1
|
Mu W, Dahmoush H. Classification and neuroimaging of ependymal tumors. Front Pediatr 2023; 11:1181211. [PMID: 37287627 PMCID: PMC10242666 DOI: 10.3389/fped.2023.1181211] [Citation(s) in RCA: 1] [Impact Index Per Article: 1.0] [Reference Citation Analysis] [Abstract] [Key Words] [Track Full Text] [Figures] [Journal Information] [Submit a Manuscript] [Subscribe] [Scholar Register] [Received: 03/07/2023] [Accepted: 04/14/2023] [Indexed: 06/09/2023] Open
Abstract
Ependymal tumors arise from the ependymal cell remnants of the cerebral ventricles, the central canal of the spinal cord, or the filum terminale or conus medullaris, although most pediatric supratentorial ependymomas do not exhibit clear communication or abutment of the ventricles. In this article, we discuss the classification, imaging characteristics, and clinical settings of these tumors. The WHO 2021 classification system has categorized ependymal tumors based on histopathologic and molecular features and location, in which they are grouped as supratentorial, posterior fossa (PF), and spinal. The supratentorial tumors are defined by either the ZFTA (formerly RELA) fusion or the YAP1 fusion. Posterior fossa tumors are divided into group A and group B based on methylation. On imaging, supratentorial and infratentorial ependymomas may arise from the ventricles and commonly contain calcifications and cystic components, with variable hemorrhage and heterogeneous enhancement. Spinal ependymomas are defined by MYCN amplification. These tumors are less commonly calcified and may present with the "cap sign," with T2 hypointensity due to hemosiderin deposition. Myxopapillary ependymoma and subependymoma remain tumor subtypes, with no change related to molecular classification as this does not provide additional clinical utility. Myxopapillary ependymomas are intradural and extramedullary tumors at the filum terminale and/or conus medullaris and may also present the cap sign. Subependymomas are homogeneous when small and may be heterogeneous and contain calcifications when larger. These tumors typically do not demonstrate enhancement. Clinical presentation and prognosis vary depending on tumor location and type. Knowledge of the updated WHO classification of the central nervous system in conjunction with imaging features is critical for accurate diagnosis and treatment.
Collapse
Affiliation(s)
- Weiya Mu
- Department of Radiology, Stanford Health Care, Stanford, CA, United States
| | - Hisham Dahmoush
- Department of Radiology, Lucile Packard Children’s Hospital, Stanford, CA, United States
| |
Collapse
|
2
|
Gonçalves FG, Zandifar A, Ub Kim JD, Tierradentro-García LO, Ghosh A, Khrichenko D, Andronikou S, Vossough A. Application of Apparent Diffusion Coefficient Histogram Metrics for Differentiation of Pediatric Posterior Fossa Tumors : A Large Retrospective Study and Brief Review of Literature. Clin Neuroradiol 2022; 32:1097-1108. [PMID: 35674799 DOI: 10.1007/s00062-022-01179-6] [Citation(s) in RCA: 1] [Impact Index Per Article: 0.5] [Reference Citation Analysis] [Abstract] [Key Words] [MESH Headings] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 02/25/2022] [Accepted: 05/08/2022] [Indexed: 02/07/2023]
Abstract
PURPOSE This study aimed to evaluate the application of apparent diffusion coefficient (ADC) histogram analysis to differentiate posterior fossa tumors (PFTs) in children. METHODS A total of 175 pediatric patients with PFT, including 75 pilocytic astrocytomas (PA), 59 medulloblastomas, 16 ependymomas, and 13 atypical teratoid rhabdoid tumors (ATRT), were analyzed. Tumors were visually assessed using DWI trace and conventional MRI images and manually segmented and post-processed using parametric software (pMRI). Furthermore, tumor ADC values were normalized to the thalamus and cerebellar cortex. The following histogram metrics were obtained: entropy, minimum, 10th, and 90th percentiles, maximum, mean, median, skewness, and kurtosis to distinguish the different types of tumors. Kruskal Wallis and Mann-Whitney U tests were used to evaluate the differences. Finally, receiver operating characteristic (ROC) curves were utilized to determine the optimal cut-off values for differentiating the various PFTs. RESULTS Most ADC histogram metrics showed significant differences between PFTs (p < 0.001) except for entropy, skewness, and kurtosis. There were significant pairwise differences in ADC metrics for PA versus medulloblastoma, PA versus ependymoma, PA versus ATRT, medulloblastoma versus ependymoma, and ependymoma versus ATRT (all p < 0.05). Our results showed no significant differences between medulloblastoma and ATRT. Normalized ADC data showed similar results to the absolute ADC value analysis. ROC curve analysis for normalized ADCmedian values to thalamus showed 94.9% sensitivity (95% CI: 85-100%) and 93.3% specificity (95% CI: 87-100%) for differentiating medulloblastoma from ependymoma. CONCLUSION ADC histogram metrics can be applied to differentiate most types of posterior fossa tumors in children.
