edema/swelling which necessitates casting of additional immobilization masks after the first simulation and re-simulation (which may include additional anesthesia sessions). AIMS: To evaluate the impact of steroids on re-simulation during radiotherapy process of CSI in children. Materials and Methods: This retrospective study was performed at Tel Aviv Medical Center 2009-2011. 24 children were treated. Median age was 8.5 years (range 2-15). Primary tumor : Medulloblastoma 7, Diffuse Pontine Glioma 9, Ependymoma 3, Craniopharyngioma 3, Glioblastoma 1, PNET 1 patient.15 patients(pts) underwent surgical procedures prior to radiotherapy. 17 pts received local field irradiation and 7 CSI. The radiotherapy technique was either IMRT or 3DCRT. Seven children were treated with daily anesthesia. Twenty (83%) children received steroid treatment with dexamethasone. RESULTS: 17 children (70%) underwent re-simulation of which 16 required a new face mask (2 required 2 re-simulations with new masks). All 16 patients were treated with steroids. Of 7 other patients who were not re-simulated, only 4 received steroids. 11 patients started steroids before radiotherapy and 9 during RT. CONCLUSIONS: Steroids are widely used during CSI to alleviate symptoms such as headache and vomiting.. The weight gain and changes in body and facial habitus attributed to steroid use mandated re-simulation with new masks in two thirds of children regardless of anesthesia use, age or diagnosis. No changes in face masks were needed in children who did not receive steroids. Treatment with steroids during radiotherapy has a significant impact on the technical approach to patients requiring cranial radiation and may have attendant consequences such as the need for additional anesthesia and treatment delays. Judicious use of steroids is therefore critical in this population.

Neuro Oncol. 2012 Jun;14(Suppl 1):i140–i147.

RA-26. INFLUENCE OF PROTON BEAM THERAPY ON COGNITIVE FUNCTIONING IN CHILDREN WITH BRAIN TUMORS: PRELIMINARY RESULTS OF A PROSPECTIVE LONGITUDINAL STUDY

Eun-Seung Yu 1, Yeon-Joo Kim 2, Hyeon Jin Park 3, Ho Jin Kim 4, Sang Hoon Shin 3, Jong-Heun Kim 1, Joo-Young Kim 2

Abstract

PURPOSE: Children with brain tumors show changes in neurocognitive function with various reasons. This study was performed to examine any possible neurocognitive changes which may occur after proton beam radiotherapy (PBT) in children with brain tumors. METHODS: We designed a prospective longitudinal study which compared pre-radiation neurocognitive function to post-radiation. Thirty-three patients diagnosed with brain tumors undertook neurocognitive tests before PBT, and 10 of those were subject to follow up test at 1 year post-PBT. The mean age at the time of radiation was 11. 5 years (range 6 to 19, SD = 3.78). 78.8% received chemotherapy (n = 26); 75.8% operation (n = 25); 57.6% only proton therapy (n = 19). The mean duration of radiation was 43.1 days (range 22 to 59, SD = 9.75) and Radiation dose to the primary site ranged 30-76 Gy/17-38 fractions. Administered tests were Korean version of Wechsler Intelligence Scale for Children-3rd Edition, Rey-Kim Memory Test, and Kim's Frontal-Executive Function Test. All test scores were converted to standard scores based on age-matched standardization norms. RESULTS: At the time of pre-radiation evaluation (n = 33), the mean of the full scale intellectual Quotient (p = .007), Performance IQ (p = .000), Perceptual Organization Index (p = .021), Freedom from Distractibility Index (p = .003), Processing Speed Index (p = .000), and Memory Quotient (p = .002) were significantly lower than normative mean. These basal neurocognitive functions didn't show significant differences in various treatment conditions -operation, V-P shunt, type of operation, chemotherapy (p > .050 for all). Comparing measures of pre-PBT to post-PBT evaluation (n = 10), there were no significant changes in all indices (p > .005 for all). CONCLUSIONS: These results show that most neurocognitive function, especially perceptual organization, attention, processing speed, memory, and executive functioning is impaired in children with brain tumors even before PBT was given. With relatively short follow up period of 1year, neurocognitive function does not seem to change significantly by PBT.

