1
|
Wiestler B, Bison B, Behrens L, Tüchert S, Metz M, Griessmair M, Jakob M, Schlegel PG, Binder V, von Luettichau I, Metzler M, Johann P, Hau P, Frühwald M. Human-Level Differentiation of Medulloblastoma from Pilocytic Astrocytoma: A Real-World Multicenter Pilot Study. Cancers (Basel) 2024; 16:1474. [PMID: 38672556 PMCID: PMC11048511 DOI: 10.3390/cancers16081474] [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: 03/12/2024] [Revised: 04/05/2024] [Accepted: 04/08/2024] [Indexed: 04/28/2024] Open
Abstract
Medulloblastoma and pilocytic astrocytoma are the two most common pediatric brain tumors with overlapping imaging features. In this proof-of-concept study, we investigated using a deep learning classifier trained on a multicenter data set to differentiate these tumor types. We developed a patch-based 3D-DenseNet classifier, utilizing automated tumor segmentation. Given the heterogeneity of imaging data (and available sequences), we used all individually available preoperative imaging sequences to make the model robust to varying input. We compared the classifier to diagnostic assessments by five readers with varying experience in pediatric brain tumors. Overall, we included 195 preoperative MRIs from children with medulloblastoma (n = 69) or pilocytic astrocytoma (n = 126) across six university hospitals. In the 64-patient test set, the DenseNet classifier achieved a high AUC of 0.986, correctly predicting 62/64 (97%) diagnoses. It misclassified one case of each tumor type. Human reader accuracy ranged from 100% (expert neuroradiologist) to 80% (resident). The classifier performed significantly better than relatively inexperienced readers (p < 0.05) and was on par with pediatric neuro-oncology experts. Our proof-of-concept study demonstrates a deep learning model based on automated tumor segmentation that can reliably preoperatively differentiate between medulloblastoma and pilocytic astrocytoma, even in heterogeneous data.
Collapse
Affiliation(s)
- Benedikt Wiestler
- Department of Neuroradiology, School of Medicine and Health, Technical University of Munich, 81675 Munich, Germany (M.G.)
- TranslaTUM, Center for Translational Cancer Research, Technical University of Munich, 81675 Munich, Germany
- Study Groups on CNS Tumors Within the Bavarian Cancer Research Center (BZKF)
| | - Brigitte Bison
- Study Groups on CNS Tumors Within the Bavarian Cancer Research Center (BZKF)
- KIONET, Kinderonkologisches Netzwerk Bayern
- Diagnostic and Interventional Neuroradiology, Faculty of Medicine, University Hospital Augsburg, 86156 Augsburg, Germany; (B.B.); (L.B.)
- Neuroradiological Reference Center for the Pediatric Brain Tumor (HIT) Studies of the German Society of Pediatric Oncology and Hematology, Faculty of Medicine, University Hospital Augsburg, 86156 Augsburg, Germany
| | - Lars Behrens
- Study Groups on CNS Tumors Within the Bavarian Cancer Research Center (BZKF)
- KIONET, Kinderonkologisches Netzwerk Bayern
- Diagnostic and Interventional Neuroradiology, Faculty of Medicine, University Hospital Augsburg, 86156 Augsburg, Germany; (B.B.); (L.B.)
- Neuroradiological Reference Center for the Pediatric Brain Tumor (HIT) Studies of the German Society of Pediatric Oncology and Hematology, Faculty of Medicine, University Hospital Augsburg, 86156 Augsburg, Germany
| | - Stefanie Tüchert
- Study Groups on CNS Tumors Within the Bavarian Cancer Research Center (BZKF)
- Department of Diagnostic and Interventional Radiology, University Hospital Augsburg, 86156 Augsburg, Germany
| | - Marie Metz
- Department of Neuroradiology, School of Medicine and Health, Technical University of Munich, 81675 Munich, Germany (M.G.)
- Study Groups on CNS Tumors Within the Bavarian Cancer Research Center (BZKF)
| | - Michael Griessmair
- Department of Neuroradiology, School of Medicine and Health, Technical University of Munich, 81675 Munich, Germany (M.G.)