Collapse
Affiliation(s)
- Fabrício Guimarães Gonçalves
- Department of Radiology, Division of Neuroradiology, The Children's Hospital of Philadelphia, Philadelphia, PA, USA
| | - Alireza Zandifar
- Department of Radiology, Division of Neuroradiology, The Children's Hospital of Philadelphia, Philadelphia, PA, USA.
| | - Jorge Du Ub Kim
- Department of Radiology, Division of Neuroradiology, The Children's Hospital of Philadelphia, Philadelphia, PA, USA
| | | | - Adarsh Ghosh
- Department of Radiology, Division of Neuroradiology, The Children's Hospital of Philadelphia, Philadelphia, PA, USA
| | - Dmitry Khrichenko
- Department of Radiology, Division of Neuroradiology, The Children's Hospital of Philadelphia, Philadelphia, PA, USA
| | - Savvas Andronikou
- Department of Radiology, Division of Neuroradiology, The Children's Hospital of Philadelphia, Philadelphia, PA, USA
- Department of Radiology, Perelman School of Medicine, University of Pennsylvania, Philadelphia, PA, USA
| | - Arastoo Vossough
- Department of Radiology, Division of Neuroradiology, The Children's Hospital of Philadelphia, Philadelphia, PA, USA
- Department of Radiology, Perelman School of Medicine, University of Pennsylvania, Philadelphia, PA, USA
| |
Collapse
|
3
|
Machine Learning in the Classification of Pediatric Posterior Fossa Tumors: A Systematic Review. Cancers (Basel) 2022; 14:cancers14225608. [PMID: 36428701 PMCID: PMC9688156 DOI: 10.3390/cancers14225608] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Abstract] [Key Words] [Track Full Text] [Figures] [Journal Information] [Subscribe] [Scholar Register] [Received: 09/29/2022] [Revised: 11/02/2022] [Accepted: 11/11/2022] [Indexed: 11/17/2022] Open
Abstract
Background: Posterior fossa tumors (PFTs) are a morbid group of central nervous system tumors that most often present in childhood. While early diagnosis is critical to drive appropriate treatment, definitive diagnosis is currently only achievable through invasive tissue collection and histopathological analyses. Machine learning has been investigated as an alternative means of diagnosis. In this systematic review and meta-analysis, we evaluated the primary literature to identify all machine learning algorithms developed to classify and diagnose pediatric PFTs using imaging or molecular data. Methods: Of the 433 primary papers identified in PubMed, EMBASE, and Web of Science, 25 ultimately met the inclusion criteria. The included papers were extracted for algorithm architecture, study parameters, performance, strengths, and limitations. Results: The algorithms exhibited variable performance based on sample size, classifier(s) used, and individual tumor types being investigated. Ependymoma, medulloblastoma, and pilocytic astrocytoma were the most studied tumors with algorithm accuracies ranging from 37.5% to 94.5%. A minority of studies compared the developed algorithm to a trained neuroradiologist, with three imaging-based algorithms yielding superior performance. Common algorithm and study limitations included small sample sizes, uneven representation of individual tumor types, inconsistent performance reporting, and a lack of application in the clinical environment. Conclusions: Artificial intelligence has the potential to improve the speed and accuracy of diagnosis in this field if the right algorithm is applied to the right scenario. Work is needed to standardize outcome reporting and facilitate additional trials to allow for clinical uptake.
