Radiomics Can Distinguish Pediatric Supratentorial Embryonal Tumors, High-Grade Gliomas, and Ependymomas.
Journal
AJNR. American journal of neuroradiology
ISSN: 1936-959X
Titre abrégé: AJNR Am J Neuroradiol
Pays: United States
ID NLM: 8003708
Informations de publication
Date de publication:
04 2022
04 2022
Historique:
received:
17
09
2021
accepted:
25
01
2022
pubmed:
2
4
2022
medline:
13
4
2022
entrez:
1
4
2022
Statut:
ppublish
Résumé
Pediatric supratentorial tumors such as embryonal tumors, high-grade gliomas, and ependymomas are difficult to distinguish by histopathology and imaging because of overlapping features. We applied machine learning to uncover MR imaging-based radiomics phenotypes that can differentiate these tumor types. Our retrospective cohort of 231 patients from 7 participating institutions had 50 embryonal tumors, 127 high-grade gliomas, and 54 ependymomas. For each tumor volume, we extracted 900 Image Biomarker Standardization Initiative-based PyRadiomics features from T2-weighted and gadolinium-enhanced T1-weighted images. A reduced feature set was obtained by sparse regression analysis and was used as input for 6 candidate classifier models. Training and test sets were randomly allocated from the total cohort in a 75:25 ratio. The final classifier model for embryonal tumor-versus-high-grade gliomas identified 23 features with an area under the curve of 0.98; the sensitivity, specificity, positive predictive value, negative predictive value, and accuracy were 0.85, 0.91, 0.79, 0.94, and 0.89, respectively. The classifier for embryonal tumor-versus-ependymomas identified 4 features with an area under the curve of 0.82; the sensitivity, specificity, positive predictive value, negative predictive value, and accuracy were 0.93, 0.69, 0.76, 0.90, and 0.81, respectively. The classifier for high-grade gliomas-versus-ependymomas identified 35 features with an area under the curve of 0.96; the sensitivity, specificity, positive predictive value, negative predictive value, and accuracy were 0.82, 0.94, 0.82, 0.94, and 0.91, respectively. In this multi-institutional study, we identified distinct radiomic phenotypes that distinguish pediatric supratentorial tumors, high-grade gliomas, and ependymomas with high accuracy. Incorporation of this technique in diagnostic algorithms can improve diagnosis, risk stratification, and treatment planning.
Sections du résumé
BACKGROUND AND PURPOSE
Pediatric supratentorial tumors such as embryonal tumors, high-grade gliomas, and ependymomas are difficult to distinguish by histopathology and imaging because of overlapping features. We applied machine learning to uncover MR imaging-based radiomics phenotypes that can differentiate these tumor types.
MATERIALS AND METHODS
Our retrospective cohort of 231 patients from 7 participating institutions had 50 embryonal tumors, 127 high-grade gliomas, and 54 ependymomas. For each tumor volume, we extracted 900 Image Biomarker Standardization Initiative-based PyRadiomics features from T2-weighted and gadolinium-enhanced T1-weighted images. A reduced feature set was obtained by sparse regression analysis and was used as input for 6 candidate classifier models. Training and test sets were randomly allocated from the total cohort in a 75:25 ratio.
RESULTS
The final classifier model for embryonal tumor-versus-high-grade gliomas identified 23 features with an area under the curve of 0.98; the sensitivity, specificity, positive predictive value, negative predictive value, and accuracy were 0.85, 0.91, 0.79, 0.94, and 0.89, respectively. The classifier for embryonal tumor-versus-ependymomas identified 4 features with an area under the curve of 0.82; the sensitivity, specificity, positive predictive value, negative predictive value, and accuracy were 0.93, 0.69, 0.76, 0.90, and 0.81, respectively. The classifier for high-grade gliomas-versus-ependymomas identified 35 features with an area under the curve of 0.96; the sensitivity, specificity, positive predictive value, negative predictive value, and accuracy were 0.82, 0.94, 0.82, 0.94, and 0.91, respectively.