- Study Groups on CNS Tumors Within the Bavarian Cancer Research Center (BZKF)
| | - Marcus Jakob
- Study Groups on CNS Tumors Within the Bavarian Cancer Research Center (BZKF)
- KIONET, Kinderonkologisches Netzwerk Bayern
- Department of Pediatric Hematology, Oncology and Stem Cell Transplantation, University of Regensburg, 93053 Regensburg, Germany;
| | - Paul-Gerhardt Schlegel
- Study Groups on CNS Tumors Within the Bavarian Cancer Research Center (BZKF)
- KIONET, Kinderonkologisches Netzwerk Bayern
- Department of Pediatric Hematology, Oncology and Stem Cell Transplantation, University Children’s Hospital Würzburg, 97080 Würzburg, Germany;
| | - Vera Binder
- Study Groups on CNS Tumors Within the Bavarian Cancer Research Center (BZKF)
- KIONET, Kinderonkologisches Netzwerk Bayern
- Department of Pediatrics, Dr. Von Hauner Children’s Hospital, University Hospital, LMU Munich, 80539 Munich, Germany;
| | - Irene von Luettichau
- Study Groups on CNS Tumors Within the Bavarian Cancer Research Center (BZKF)
- KIONET, Kinderonkologisches Netzwerk Bayern
- Division of Pediatric Hematology and Oncology, Department of Pediatrics, Kinderklinik München Schwabing, Children’s Cancer Research Center, TUM School of Medicine and Health, Technical University of Munich, 80333 Munich, Germany;
| | - Markus Metzler
- Study Groups on CNS Tumors Within the Bavarian Cancer Research Center (BZKF)
- KIONET, Kinderonkologisches Netzwerk Bayern
- Pediatric Oncology and Hematology, Department of Pediatrics and Adolescent Medicine, University Hospital Erlangen, Comprehensive Cancer Center Erlangen-EMN (CCC ER-EMN), 91054 Erlangen, Germany;
| | - Pascal Johann
- Study Groups on CNS Tumors Within the Bavarian Cancer Research Center (BZKF)
- KIONET, Kinderonkologisches Netzwerk Bayern
- Swabian Children’s Cancer Center, Pediatrics and Adolescent Medicine, University Hospital Augsburg, 86156 Augsburg, Germany; (P.J.); (M.F.)
| | - Peter Hau
- Study Groups on CNS Tumors Within the Bavarian Cancer Research Center (BZKF)
- Department of Neurology and Wilhelm Sander-NeuroOncology Unit, University Hospital Regensburg, 93053 Regensburg, Germany;
| | - Michael Frühwald
- Study Groups on CNS Tumors Within the Bavarian Cancer Research Center (BZKF)
- KIONET, Kinderonkologisches Netzwerk Bayern
- Swabian Children’s Cancer Center, Pediatrics and Adolescent Medicine, University Hospital Augsburg, 86156 Augsburg, Germany; (P.J.); (M.F.)
| |
Collapse
|
2
|
Jarry VDM, Pereira FV, Dalaqua M, Duarte JÁ, França Junior MC, Reis F. Common and uncommon neuroimaging manifestations of ataxia: an illustrated guide for the trainee radiologist. Part 2 - neoplastic, congenital, degenerative, and hereditary diseases. Radiol Bras 2022; 55:259-266. [PMID: 35983347 PMCID: PMC9380611 DOI: 10.1590/0100-3984.2021.0112] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Abstract] [Track Full Text] [Download PDF] [Figures] [Journal Information] [Subscribe] [Scholar Register] [Received: 07/05/2021] [Accepted: 12/09/2021] [Indexed: 11/22/2022] Open
Abstract
Abstract Ataxia is defined as a lack of coordination of voluntary movement, caused by a variety of factors. Ataxia can be classified by the age at onset and type (chronic or acute). The causative lesions involve the cerebellum and cerebellar connections. The correct, appropriate use of neuroimaging, particularly magnetic resonance imaging, can make the diagnosis relatively straightforward and facilitate implementation of the appropriate clinical management. The purpose of this pictorial essay is to describe the imaging findings of ataxia, based on cases obtained from the archives of a tertiary care hospital, with a review of the most important findings. We also discuss and review the imaging aspects of neoplastic diseases, malformations, degenerative diseases, and hereditary diseases related to ataxia.