Collapse
|
4
|
Wang C, Uh J, Patni T, Merchant T, Li Y, Hua CH, Acharya S. Toward MR-only proton therapy planning for pediatric brain tumors: synthesis of relative proton stopping power images with multiple sequence MRI and development of an online quality assurance tool. Med Phys 2022; 49:1559-1570. [PMID: 35075670 DOI: 10.1002/mp.15479] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Abstract] [Key Words] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 10/22/2021] [Revised: 12/23/2021] [Accepted: 01/11/2022] [Indexed: 11/11/2022] Open
Abstract
PURPOSE To generate synthetic relative proton-stopping-power (sRPSP) images from MRI sequence(s) and develop an online quality assurance (QA) tool for sRPSP to facilitate safe integration of MR-only proton planning into clinical practice. MATERIALS AND METHODS Planning CT and MR images of 195 pediatric brain tumor patients were utilized (training: 150, testing: 45). Seventeen consistent-cycle Generative Adversarial Network (ccGAN) models were trained separately using paired CT-converted RPSP and MRI datasets to transform a subject's MRI into sRPSP. T1-weighted (T1W), T2-weighted (T2W), and FLAIR MRI were permutated to form 17 combinations, with or without preprocessing, for determining the optimal training sequence(s). For evaluation, sRPSP images were converted to synthetic CT (sCT) and compared to the real CT in terms of mean absolute error (MAE) in HU. For QA, sCT was deformed and compared to a reference template built from training dataset to produce a flag map, highlighting pixels that deviate by >100 HU and fall outside the mean ± standard deviation reference intensity. The gamma intensity analysis (10%/3mm) of the deformed sCT against the QA template on the intensity difference was investigated as a surrogate of sCT accuracy. RESULTS The sRPSP images generated from a single T1W or T2W sequence outperformed that generated from multi-MRI sequences in terms of MAE (all P<0.05). Preprocessing with N4 bias and histogram matching reduced MAE of T2W MRI-based sCT (54±21 HU vs. 42±13 HU, P = .002). The gamma intensity analysis of sCT against the QA template was highly correlated with the MAE of sCT against the real CT in the testing cohort (r = -0.89 for T1W sCT; r = -0.93 for T2W sCT). CONCLUSION Accurate sRPSP images can be generated from T1W/T2W MRI for proton planning. A QA tool highlights regions of inaccuracy, flagging problematic cases unsuitable for clinical use. This article is protected by copyright. All rights reserved.
Collapse
Affiliation(s)
- Chuang Wang
- Department of Radiation Oncology, St. Jude Children's Research Hospital, Memphis, TN, United States Of America
| | - Jinsoo Uh
- Department of Radiation Oncology, St. Jude Children's Research Hospital, Memphis, TN, United States Of America
| | - Tushar Patni
- Department of Biostatistics, St. Jude Children's Research Hospital, Memphis, TN, United States Of America
| | - Thomas Merchant
- Department of Radiation Oncology, St. Jude Children's Research Hospital, Memphis, TN, United States Of America
| | - Yimei Li
- Department of Biostatistics, St. Jude Children's Research Hospital, Memphis, TN, United States Of America
| | - Chia-Ho Hua
- Department of Radiation Oncology, St. Jude Children's Research Hospital, Memphis, TN, United States Of America
| | - Sahaja Acharya
- Department of Radiation Oncology, St. Jude Children's Research Hospital, Memphis, TN, United States Of America.,Department of Radiation Oncology and Molecular Radiation Sciences, Johns Hopkins Medicine, Baltimore, MD, United States Of America
| |
Collapse
|
5
|
Zhang M, Wang E, Yecies D, Tam LT, Han M, Toescu S, Wright JN, Altinmakas E, Chen E, Radmanesh A, Nemelka J, Oztekin O, Wagner MW, Lober RM, Ertl-Wagner B, Ho CY, Mankad K, Vitanza NA, Cheshier SH, Jacques TS, Fisher PG, Aquilina K, Said M, Jaju A, Pfister S, Taylor MD, Grant GA, Mattonen S, Ramaswamy V, Yeom KW. Radiomic Signatures of Posterior Fossa Ependymoma: Molecular Subgroups and Risk Profiles. Neuro Oncol 2021; 24:986-994. [PMID: 34850171 DOI: 10.1093/neuonc/noab272] [Citation(s) in RCA: 7] [Impact Index Per Article: 2.3] [Reference Citation Analysis] [Abstract] [Key Words] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Indexed: 11/13/2022] Open
