Evaluation of magnetic resonance imaging-based radiomics characteristics for differentiation of benign and malignant peripheral nerve sheath tumors in neurofibromatosis type 1.
NF1
atypical neurofibroma
machine learning
magnetic resonance imaging
malignant peripheral nerve sheath tumor
Journal
Neuro-oncology
ISSN: 1523-5866
Titre abrégé: Neuro Oncol
Pays: England
ID NLM: 100887420
Informations de publication
Date de publication:
03 10 2022
03 10 2022
Historique:
pubmed:
16
4
2022
medline:
5
10
2022
entrez:
15
4
2022
Statut:
ppublish
Résumé
Patients with neurofibromatosis type 1 (NF1) develop benign (BPNST), premalignant atypical (ANF), and malignant (MPNST) peripheral nerve sheath tumors. Radiological differentiation of these entities is challenging. Therefore, we aimed to evaluate the value of a magnetic resonance imaging (MRI)-based radiomics machine-learning (ML) classifier for differentiation of these three entities of internal peripheral nerve sheath tumors in NF1 patients. MRI was performed at 3T in 36 NF1 patients (20 male; age: 31 ± 11 years). Segmentation of 117 BPNSTs, 17 MPNSTs, and 8 ANFs was manually performed using T2w spectral attenuated inversion recovery sequences. One hundred seven features per lesion were extracted using PyRadiomics and applied for BPNST versus MPNST differentiation. A 5-feature radiomics signature was defined based on the most important features and tested for signature-based BPNST versus MPNST classification (random forest [RF] classification, leave-one-patient-out evaluation). In a second step, signature feature expressions for BPNSTs, ANFs, and MPNSTs were evaluated for radiomics-based classification for these three entities. The mean area under the receiver operator characteristic curve (AUC) for the radiomics-based BPNST versus MPNST differentiation was 0.94, corresponding to correct classification of on average 16/17 MPNSTs and 114/117 BPNSTs (sensitivity: 94%, specificity: 97%). Exploratory analysis with the eight ANFs revealed intermediate radiomic feature characteristics in-between BPNST and MPNST tumor feature expression. In this proof-of-principle study, ML using MRI-based radiomics characteristics allows sensitive and specific differentiation of BPNSTs and MPNSTs in NF1 patients. Feature expression of premalignant atypical tumors was distributed in-between benign and malignant tumor feature expressions, which illustrates biological plausibility of the considered radiomics characteristics.
Sections du résumé
BACKGROUND
Patients with neurofibromatosis type 1 (NF1) develop benign (BPNST), premalignant atypical (ANF), and malignant (MPNST) peripheral nerve sheath tumors. Radiological differentiation of these entities is challenging. Therefore, we aimed to evaluate the value of a magnetic resonance imaging (MRI)-based radiomics machine-learning (ML) classifier for differentiation of these three entities of internal peripheral nerve sheath tumors in NF1 patients.
METHODS
MRI was performed at 3T in 36 NF1 patients (20 male; age: 31 ± 11 years). Segmentation of 117 BPNSTs, 17 MPNSTs, and 8 ANFs was manually performed using T2w spectral attenuated inversion recovery sequences. One hundred seven features per lesion were extracted using PyRadiomics and applied for BPNST versus MPNST differentiation. A 5-feature radiomics signature was defined based on the most important features and tested for signature-based BPNST versus MPNST classification (random forest [RF] classification, leave-one-patient-out evaluation). In a second step, signature feature expressions for BPNSTs, ANFs, and MPNSTs were evaluated for radiomics-based classification for these three entities.
RESULTS
The mean area under the receiver operator characteristic curve (AUC) for the radiomics-based BPNST versus MPNST differentiation was 0.94, corresponding to correct classification of on average 16/17 MPNSTs and 114/117 BPNSTs (sensitivity: 94%, specificity: 97%). Exploratory analysis with the eight ANFs revealed intermediate radiomic feature characteristics in-between BPNST and MPNST tumor feature expression.
CONCLUSION
In this proof-of-principle study, ML using MRI-based radiomics characteristics allows sensitive and specific differentiation of BPNSTs and MPNSTs in NF1 patients. Feature expression of premalignant atypical tumors was distributed in-between benign and malignant tumor feature expressions, which illustrates biological plausibility of the considered radiomics characteristics.