Neuro Oncol. 2012 Jun;14(Suppl 1):i140–i147.

RA-27. REDUCING DOSE TO ORGANS-AT-RISK IN RADIOTHERAPY OF BRAIN TUMORS ARISING IN POSTERIOR FOSSA

Young K Lee 1, Maria R Fiore 2, Chantal Sanne 3, Henry C Mandeville 4, Frank H Saran 4

Abstract

Delivering a radical radiation dose of 54-60Gy to posterior fossa brain tumours can produce significant late effects due to the proximity of the cochlea and other organs-at-risk (OAR). The aim of this study is to reduce the OAR dose by using intensity-modulated radiotherapy (IMRT) and volumetric modulated arc therapy (VMAT) for posterior fossa tumours. 10 (medulloblastoma = 5, ependymoma = 5) patients treated with conventional radiotherapy (CRT) were retrospectively identified. Clinical target volume (CTV) was defined as the tumor bed and a uniform 10mm margin was used to create a planning target volume (PTV). Proximal OAR were delineated and planning-organ-at-risk-volume (PRV) created for cochlea and optic chiasm with a uniform 3mm margin. Medulloblastoma prescription dose was 30.6Gy/17# assuming a craniospinal dose of 23.4Gy/13# was delivered. Ependymoma prescription dose was 59.4Gy/33#. Using Philips Pinnacle3 planning system, IMRT and VMAT plans were inversely-planned with to achieve a uniform dose to the PTV and meet OAR constraints. OAR dose-volume-histograms and PTV coverage and heterogeneity in CRT, IMRT and VMAT plans were compared. PTV coverage and heterogeneity were comparable for all plans. Large reductions in mean dose were observed to the right/left cochlear PRV from CRT(median 12.2/15.7Gy) to IMRT(8.3/11.5Gy) and to VMAT(6.8/8.4Gy) for medulloblastoma patients. Smaller cochlea PRV dose reductions were observed for ependymoma patients. ∼40% and ∼60% reductions in mean parotid dose was observed with IMRT and VMAT respectively when compared to CRT for medulloblastoma patients. For ependymoma patients, parotid doses were within acceptable clinical thresholds for all plans. Doses to other OAR for all plans were within the dose constraints. IMRT and VMAT can reduce OAR doses, compared to CRT without compromising the dose to PTV. This reduction was more marked in medulloblastoma patients receiving craniospinal irradiation than in ependymoma patients receiving localised radiotherapy only. The OAR dose reduction was greater using VMAT compared to IMRT.

Neuro Oncol. 2012 Jun;14(Suppl 1):i140–i147.

RA-28. FRACTIONATED STEREOTACTIC RADIOSURGERY FOR RELAPSED PAEDIATRIC PRIMITIVE NEUROECTODERMAL TUMOURS: A SINGLE INSTITUTION EXPERIENCE

Jeffrey Greenspoon 1, JoAnn Duckworth 2, Sheila Singh 2, Katrin Scheinemann 2, Anthony Whitton 1