Abstract
BACKGROUND The risk profile for posterior fossa ependymoma (EP) depends on surgical and molecular status [Group A (PFA) versus Group B (PFB)]. While subtotal tumor resection is known to confer worse prognosis, MRI-based EP risk-profiling is unexplored. We aimed to apply machine learning strategies to link MRI-based biomarkers of high-risk EP and also to distinguish PFA from PFB. METHODS We extracted 1800 quantitative features from presurgical T2-weighted (T2-MRI) and gadolinium-enhanced T1-weighted (T1-MRI) imaging of 157 EP patients. We implemented nested cross-validation to identify features for risk score calculations and apply a Cox model for survival analysis. We conducted additional feature selection for PFA versus PFB and examined performance across three candidate classifiers. RESULTS For all EP patients with GTR, we identified four T2-MRI-based features and stratified patients into high- and low-risk groups, with 5-year overall survival rates of 62% and 100%, respectively (p < 0.0001). Among presumed PFA patients with GTR, four T1-MRI and five T2-MRI features predicted divergence of high- and low-risk groups, with 5-year overall survival rates of 62.7% and 96.7%, respectively (p = 0.002). T1-MRI-based features showed the best performance distinguishing PFA from PFB with an AUC of 0.86. CONCLUSIONS We present machine learning strategies to identify MRI phenotypes that distinguish PFA from PFB, as well as high- and low-risk PFA. We also describe quantitative image predictors of aggressive EP tumors that might assist risk-profiling after surgery. Future studies could examine translating radiomics as an adjunct to EP risk assessment when considering therapy strategies or trial candidacy.
Collapse
Affiliation(s)
- Michael Zhang
- Department of Neurosurgery, Stanford Hospital and Clinics, Stanford, CA, USA.,Department of Radiology, Lucile Packard Children's Hospital, Stanford, CA, USA
| | - Edward Wang
- Department of Medical Biophysics, Western University, London, ON, Canada
| | - Derek Yecies
- Department of Neurosurgery, Stanford Hospital and Clinics, Stanford, CA, USA.,Department of Radiology, Lucile Packard Children's Hospital, Stanford, CA, USA
| | - Lydia T Tam
- Stanford School of Medicine, Stanford University, Stanford, CA, USA
| | - Michelle Han
- Department of Pediatrics, Children's Hospital of Philadelphia, Philadelphia, PA
| | - Sebastian Toescu
- Department of Neurosurgery, Great Ormond Street Institute of Child Health, London, UK
| | - Jason N Wright
- Department of Radiology, Seattle Children's Hospital, and Harborview Medical Center, Seattle, WA, USA
| | - Emre Altinmakas
- Department of Radiology, Koç University School of Medicine, Istanbul, Turkey
| | - Eric Chen
- Department of Clinical Radiology & Imaging Sciences, Riley Children's Hospital, Indianapolis, IA, USA
| | - Alireza Radmanesh
- Department of Radiology, New York University Grossman School of Medicine, New York, NY, USA
| | - Jordan Nemelka
- Division of Pediatric Neurosurgery, Department of Neurosurgery, Huntsman Cancer Institute, University of Utah School of Medicine, Intermountain Healthcare Primary Children's Hospital, Salt Lake City, UT, USA
| | - Ozgur Oztekin
- Department of Neuroradiology, Cigli Education and Research Hospital, and Tepecik Education and Research Hospital, Izmir, Turkey
| | - Matthias W Wagner
- Department of Diagnostic Imaging, The Hospital for Sick Children, ON, Canada
| | - Robert M Lober
- Division of Neurosurgery, Dayton Children's Hospital, Dayton, OH, USA
| | - Birgit Ertl-Wagner
- Department of Diagnostic Imaging, The Hospital for Sick Children, ON, Canada
| | - Chang Y Ho
- Department of Clinical Radiology & Imaging Sciences, Riley Children's Hospital, Indianapolis, IA, USA
| | - Kshitij Mankad
- Department of Radiology, Great Ormond Street Institute of Child Health, London, UK
| | - Nicholas A Vitanza
- Division of Pediatric Hematology/Oncology, Department of Pediatrics, Seattle Children's Hospital, Seattle WA, USA
| | - Samuel H Cheshier