Identifiants
pubmed: 35426432
pii: 6568977
doi: 10.1093/neuonc/noac100
pmc: PMC9527508
doi:
Types de publication
Journal Article
Research Support, Non-U.S. Gov't
Langues
eng
Sous-ensembles de citation
IM
Pagination
1790-1798Informations de copyright
© The Author(s) 2022. Published by Oxford University Press on behalf of the Society for Neuro-Oncology.
Références
Invest Radiol. 2019 Apr;54(4):221-228
pubmed: 30433891
J Clin Oncol. 2016 Jun 10;34(17):1978-86
pubmed: 26926675
J Neurol Surg A Cent Eur Neurosurg. 2016 Nov;77(6):495-504
pubmed: 27276118
PLoS One. 2015 May 04;10(5):e0124165
pubmed: 25938522
Hum Pathol. 2017 Sep;67:1-10
pubmed: 28551330
AJNR Am J Neuroradiol. 2014 Aug;35(8):1615-20
pubmed: 24763412
Neuro Oncol. 2008 Aug;10(4):593-8
pubmed: 18559970
Clin Nucl Med. 2013 Jan;38(1):e19-25
pubmed: 23242059
Neurology. 2020 Jun 16;94(24):e2521-e2531
pubmed: 32345730
AJR Am J Roentgenol. 2010 Jun;194(6):1568-74
pubmed: 20489098
Neurology. 2005 Jul 26;65(2):205-11
pubmed: 16043787
Int J Cancer. 1984 Jan 15;33(1):37-42
pubmed: 6693192
Neuro Oncol. 2019 Mar 18;21(4):508-516
pubmed: 30496452
J Med Genet. 2002 May;39(5):311-4
pubmed: 12011145
J Cancer Res Clin Oncol. 2009 Jul;135(7):891-900
pubmed: 19101731
Oncologist. 2014 Feb;19(2):193-201
pubmed: 24470531
PLoS Genet. 2021 May 5;17(5):e1009517
pubmed: 33951044
J Neuroradiol. 2019 May;46(3):179-185
pubmed: 29958847
Cancers (Basel). 2013 May 08;5(2):519-28
pubmed: 24216989
Eur J Nucl Med Mol Imaging. 2017 Oct;44(11):1845-1852
pubmed: 28589254
Acta Neuropathol. 2012 Mar;123(3):295-319
pubmed: 22327363
Arch Dermatol. 2005 Jan;141(1):71-4
pubmed: 15655144
J Neurooncol. 2014 Jan;116(2):307-13
pubmed: 24166582
N Engl J Med. 2020 Apr 9;382(15):1430-1442
pubmed: 32187457
Cancer Res. 2017 Nov 1;77(21):e104-e107
pubmed: 29092951
Eur Radiol. 2014 Feb;24(2):405-12
pubmed: 24097302
Genes Chromosomes Cancer. 2011 Dec;50(12):1021-32
pubmed: 21987445
Neurology. 2016 Aug 16;87(7 Suppl 1):S31-9
pubmed: 27527647
J Med Genet. 2009 Feb;46(2):81-5
pubmed: 18930997
PLoS One. 2009 Sep 18;4(9):e7087
pubmed: 19763254
Neurology. 2019 Sep 10;93(11):e1076-e1084
pubmed: 31395668
Neuro Oncol. 2021 Apr 12;23(4):557-571
pubmed: 33326583
Neuroimage. 2006 Jul 1;31(3):1116-28
pubmed: 16545965
Eur Spine J. 2015 Oct;24(10):2119-25
pubmed: 25015181
Arch Neurol. 1988 May;45(5):575-8
pubmed: 3128965
J Med Genet. 1989 Nov;26(11):712-21
pubmed: 2511319
Eur J Nucl Med Mol Imaging. 2013 May;40(5):685-92
pubmed: 23232507
Korean J Radiol. 2019 Jul;20(7):1124-1137
pubmed: 31270976
Nat Rev Clin Oncol. 2017 Dec;14(12):749-762
pubmed: 28975929