Interpretation of magnetic resonance images in diffuse intrinsic pontine glioma: a survey of pediatric neurosurgeons

Abstract

The current management paradigm for children with newly diagnosed diffuse intrinsic pontine glioma (DIPG) is to establish a diagnosis and begin therapy based on MR imaging findings correlated with an appropriate clinical presentation, and without a tissue diagnosis. This strategy assumes that pediatric neurosurgeons and neurooncologists uniformly interpret MR imaging findings in this population. This study sought to examine the consistency of North American pediatric neurosurgeons in assessing MR images in this patient population, and in their surgical plans based on the interpretation of those images. The authors created an online survey and invited all members of the American Society of Pediatric Neurosurgeons and the American Association of Neurological Surgeons/Congress of Neurological Surgeons Joint Section for Pediatric Neurosurgery to participate. The survey consisted of 58 questions, and 48 pertained to representative MR images from 16 children who presented to The Children's Hospital of Colorado with diffuse pontine tumors. Based on the imaging presented and a standard clinical scenario, the respondent was asked if she or he believed a lesion to be "typical" or "atypical," whether she or he would biopsy the lesion, and what surgical approach would be chosen. The remaining 10 questions pertained to respondent demographics and his or her practice regarding tissue preservation and interest in participating in a multicenter trial that included tumor biopsy in selected cases. Rates at which each lesion was considered to be typical or atypical and rates of recommended biopsy were calculated. Surveys were received by 269 individuals. Eighty-six responses were received (32.0%). No tumor was judged to be either typical or atypical by all respondents. Individual surgeons varied widely regarding how many of the tumors were judged as typical or warranted a biopsy. The percentage of respondents who disagreed with the majority opinion regarding whether a tumor qualified as typical ranged from 2.3% to 48.8%, with a median of 28.6%. More than 75% agreement regarding whether a tumor was typical or atypical was found in 7 (43.8%) of 16 cases. The κ statistic regarding typicality was 0.297 ± 0.0004 (mean ± SEM), implying only fair agreement. For every tumor, at least 1 respondent who believed the lesion to be typical in appearance would nevertheless biopsy the lesion. Of those respondents who considered a lesion to be typical, a median of 5.1% (range 1.2%-66.7%) would choose to biopsy this lesion. Of those respondents who considered a lesion atypical, a median of 18.3% (range 3.7%-100%) would choose not to biopsy the lesion. Of 85 responses to the question, "Would you be willing to biopsy a typical diffuse pontine glioma as a part of a multicenter trial?," 59 (69.4%) of 85 respondents answered that they would. Although making a diagnosis based on radiographic evidence alone represents a well-established management paradigm for children with suspected DIPG, this study demonstrates considerable inconsistency on the part of pediatric neurosurgeons in the application of this strategy to individual patients. As such, the practice of diagnosing DIPG based on imaging characteristics and clinical history alone does not reach the appropriate threshold to be considered a standard of care.