- Division of Pediatric Neurosurgery, Department of Neurosurgery, Huntsman Cancer Institute, University of Utah School of Medicine, Intermountain Healthcare Primary Children's Hospital, Salt Lake City, UT, USA
| | - Tom S Jacques
- Department of Developmental Biology & Cancer, University College London Great Ormond Street Institute of Child Health, and Great Ormond Street Hospital for Children NHS Foundation Trust, London, UK
| | - Paul G Fisher
- Department of Neurology, Lucile Packard Children's Hospital, Stanford University, Palo Alto, CA, USA
| | - Kristian Aquilina
- Department of Neurosurgery, Great Ormond Street Institute of Child Health, London, UK
| | - Mourad Said
- Radiology Department Centre International Carthage Médicale, Monastir, Tunisia
| | - Alok Jaju
- Department of Medical Imaging, Ann and Robert H. Lurie Children's Hospital of Chicago, Chicago, IL, USA
| | - Stefan Pfister
- Department of Pediatrics, Hopp Children' Cancer Center, Heidelberg, Germany
| | - Michael D Taylor
- Division of Neurosurgery, The Hospital for Sick Children, Toronto, ON, Canada
| | - Gerald A Grant
- Department of Neurosurgery, Lucile Packard Children's Hospital, Stanford, CA, USA
| | - Sarah Mattonen
- Department of Medical Biophysics, Western University, London, ON, Canada
| | - Vijay Ramaswamy
- Division of Haematology/Oncology, Programme in Developmental and Stem Cell Biology, The Hospital for Sick Children, Toronto, ON, Canada.,Department of Medical Biophysics, University of Toronto, Toronto, ON, Canada
| | - Kristen W Yeom
- Department of Radiology, Lucile Packard Children's Hospital, Stanford, CA, USA
| |
Collapse
|
6
|
Zhang M, Wong SW, Wright JN, Toescu S, Mohammadzadeh M, Han M, Lummus S, Wagner MW, Yecies D, Lai H, Eghbal A, Radmanesh A, Nemelka J, Harward S, Malinzak M, Laughlin S, Perreault S, Braun KRM, Vossough A, Poussaint T, Goetti R, Ertl-Wagner B, Ho CY, Oztekin O, Ramaswamy V, Mankad K, Vitanza NA, Cheshier SH, Said M, Aquilina K, Thompson E, Jaju A, Grant GA, Lober RM, Yeom KW. Machine Assist for Pediatric Posterior Fossa Tumor Diagnosis: A Multinational Study. Neurosurgery 2021; 89:892-900. [PMID: 34392363 DOI: 10.1093/neuros/nyab311] [Citation(s) in RCA: 7] [Impact Index Per Article: 2.3] [Reference Citation Analysis] [Abstract] [Key Words] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 02/24/2021] [Accepted: 06/09/2021] [Indexed: 01/06/2023] Open
Abstract
BACKGROUND Clinicians and machine classifiers reliably diagnose pilocytic astrocytoma (PA) on magnetic resonance imaging (MRI) but less accurately distinguish medulloblastoma (MB) from ependymoma (EP). One strategy is to first rule out the most identifiable diagnosis. OBJECTIVE To hypothesize a sequential machine-learning classifier could improve diagnostic performance by mimicking a clinician's strategy of excluding PA before distinguishing MB from EP. METHODS We extracted 1800 total Image Biomarker Standardization Initiative (IBSI)-based features from T2- and gadolinium-enhanced T1-weighted images in a multinational cohort of 274 MB, 156 PA, and 97 EP. We designed a 2-step sequential classifier - first ruling out PA, and next distinguishing MB from EP. For each step, we selected the best performing model from 6-candidate classifier using a reduced feature set, and measured performance on a holdout test set with the microaveraged F1 score. RESULTS Optimal diagnostic performance was achieved using 2 decision steps, each with its own distinct imaging features and classifier method. A 3-way logistic regression classifier first distinguished PA from non-PA, with T2 uniformity and T1 contrast as the most relevant IBSI features (F1 score 0.8809). A 2-way neural net classifier next distinguished MB from EP, with T2 sphericity and T1 flatness as most relevant (F1 score 0.9189). The combined, sequential classifier was with F1 score 0.9179. CONCLUSION An MRI-based sequential machine-learning classifiers offer high-performance prediction of pediatric posterior fossa tumors across a large, multinational cohort. Optimization of this model with demographic, clinical, imaging, and molecular predictors could provide significant advantages for family counseling and surgical planning.