CONCLUSIONS
In this multi-institutional study, we identified distinct radiomic phenotypes that distinguish pediatric supratentorial tumors, high-grade gliomas, and ependymomas with high accuracy. Incorporation of this technique in diagnostic algorithms can improve diagnosis, risk stratification, and treatment planning.
Identifiants
pubmed: 35361575
pii: ajnr.A7481
doi: 10.3174/ajnr.A7481
pmc: PMC8993189
doi:
Types de publication
Journal Article
Multicenter Study
Langues
eng
Sous-ensembles de citation
IM
Pagination
603-610Informations de copyright
© 2022 by American Journal of Neuroradiology.
Références
J Negat Results Biomed. 2011 May 31;10:7
pubmed: 21627842
Cell. 2016 Feb 25;164(5):1060-1072
pubmed: 26919435
Mod Pathol. 2020 Jan;33(1):47-56
pubmed: 31375768
Neuro Oncol. 2015 Aug;17(8):1157-65
pubmed: 25916887
Pediatr Blood Cancer. 2015 May;62(5):776-83
pubmed: 25704363
AJNR Am J Neuroradiol. 2021 Sep;42(9):1702-1708
pubmed: 34266866
Radiology. 2020 May;295(2):328-338
pubmed: 32154773
Asian J Neurosurg. 2015 Oct-Dec;10(4):276-81
pubmed: 26425155
AJNR Am J Neuroradiol. 2020 Jul;41(7):1279-1285
pubmed: 32661052
Neuroradiology. 2000 Jun;42(6):393-8
pubmed: 10929296
Cancer. 2003 Sep 15;98(6):1243-52
pubmed: 12973849
Clin Radiol. 2007 Jan;62(1):43-52
pubmed: 17145263
Neuro Oncol. 2021 Aug 2;23(8):1231-1251
pubmed: 34185076
Neurosurgery. 2019 Jul 1;85(1):41-49
pubmed: 29917116
Acta Neuropathol. 2016 Jun;131(6):803-20
pubmed: 27157931
J Neurooncol. 2017 Sep;134(3):541-549
pubmed: 28357536
Front Oncol. 2019 May 21;9:374
pubmed: 31165039
Neuroimaging Clin N Am. 2017 Feb;27(1):39-67
pubmed: 27889023
Cancers (Basel). 2020 Dec 18;12(12):
pubmed: 33353180
Pediatr Neurol. 2003 Nov;29(5):430-4
pubmed: 14684239
Cancer Cell. 2017 Oct 9;32(4):520-537.e5
pubmed: 28966033
AJNR Am J Neuroradiol. 2019 Nov;40(11):1796-1803
pubmed: 31601576
Radiographics. 2018 Nov-Dec;38(7):2102-2122
pubmed: 30422762
Crit Rev Oncol Hematol. 2007 Jul;63(1):81-9
pubmed: 17482475
Acad Radiol. 2021 Mar;28(3):318-327
pubmed: 32222329
J Med Imaging (Bellingham). 2020 Jul;7(4):042803
pubmed: 32206688
AJNR Am J Neuroradiol. 2021 Mar;42(3):448-456
pubmed: 33509914
Cancer. 2018 Dec 15;124(24):4633-4649
pubmed: 30383900
Cancer Res. 2017 Nov 1;77(21):e104-e107
pubmed: 29092951
AJR Am J Roentgenol. 2013 May;200(5):W483-503
pubmed: 23617516
AJNR Am J Neuroradiol. 2020 Sep;41(9):1718-1725
pubmed: 32816765
Front Oncol. 2012 Aug 24;2:105
pubmed: 22937526
J Clin Oncol. 2018 Oct 17;:JCO2017764720
pubmed: 30332335
Nat Rev Clin Oncol. 2017 Dec;14(12):749-762
pubmed: 28975929
AJNR Am J Neuroradiol. 2014 May;35(5):1009-15
pubmed: 24309122