M Iv1, M Zhou1,2, K Shpanskaya1, S Perreault3, Z Wang2, E Tranvinh1, B Lanzman1, S Vajapeyam4, N A Vitanza5, P G Fisher6, Y J Cho7, S Laughlin8, V Ramaswamy8, M D Taylor8, S H Cheshier9, G A Grant10, T Young Poussaint4, O Gevaert2, K W Yeom11,12. 1. From the Department of Radiology (M.I., M.Z., K.S., E.T., B.L., K.W.Y.). 2. Stanford Center for Biomedical Informatics (M.Z., O.G., Z.W.). 3. Department of Pediatrics (S.P.), Pediatric Neurology, Centre Hospitalier Universitaire Sainte Justine, University of Montréal, Montreal, Quebec, Canada. 4. Department of Radiology (S.V., T.Y.P.), Boston Children's Hospital, Harvard University, Boston, Massachusetts. 5. Department Pediatrics Hematology-Oncology (N.A.V.), Seattle Children's Hospital, University of Washington, Seattle, Washington. 6. Department of Pediatrics (P.G.F.), Pediatric Neurology. 7. Department of Pediatrics (Y.J.C.), Pediatric Neurology, Oregon Health & Science University, Portland, Oregon. 8. Departments of Radiology, Neuro-Oncology, and Neurosurgery (S.L., V.R., M.D.T.), Hospital for Sick Children, Toronto, Ontario, Canada. 9. Department of Neurosurgery (S.H.C.), Pediatric Neurosurgery, University of Utah, Salt Lake City, Utah. 10. Department of Neurosurgery (G.A.G.), Pediatric Neurosurgery, Lucile Packard Children's Hospital, Stanford University, Palo Alto, California. 11. From the Department of Radiology (M.I., M.Z., K.S., E.T., B.L., K.W.Y.) kyeom@stanford.edu. 12. Department of Radiology (K.W.Y.), Artificial Intelligence in Medicine and Imaging, Stanford University, Stanford, California.
Abstract
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.
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.
Authors: Radka Stoyanova; Mandeep Takhar; Yohann Tschudi; John C Ford; Gabriel Solórzano; Nicholas Erho; Yoganand Balagurunathan; Sanoj Punnen; Elai Davicioni; Robert J Gillies; Alan Pollack Journal: Transl Cancer Res Date: 2016-08 Impact factor: 1.241
Authors: Kristian W Pajtler; Hendrik Witt; Martin Sill; David T W Jones; Volker Hovestadt; Fabian Kratochwil; Khalida Wani; Ruth Tatevossian; Chandanamali Punchihewa; Pascal Johann; Jüri Reimand; Hans-Jörg Warnatz; Marina Ryzhova; Steve Mack; Vijay Ramaswamy; David Capper; Leonille Schweizer; Laura Sieber; Andrea Wittmann; Zhiqin Huang; Peter van Sluis; Richard Volckmann; Jan Koster; Rogier Versteeg; Daniel Fults; Helen Toledano; Smadar Avigad; Lindsey M Hoffman; Andrew M Donson; Nicholas Foreman; Ekkehard Hewer; Karel Zitterbart; Mark Gilbert; Terri S Armstrong; Nalin Gupta; Jeffrey C Allen; Matthias A Karajannis; David Zagzag; Martin Hasselblatt; Andreas E Kulozik; Olaf Witt; V Peter Collins; Katja von Hoff; Stefan Rutkowski; Torsten Pietsch; Gary Bader; Marie-Laure Yaspo; Andreas von Deimling; Peter Lichter; Michael D Taylor; Richard Gilbertson; David W Ellison; Kenneth Aldape; Andrey Korshunov; Marcel Kool; Stefan M Pfister Journal: Cancer Cell Date: 2015-05-11 Impact factor: 31.743