Collapse
Affiliation(s)
- Michael Zhang
- Department of Neurosurgery, Stanford Hospital and Clinics, Stanford, California, USA.,Department of Radiology, Lucile Packard Children's Hospital, Stanford, California, USA
| | - Samuel W Wong
- Department of Statistics, Stanford University, Stanford, California, USA
| | - Jason N Wright
- Department of Radiology, Seattle Children's Hospital, Seattle, Washington, USA.,Department of Radiology, Harborview Medical Center, Seattle, Washington, USA
| | - Sebastian Toescu
- Department of Neurosurgery, Great Ormond Street Hospital, London, United Kingdom
| | | | - Michelle Han
- Department of Pediatrics, Children's Hospital of Philadelphia, Philadelphia, Pennsylvania, USA
| | - Seth Lummus
- Department of Physiology and Nutrition, University of Colorado Colorado Springs, Colorado Springs, Colorado, USA
| | - Matthias W Wagner
- Department of Diagnostic Imaging, The Hospital for Sick Children, Toronto, Canada
| | - Derek Yecies
- Department of Neurosurgery, Lucile Packard Children's Hospital, Stanford, California, USA
| | - Hollie Lai
- Department of Radiology, Children's Hospital of Orange County, Orange, California, USA
| | - Azam Eghbal
- Department of Radiology, Children's Hospital of Orange County, Orange, California, USA
| | - Alireza Radmanesh
- Department of Radiology, New York University Grossman School of Medicine, New York, New York, USA
| | - Jordan Nemelka
- Department of Neurosurgery, University of Utah School of Medicine, Salt Lake City, Utah, USA
| | - Stephen Harward
- Department of Neurosurgery, Duke Children's Hospital & Health Center, Durham, North Carolina, USA
| | - Michael Malinzak
- Department of Radiology, Duke Children's Hospital & Health Center, Durham, North Carolina, USA
| | - Suzanne Laughlin
- Department of Diagnostic Imaging, The Hospital for Sick Children, Toronto, Canada
| | - Sebastien Perreault
- Division of Child Neurology, Department of Pediatrics, Centre Hospitalier Universitaire Sainte-Justine, Université de Montréal, Montreal, Canada
| | - Kristina R M Braun
- Department of Clinical Radiology & Imaging Sciences, Riley Children's Hospital, Indianapolis, Iowa, USA
| | - Arastoo Vossough
- Department of Radiology, Children's Hospital of Philadelphia, Philadelphia, Pennsylvania, USA
| | - Tina Poussaint
- Department of Radiology, Boston Children's Hospital, Boston, Massachusetts, USA
| | - Robert Goetti
- Department of Medical Imaging, The Children's Hospital at Westmead, The University of Sydney, Sydney, Australia
| | - Birgit Ertl-Wagner
- Department of Diagnostic Imaging, The Hospital for Sick Children, Toronto, Canada
| | - Chang Y Ho
- Department of Clinical Radiology & Imaging Sciences, Riley Children's Hospital, Indianapolis, Iowa, USA
| | - Ozgur Oztekin
- Department of Neuroradiology, Cigli Education and Research Hospital, Izmir, Turkey.,Department of Neuroradiology, Tepecik Education and Research Hospital, Izmir, Turkey
| | - Vijay Ramaswamy
- Division of Haematology/Oncology, Department of Pediatrics, The Hospital for Sick Children, Toronto, Canada
| | - Kshitij Mankad
- Department of Radiology, Great Ormond Street Hospital, London, United Kingdom
| | - Nicholas A Vitanza
- Division of Pediatric Hematology/Oncology, Department of Pediatrics, Seattle Children's Hospital, Seattle Washington, USA
| | - Samuel H Cheshier
- Department of Neurosurgery, University of Utah School of Medicine, Salt Lake City, Utah, USA
| | - Mourad Said
- Radiology Department, Centre International Carthage Médicale, Monastir, Tunisia
| | - Kristian Aquilina
- Department of Neurosurgery, Great Ormond Street Hospital, London, United Kingdom
| | - Eric Thompson
- Department of Neurosurgery, Duke Children's Hospital & Health Center, Durham, North Carolina, USA
| | - Alok Jaju
- Department of Medical Imaging, Ann and Robert H. Lurie Children's Hospital of Chicago, Chicago, Illinois, USA
| | - Gerald A Grant
- Department of Neurosurgery, Lucile Packard Children's Hospital, Stanford, California, USA
| | - Robert M Lober
- Division of Neurosurgery, Dayton Children's Hospital, Dayton, Ohio, USA
| | - Kristen W Yeom
- Department of Radiology, Lucile Packard Children's Hospital, Stanford, California, USA
| |
Collapse
|
7
|
Zhang M, Wong SW, Lummus S, Han M, Radmanesh A, Ahmadian SS, Prolo LM, Lai H, Eghbal A, Oztekin O, Cheshier SH, Fisher PG, Ho CY, Vogel H, Vitanza NA, Lober RM, Grant GA, Jaju A, Yeom KW. Radiomic Phenotypes Distinguish Atypical Teratoid/Rhabdoid Tumors from Medulloblastoma. AJNR Am J Neuroradiol 2021; 42:1702-1708. [PMID: 34266866 DOI: 10.3174/ajnr.a7200] [Citation(s) in RCA: 5] [Impact Index Per Article: 1.7] [Reference Citation Analysis] [Abstract] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 02/13/2021] [Accepted: 04/05/2021] [Indexed: 01/06/2023]
Abstract
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.