Collapse
Affiliation(s)
| | | | | | | | | | - Fabiano Reis
- Universidade Estadual de Campinas (Unicamp), Brazil
| |
Collapse
|
3
|
Attia NM, Sayed SAA, Riad KF, Korany GM. Magnetic resonance spectroscopy in pediatric brain tumors: how to make a more confident diagnosis. THE EGYPTIAN JOURNAL OF RADIOLOGY AND NUCLEAR MEDICINE 2020. [DOI: 10.1186/s43055-020-0135-3] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Abstract] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Indexed: 11/10/2022] Open
Abstract
Abstract
Background
Non-invasive diagnosis of pediatric brain tumors can be challenging due to diverse tumor pathologies and similar imaging appearances. Magnetic resonance spectroscopy (MRS), when combined with high spatial resolution anatomic imaging obtained with conventional magnetic resonance imaging (MRI), provides metabolic information within the lesion as well as the surrounding tissue. The differentiation of neoplastic from non-neoplastic lesions and low-grade from high-grade neoplasms is essential for determining the choice of treatment and the best treatment plan. We aimed to measure specific metabolic ratios and evaluate metabolic profiles of various lesions by MRS to assist in making a more confident diagnosis.
Results
The choline/creatine (Cho/Cr), choline/N-acetylaspartate (Cho/NAA), and Cho/NAA+Cr ratios all had statistically significant values for the differentiation between neoplastic and non-neoplastic lesions at cutoffs 1.8, 2, and 0.8 respectively. The Cho/NAA, Cho/Cr, Cho/NAA+Cr, and myo-inositol/creatine (mI/Cr) ratios all had statistically significant values for the differentiation of high-grade from low-grade neoplasms at cutoffs 3.3, 3.5, 1.3, and 1.5 respectively. The presence of a lipid lactate peak was only significant for differentiating high-grade from low-grade neoplasms. Medulloblastomas, diffuse pontine gliomas, and choroid plexus carcinoma all showed characteristic metabolic profiles on MRS. Metastasis showed lower Cho/NAA and Cho/Cr ratios outside the tumor margin than high-grade neoplasms.
Conclusion
The use of certain metabolite ratios with high sensitivity and specificity to distinguish neoplastic from non-neoplastic lesions and low-grade from high-grade neoplasms while assessing the metabolic profile of the lesion aids in the non-invasive diagnosis of pediatric brain tumors. MRS facilitates earlier treatment planning by determining tumor spatial extent and predicting tumor behavior with potential to solve sampling problems of inaccessible and heterogenous lesions as well as unnecessary sampling of benign lesions.
Collapse
|
4
|
Magnetic resonance imaging spectroscopy in pediatric atypical teratoid rhabdoid tumors of the brain. J Pediatr Hematol Oncol 2014; 36:e341-5. [PMID: 24072251 DOI: 10.1097/mph.0000000000000041] [Citation(s) in RCA: 13] [Impact Index Per Article: 1.3] [Reference Citation Analysis] [Abstract] [Track Full Text] [Journal Information] [Submit a Manuscript] [Subscribe] [Scholar Register] [Indexed: 11/26/2022]
Abstract
BACKGROUND Pediatric central nervous system (CNS) atypical teratoid rhabdoid tumors (ATRT) are highly malignant tumors characterized by SMARCB1 gene abnormalities. Despite chemoradiation responsiveness, most children die of disease. No imaging findings distinguish ATRT from other malignant brain tumors. This study sought to describe magnetic resonance spectroscopy (MRS) of childhood CNS ATRT and identify metabolite patterns for diagnosis and disease status monitoring. METHODS Data from 7 children diagnosed with CNS ATRT from 2007 to 2010, whose imaging included MRS, were retrospectively reviewed. RESULTS Age at diagnosis ranged from 2.5 to 54 months. Tumors were large with calcium and cysts and avid gadolinium enhancement. All were isointense on T1-weighted imaging and mildly hyperintense on T2-weighted imaging. Short-TE MRS showed prominent lactate+lipid and choline, minimal N-acetyl acetate (NAA), and rarely minimal myoinositol and low creatine peaks. Long TE showed prominent choline, minimal NAA, and rarely low lactate peaks. CONCLUSIONS The combination of prominent choline and lactate+lipids peaks, and generally absent NAA and myoinositol peaks by MRS in this panel of ATRT expands existing information and provides a potentially distinct metabolite profile from other malignant pediatric brain tumors, including medulloblastoma. Prospective, comparative quantitative MRS of ATRT with other pediatric CNS tumors is warranted.
Collapse
|