Authors: Kristen W Yeom; Bret C Mobley; Robert M Lober; Jalal B Andre; Sonia Partap; Hannes Vogel; Patrick D Barnes Journal: AJR Am J Roentgenol Date: 2013-04 Impact factor: 3.959
Authors: Patrick Grossmann; Vivek Narayan; Ken Chang; Rifaquat Rahman; Lauren Abrey; David A Reardon; Lawrence H Schwartz; Patrick Y Wen; Brian M Alexander; Raymond Huang; Hugo J W L Aerts Journal: Neuro Oncol Date: 2017-11-29 Impact factor: 12.300
Authors: A Sorana Morrissy; Florence M G Cavalli; Marc Remke; Vijay Ramaswamy; David J H Shih; Borja L Holgado; Hamza Farooq; Laura K Donovan; Livia Garzia; Sameer Agnihotri; Erin N Kiehna; Eloi Mercier; Chelsea Mayoh; Simon Papillon-Cavanagh; Hamid Nikbakht; Tenzin Gayden; Jonathon Torchia; Daniel Picard; Diana M Merino; Maria Vladoiu; Betty Luu; Xiaochong Wu; Craig Daniels; Stuart Horswell; Yuan Yao Thompson; Volker Hovestadt; Paul A Northcott; David T W Jones; John Peacock; Xin Wang; Stephen C Mack; Jüri Reimand; Steffen Albrecht; Adam M Fontebasso; Nina Thiessen; Yisu Li; Jacqueline E Schein; Darlene Lee; Rebecca Carlsen; Michael Mayo; Kane Tse; Angela Tam; Noreen Dhalla; Adrian Ally; Eric Chuah; Young Cheng; Patrick Plettner; Haiyan I Li; Richard D Corbett; Tina Wong; William Long; James Loukides; Pawel Buczkowicz; Cynthia E Hawkins; Uri Tabori; Brian R Rood; John S Myseros; Roger J Packer; Andrey Korshunov; Peter Lichter; Marcel Kool; Stefan M Pfister; Ulrich Schüller; Peter Dirks; Annie Huang; Eric Bouffet; James T Rutka; Gary D Bader; Charles Swanton; Yusanne Ma; Richard A Moore; Andrew J Mungall; Jacek Majewski; Steven J M Jones; Sunit Das; David Malkin; Nada Jabado; Marco A Marra; Michael D Taylor Journal: Nat Genet Date: 2017-04-10 Impact factor: 38.330
Authors: Siddhi Ramesh; Sukarn Chokkara; Timothy Shen; Ajay Major; Samuel L Volchenboum; Anoop Mayampurath; Mark A Applebaum Journal: JCO Clin Cancer Inform Date: 2021-12
Authors: Soumen Khatua; Laurence J N Cooper; David I Sandberg; Leena Ketonen; Jason M Johnson; Michael E Rytting; Diane D Liu; Heather Meador; Prashant Trikha; Robin J Nakkula; Gregory K Behbehani; Dristhi Ragoonanan; Sumit Gupta; Aikaterini Kotrotsou; Tagwa Idris; Elizabeth J Shpall; Katy Rezvani; Rivka Colen; Wafik Zaky; Dean A Lee; Vidya Gopalakrishnan Journal: Neuro Oncol Date: 2020-08-17 Impact factor: 12.300
Authors: Rashad Jabarkheel; Nisreen Amayiri; Derek Yecies; Yuhao Huang; Sebastian Toescu; Liana Nobre; Donald J Mabbott; Sniya V Sudhakar; Prateek Malik; Suzanne Laughlin; Maisa Swaidan; Maysa Al Hussaini; Awni Musharbash; Geeta Chacko; Leni G Mathew; Paul G Fisher; Darren Hargrave; Ute Bartels; Uri Tabori; Stefan M Pfister; Kristian Aquilina; Michael D Taylor; Gerald A Grant; Eric Bouffet; Kshitij Mankad; Kristen W Yeom; Vijay Ramaswamy Journal: Neuro Oncol Date: 2020-02-20 Impact factor: 12.300
Authors: C A P F Alves; U Löbel; J S Martin-Saavedra; S Toescu; M H Tsunemi; S R Teixeira; K Mankad; D Hargrave; T S Jacques; C da Costa Leite; F G Gonçalves; A Vossough; F D'Arco Journal: AJNR Am J Neuroradiol Date: 2021-03-04 Impact factor: 3.825