Collapse
Affiliation(s)
- M Zhang
- From the Departments of Neurosurgery (M.Z.)
| | - S W Wong
- Department of Statistics (S.W.W.), Stanford University, Stanford, California
| | - S Lummus
- Department of Physiology and Nutrition (S.L.), University of Colorado, Colorado Springs, Colorado
| | - M Han
- Department of Pediatrics (M.H.), Children's Hospital of Philadelphia, University of Pennsylvania, Philadelphia, Pennsylvania
| | - A Radmanesh
- Department of Radiology (A.R.), New York University Grossman School of Medicine, New York, New York
| | - S S Ahmadian
- Pathology (S.S.A., H.V.), Stanford Medical Center, Stanford University, Stanford, California
| | - L M Prolo
- Departments of Neurosurgery (L.M.P., G.A.G.)
| | - H Lai
- Department of Radiology (H.L., A.E.), Children's Hospital of Orange County, Orange, California and University of California, Irvine, Irvine, California
| | - A Eghbal
- Department of Radiology (H.L., A.E.), Children's Hospital of Orange County, Orange, California and University of California, Irvine, Irvine, California
| | - O Oztekin
- Department of Neuroradiology (O.O.), Cigli Education and Research Hospital, Bakircay University, Izmir, Turkey.,Department of Neuroradiology (O.O.), Tepecik Education and Research Hospital, Health Science University, Izmir, Turkey
| | - S H Cheshier
- Division of Pediatric Neurosurgery (S.H.C.), Department of Neurosurgery, Huntsman Cancer Institute, Intermountain Healthcare Primary Children's Hospital, University of Utah School of Medicine, Salt Lake City, Utah
| | | | - C Y Ho
- Departments of Clinical Radiology & Imaging Sciences (C.Y.H.), Riley Children's Hospital, Indiana University, Indianapolis, Indiana
| | - H Vogel
- Pathology (S.S.A., H.V.), Stanford Medical Center, Stanford University, Stanford, California
| | - N A Vitanza
- Division of Pediatric Hematology/Oncology (N.A.V.), Department of Pediatrics, Seattle Children's Hospital, Seattle, Washington
| | - R M Lober
- Division of Neurosurgery (R.M.L.), Department of Pediatrics, Wright State University Boonshoft School of Medicine, Dayton Children's Hospital, Dayton, Ohio
| | - G A Grant
- Departments of Neurosurgery (L.M.P., G.A.G.)
