@article{a2bc38bb1730482cb616bbdfc068421b,
title = "MRI Radiogenomics of Pediatric Medulloblastoma: A Multicenter Study",
abstract = "Background: Radiogenomics of pediatric medulloblastoma (MB) offers an opportunity for MB risk stratification, which may aid therapeutic decision making, family counseling, and selection of patient groups suitable for targeted genetic analysis. Purpose: To develop machine learning strategies that identify the four clinically significant MB molecular subgroups. Materials and Methods: In this retrospective study, consecutive pediatric patients with newly diagnosed MB at MRI at 12 international pediatric sites between July 1997 and May 2020 were identified. There were 1800 features extracted from T2- and contrast-enhanced T1-weighted preoperative MRI scans. A two-stage sequential classifier was designed-one that first identifies non-wingless (WNT) and non-sonic hedgehog (SHH) MB and then differentiates therapeutically relevant WNT from SHH. Further, a classifier that distinguishes high-risk group 3 from group 4 MB was developed. An independent, binary subgroup analysis was conducted to uncover radiomics features unique to infantile versus childhood SHH subgroups. The best-performing models from six candidate classifiers were selected, and performance was measured on holdout test sets. CIs were obtained by bootstrapping the test sets for 2000 random samples. Model accuracy score was compared with the no-information rate using the Wald test. Results: The study cohort comprised 263 patients (mean age 6 SD at diagnosis, 87 months 6 60; 166 boys). A two-stage classifier outperformed a single-stage multiclass classifier. The combined, sequential classifier achieved a microaveraged F1 score of 88% and a binary F1 score of 95% specifically for WNT. A group 3 versus group 4 classifier achieved an area under the receiver operating characteristic curve of 98%. Of the Image Biomarker Standardization Initiative features, texture and first-order intensity features were most contributory across the molecular subgroups. Conclusion: An MRI-based machine learning decision path allowed identification of the four clinically relevant molecular pediatric medulloblastoma subgroups.",
author = "Michael Zhang and Wong, {Samuel W.} and Wright, {Jason N.} and Wagner, {Matthias W.} and Sebastian Toescu and Michelle Han and Tam, {Lydia T.} and Quan Zhou and Ahmadian, {Saman S.} and Katie Shpanskaya and Seth Lummus and Hollie Lai and Azam Eghbal and Alireza Radmanesh and Jordan Nemelka and Stephen Harward and Michael Malinzak and Suzanne Laughlin and S{\'e}bastien Perreault and Braun, {Kristina R.M.} and Lober, {Robert M.} and Cho, {Yoon Jae} and Birgit Ertl-Wagner and Ho, {Chang Y.} and Kshitij Mankad and Hannes Vogel and Cheshier, {Samuel H.} and Jacques, {Thomas S.} and Kristian Aquilina and Fisher, {Paul G.} and Michael Taylor and Tina Poussaint and Vitanza, {Nicholas A.} and Grant, {Gerald A.} and Stefan Pfister and Eric Thompson and Alok Jaju and Vijay Ramaswamy and Yeom, {Kristen W.}",
note = "Funding Information: No relevant relationships. J.N.W. No relevant relationships. M.W.W. No relevant relationships. S.T. No relevant relationships. M.H. No relevant relationships. L.T. No relevant relationships. Q.Z. No relevant relationships. S.S.A. No relevant relationships. K.S. No relevant relationships. S. Lummus No relevant relationships. H.L. No relevant relationships. A.E. No relevant relationships. A.R. No relevant relationships. J.N. No relevant relationships. S.H. No relevant relationships. M.M. No relevant relationships. S. Laughlin No relevant relationships. S. Perreault No relevant relationships. K.R.M.B. No relevant relationships. R.M.L. No relevant relationships. Y.J.C. Hyundai Hope on Wheels research grant from the Ericksen Family Endowed Professorship. B.E.W. Deputy editor of Radiology; spouse owns stock in Siemens; spouse is employee of Siemens Healthineers. C.Y.H. No relevant relationships. K.M. European Course in Paediatric Neuroradiology lecture honorarium; Guerbet lecturer honorarium; Siemens lecturer