Abstract

INTRODUCTION: Inoperable local tumour recurrence following aggressive radiochemotherapy can present in many paediatric tumours of the central nervous system (CNS). Despite advances in systemic therapy, bulky local disease can be very difficult to control. Radiosurgery has the potential to treat localized lesions with high dose radiotherapy while minimizing the exposure to previously treated normal tissues. METHODS: We have performed fractionated stereotactic radiosurgery (FSRS) on two patients with locally recurrent tumours of the CNS. Patient one presented initially with a supratentorial PNET and patient two presented with a high risk medulloblastoma. Both had received previous full dose cranial-spinal and boost external beam radiotherapy as well as chemotherapy. THe lesions were felt to be resectable by the neurosurgeons. All lesions were treated with robotic radiosurgery which requires minimal immobilization. Each lesion was treated to a dose of 25Gy in 5 fractions. Dose constraints for critical structures were kept within the published limits for FSRS without adjustment for previous radiotherapy. RESULTS: Each fraction of radiosurgery was delivered in approximately 40 minutes. Both patients tolerated the procedure with no procedural based toxicity. Using the CTCAE v 4.0 no acute or delayed radiation toxicity has been encountered. Both patients experienced a good imaging and clinical response to FSRS. In patient one the lesions remained controlled for 4 months at which time there was tumour progression at the margin of the radiosurgery field and the patient died. Patient two has had a good imaging and clinical response at 4 months with no evidence of local recurrence. CONCLUSION: FSRS is a treatment option for locally recurrent paediatric CNS tumours. Although we have demonstrated good safety and efficacy in our two cases, further elucidation of its indications, dosing and toxicity through a prospective protocol is warranted.

Neuro Oncol. 2012 Jun;14(Suppl 1):i140–i147.

RA-29. BEVACIZUMAB DURING RADIATION THERAPY DECREASES TUMOR SWELLING AND IMPROVES CLINICAL SYMPTOMS: CASE REPORTS

Karen Gauvain 1, Thomas Geller 1, Samer Elbabaa 1, John Dombrowski 1

Abstract

Standard therapy for diffuse intrinsic pontine glioma (DIPG) is radiation therapy (RT) with or without chemotherapy. RT to high grade lesions may cause rapid swelling and necrosis during or several months following the treatment. We describe three cases of DIPG who were treated with bevacizumab during RT with marked clinical improvement. Bevacizumab was well tolerated without complications. CASE 1. An 8 girl presented with right hemiparesis and ataxic gait. MRI confirmed DIPG. Dexamethasone was started and RT began. 2 weeks into RT, she developed rapid neurologic deterioration despite a slow steroid wean. CT scan showed increased necrosis and surrounding pontine edema. Dexamethasone was restarted with mild clinical improvement. She failed an attempt to wean steroids and was then started on bevacizumab 10 mg/kg every 2 weeks. She had rapid clinical improvement following initiation of bevacizumab and was able to tolerate a steroid wean. The patient continued to do well until she had slow clinical progression 7 months later. CASE 2. A 5 y/o presented with fatigue, dizziness and left esotropia. MRI confirmed DIPG with marked hydrocephalus. Dexamethasone was started and VP shunt placed with improvement in symptoms. One week into RT, she developed progressive symptoms after an attempt to slowly wean steroids so bevacizumab was started. She had clinical improvement within one week and was able to tolerate a steroid wean. She continues on bevacizumab and remains asymptomatic 7 months following. CASE 3. A 6 y/o presented with bilateral nystagmus and ataxic gait. MRI confirmed DIPG. She was discharged on dexamethasone following clinical improvement. A week following she returned with marked neurologic deterioration and emergent radiation began. Bevacizumab began one week into RT. Her neurologic exam normalized 3 weeks after starting RT and a steroid wean was tolerated. She continues on bevacizumab and remains asymptomatic 6 months following.

Neuro Oncol. 2012 Jun;14(Suppl 1):i140–i147.

RA-30. RADIOTHERAPY FOR YOUNG CHILDREN AFTER HIGH DOSE CHEMOTHERAPY WITH AUTOLOGOUS HEMATOPOIETIC CELL RESCUE: QUALITY ASSESSMENT OF HEAD START III

Kenneth Wong 1, Arthur Olch 1, Tom Belle Davidson 2, Rajkumar Venkatramani 1, Kelley Haley 1, Wafik Zaky 1, Girish Dhall 1, Jonathan Finlay 1