| | - A Jaju
- Department of Medical Imaging (A.J.), Ann and Robert H. Lurie Children's Hospital of Chicago, Northwestern University Feinberg School of Medicine, Chicago, Illinois
| | - K W Yeom
- Radiology (K.W.Y.), Lucile Packard Children's Hospital, Stanford University, Palo Alto, California
| |
Collapse
|
8
|
Zhou H, Hu R, Tang O, Hu C, Tang L, Chang K, Shen Q, Wu J, Zou B, Xiao B, Boxerman J, Chen W, Huang RY, Yang L, Bai HX, Zhu C. Automatic Machine Learning to Differentiate Pediatric Posterior Fossa Tumors on Routine MR Imaging. AJNR Am J Neuroradiol 2020; 41:1279-1285. [PMID: 32661052 PMCID: PMC7357647 DOI: 10.3174/ajnr.a6621] [Citation(s) in RCA: 29] [Impact Index Per Article: 7.3] [Reference Citation Analysis] [Abstract] [MESH Headings] [Grants] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 11/30/2019] [Accepted: 04/30/2020] [Indexed: 12/20/2022]
Abstract
BACKGROUND AND PURPOSE Differentiating the types of pediatric posterior fossa tumors on routine imaging may help in preoperative evaluation and guide surgical resection planning. However, qualitative radiologic MR imaging review has limited performance. This study aimed to compare different machine learning approaches to classify pediatric posterior fossa tumors on routine MR imaging. MATERIALS AND METHODS This retrospective study included preoperative MR imaging of 288 patients with pediatric posterior fossa tumors, including medulloblastoma (n = 111), ependymoma (n = 70), and pilocytic astrocytoma (n = 107). Radiomics features were extracted from T2-weighted images, contrast-enhanced T1-weighted images, and ADC maps. Models generated by standard manual optimization by a machine learning expert were compared with automatic machine learning via the Tree-Based Pipeline Optimization Tool for performance evaluation. RESULTS For 3-way classification, the radiomics model by automatic machine learning with the Tree-Based Pipeline Optimization Tool achieved a test micro-averaged area under the curve of 0.91 with an accuracy of 0.83, while the most optimized model based on the feature-selection method χ2 score and the Generalized Linear Model classifier achieved a test micro-averaged area under the curve of 0.92 with an accuracy of 0.74. Tree-Based Pipeline Optimization Tool models achieved significantly higher accuracy than average qualitative expert MR imaging review (0.83 versus 0.54, P < .001). For binary classification, Tree-Based Pipeline Optimization Tool models achieved an area under the curve of 0.94 with an accuracy of 0.85 for medulloblastoma versus nonmedulloblastoma, an area under the curve of 0.84 with an accuracy of 0.80 for ependymoma versus nonependymoma, and an area under the curve of 0.94 with an accuracy of 0.88 for pilocytic astrocytoma versus non-pilocytic astrocytoma. CONCLUSIONS Automatic machine learning based on routine MR imaging classified pediatric posterior fossa tumors with high accuracy compared with manual expert pipeline optimization and qualitative expert MR imaging review.
Collapse
Affiliation(s)
- H Zhou
- Department of Neurology (H.Z., L.T., B.X.), Xiangya Hospital of Central South University, Changsha, Hunan, China
| | - R Hu
- From the School of Computer Science and Engineering (R.H., B.Z., C.Z.)
| | - O Tang
- Warren Alpert Medical School, Brown University (O.T.), Providence, Rhode Island
| | - C Hu
- Department of Neurology (C.H.), Hunan Provincial People's Hospital, Changsha, Hunan, China
| | - L Tang
- Department of Neurology (H.Z., L.T., B.X.), Xiangya Hospital of Central South University, Changsha, Hunan, China
| | - K Chang
- Department of Radiology (K.C.), Athinoula A. Martinos Center for Biomedical Imaging, Massachusetts General Hospital, Harvard Medical School, Boston, Massachusetts
| | - Q Shen
- Radiology (Q.S., J.W.), Second Xiangya Hospital of Central South University, Changsha, Hunan, China
| | - J Wu
- Radiology (Q.S., J.W.), Second Xiangya Hospital of Central South University, Changsha, Hunan, China
| | - B Zou
- From the School of Computer Science and Engineering (R.H., B.Z., C.Z.)
| | - B Xiao
- Department of Neurology (H.Z., L.T., B.X.), Xiangya Hospital of Central South University, Changsha, Hunan, China
| | - J Boxerman
- Department of Diagnostic Imaging (J.B., H.X.B.), Rhode Island Hospital
| | - W Chen
- Department of Pathology (W.C.), Hunan Children's Hospital, Changsha, Hunan, China
| | - R Y Huang
- Department of Radiology (R.Y.H.), Brigham and Women's Hospital, Boston, Massachusetts
| | - L Yang
- Departments of Neurology (L.Y.)
| | - H X Bai
- Department of Diagnostic Imaging (J.B., H.X.B.), Rhode Island Hospital
| | - C Zhu
- From the School of Computer Science and Engineering (R.H., B.Z., C.Z.)
- College of Literature and Journalism (C.Z.), Central South University, Changsha, Hunan, China
- Mobile Health Ministry of Education-China Mobile Joint Laboratory (C.Z.), China
| |
Collapse
|