honorarium; payment for expert medicolegal reports in the United Kingdom. H.V. Payment for deposition testimony from Miller Weisbrod; participant on a DataSafety Monitoring Board or Advisory Board for Stanford Neuropathy. S.H.C. No relevant relationships. T.S.J. Payment to institution from Cancer Research UK, The Brain Tumour Charity, Children with Cancer UK, NIHR, Great Ormond Street Hospital Children{\textquoteright}s Charity, Olivia Hodson Cancer Fund; royalties for book publishing from Elsevier; honoraria for editing from Wiley; payment for lecture to an educational event organized by Bayer; payment via Neuropath Ltd, a company owned and operated by the author and their wife, for expert witness work in the courts in the UK and Ireland for expert neuropathology work, mostly in the case of unexpected child deaths for the HM courts; Editor in Chief for Neuropathology and Applied Neurobiology (paid a share of the profits of the journal); Lead for the Childhood Solid Tumour Domain of the Genomics England Clinical Interpretation Partnership; shareholder in Repath and Neuropath. K.A. No relevant relationships. P.G.F. NIH/ NCI for co-investigator in Pediatric Brain Tumor Consortium; NIN/NHGRI for co-principal investigator in Undiagnosed Diseases Network; payment for role as Associate Editor for The Journal of Pediatrics by Elsevier Publishing; personal stock holdings in Johnson & Johnson. M.T. No relevant relationships. T.P. NIH funded Pediatric Brain Tumor Consortium Neuroimaging Center grant; payments from Springer Publishing; President of the American Society of Neuroradiology. N.A.V. No relevant relationships. G.A.G. Participant on a DataSa-fety Monitoring Board or Advisory Board for Stanford University. S. Pfister Grants from EU (IMI-2, ERC), BMBF, Brain Tumor Charity, DFG, Deutsche Kinderkrebsstiftung, Deutsche Krebshilfe; three patents in DNA methylation– based tumor classification; leadership or fiduciary role in SAB Princess Maxima Center, SAB INCA, and SAB BioSkryb. E.T. Grants or contracts from FDA, Cure Starts Now, American Brain Tumor Foundation, NIH; consulting fees for being a scientific advisor for Oncoheroes Biosciences; support for attending meetings and/or travel for Cure Starts Now; provisional patent on a drill for craniosynostosis surgery; participant on a DataSafety Monitoring Board or Advisory Board for University of Alabama for a Phase I clinical trial of malignant pediatric brain tumors. A.J. No relevant relationships. V.R. No relevant relationships. K.W.Y. Participant on a DataSafety Monitoring Board or Advisory Board for Stanford University. Funding Information: M.Z. is funded by the National Institutes of Health (5T32CA009695-27). S.T. was supported by payment to the University College London (UCL) by Great Ormond Street Hospital (GOSH) Children{\textquoteright}s Charity. S.H.C. was supported by the Kathryn S.R. Lowry Endowed Chair at the University of Utah, Department of Neurosurgery. T.S.J. is supported by the Institute of Child Health – Newcastle University – Institute of Cancer Research High-Risk Childhood Brain Tumour (INSTINCT) Network and the Everest Centre, which are funded by The Brain Tumour Charity, as well as supported by the GOSH Children{\textquoteright}s Charity, Children with Cancer UK, Cancer Research UK, and the Olivia Hodson Cancer Fund. All research at the GOSH National Health Service Foundation Trust and UCL Great Ormond Street Institute of Child Health is made possible by the National Institute for Health Research GOSH Biomedical Research Centre. V.R. is supported by the Canadian Institutes for Health Research, the Canadian Cancer Society Emerging Scholars Award, the Garron Family Cancer Centre, and the C.R. Younger Foundation. K.W.Y. is supported by the American Brain Tumor Association (DG1800019). Publisher Copyright: {\textcopyright} RSNA, 2022.",
year = "2022",
month = aug,
doi = "10.1148/radiol.212137",
language = "English (US)",
volume = "304",
pages = "406--416",
journal = "Radiology",
issn = "0033-8419",
publisher = "Radiological Society of North America Inc.",
number = "2",
}