Abstract

BACKGROUND: The use of high dose chemotherapy with autologous hematopoietic cell rescue (AuHCR) in Head Start III is a potentially curative approach to the management of young children with newly diagnosed CNS neoplasms, with the additional benefit of reducing or avoiding radiation therapy. We report the potential influence of timing and quality of radiotherapy on survival of patients. METHODS: Between April 2003 and December 2009, 42 children were irradiated at 16 different institutions. All children were to receive 5 induction cycles of chemotherapy followed by 1 consolidation cycle of myeloablative chemotherapy and AuHCR. Only children between 6-10 years old or with residual tumor pre-consolidation were to receive radiotherapy at least 42 days after consolidation. Radiotherapy records were reviewed to determine adherence to protocol guidelines for treatment volume and dose prescription. Of 42 patients, 7 were irradiated prior to consolidation (six for progression during induction and one refused consolidation), and an additional 5 patients without tumor pre-consolidation were treated at relapse. Overall survival was calculated. A log-rank test was used to determine the significance of potential survival differences between patients with and without major deviations in radiotherapy. RESULTS: Of the 30 patients who were fully evaluable for all treatment quality parameters, 14 (47%) had 1 or more deviations in their treatment schedule, dose, or treatment volume. Most deviations (78%) occurred in the less than 6 age group, and only 2 dose deviations involved higher than protocol dose. Major treatment volume or total dose deviations did not significantly influence overall survival. Time to start radiotherapy of greater than 70 days from AuHCR negatively impacted overall survival. CONCLUSION: Despite major treatment deviations in half of fully evaluable patients, underdosage was not associated with a worse overall survival. When indicated, starting radiotherapy soon after bone marrow recovery may improve the outcome for high-risk children.

Neuro Oncol. 2012 Jun;14(Suppl 1):i140–i147.

RA-31. RADIATION INJURY IN PEDIATRIC PATIENTS WITH CNS TUMORS TREATED WITH PROTON BEAM RADIATION THERAPY: A CASE SERIES

Michael W Bishop 1, Trent R Hummel 1, James Leach 1, Jane Minturn 2, John Breneman 1, Charles Stevenson 1, Lars Wagner 1, Mary Sutton 1, Lili Miles 1, Maryam Fouladi 1, Stewart Goldman 3

Abstract

INTRODUCTION: Radiation injury (RI) may result from injury to oligodendrocytes and vascular endothelial cells, with local cytokine release, vasogenic edema, and demyelination. RI occurs in a subset of children with brain tumors following conventional radiation. Proton beam therapy (PBT) is being used increasingly frequently in pediatrics to attenuate the cumulative dose and damage to adjacent structures, and to minimize long-term radiation-related sequelae. Little data exist regarding incidence and course of RI in children receiving PBT. METHODS: We performed a retrospective chart review to describe the clinical and radiographic characteristics and course of children who developed RI post PBT for treatment of CNS tumors at three pediatric tertiary hospitals. RI was diagnosed based on MRI findings of increased T2 and FLAIR signal to surrounding structures ± clinical symptoms. RESULTS: Eight cases, median age 6 years (range 23 months - 27 years), were identified among 132 patients who received PBT from 1995 to 2012. Diagnoses included medulloblastoma (3), ependymoma (2), AT/RT, anaplastic oligodendroglioma, and optic glioma. PBT was administered as part of upfront therapy in six patients (1 preceding high dose chemotherapy, 3 with platinum-based chemotherapy), and at recurrence in two. Median cumulative dose was 54.1 cobalt grey equivalent (range 30.6-59.4 Gy). Median time to RI onset following PBT was 142 days (range 50-234 days). Treatments included corticosteroids (6/8), bevacizumab (4/8), and hyperbaric oxygen (3/8). Rapid radiographic response and symptomatic improvement was seen in 3/4 patients receiving bevacizumab. Four patients experienced a secondary increase in signal without enhancement at 8 to 13 months post-PBT. Among 4 who remain alive (median follow-up24 months), two have persistent radiation-associated neurologic deficits. CONCLUSION: RI developed in 6% of children following PBT. Patients may initially respond to steroids and bevacizumab; later recurrence of RI can occur and is less responsive to therapy.

Articles from Neuro-Oncology are provided here courtesy of Society for Neuro-Oncology and Oxford University Press

RESOURCES