Characterization of Mediastinal Bulky Lymphomas with FDG-PET-Based Radiomics and Machine Learning Techniques.
FDG-PET
bulky lymphoma
diagnosis
machine learning
personalized treatment
precision medicine
radiomics
textural analysis
Journal
Cancers
ISSN: 2072-6694
Titre abrégé: Cancers (Basel)
Pays: Switzerland
ID NLM: 101526829
Informations de publication
Date de publication:
23 Mar 2023
23 Mar 2023
Historique:
received:
30
01
2023
revised:
11
03
2023
accepted:
20
03
2023
medline:
14
4
2023
entrez:
13
4
2023
pubmed:
14
4
2023
Statut:
epublish
Résumé
This study tested the diagnostic value of 18F-FDG PET/CT (FDG-PET) volumetric and texture parameters in the histological differentiation of mediastinal bulky disease due to classical Hodgkin lymphoma (cHL), primary mediastinal B-cell lymphoma (PMBCL) and grey zone lymphoma (GZL), using machine learning techniques. We reviewed 80 cHL, 29 PMBCL and 8 GZL adult patients with mediastinal bulky disease and histopathological diagnoses who underwent FDG-PET pre-treatment. Volumetric and radiomic parameters were measured using FDG-PET both for bulky lesions (BL) and for all lesions (AL) using LIFEx software (threshold SUV ≥ 2.5). Binary and multiclass classifications were performed with various machine learning techniques fed by a relevant subset of radiomic features. The analysis showed significant differences between the lymphoma groups in terms of SUVmax, SUVmean, MTV, TLG and several textural features of both first- and second-order grey level. Among machine learning classifiers, the tree-based ensembles achieved the best performance both for binary and multiclass classifications in histological differentiation. Our results support the value of metabolic heterogeneity as an imaging biomarker, and the use of radiomic features for early characterization of mediastinal bulky lymphoma.
Sections du résumé
BACKGROUND
BACKGROUND
This study tested the diagnostic value of 18F-FDG PET/CT (FDG-PET) volumetric and texture parameters in the histological differentiation of mediastinal bulky disease due to classical Hodgkin lymphoma (cHL), primary mediastinal B-cell lymphoma (PMBCL) and grey zone lymphoma (GZL), using machine learning techniques.
METHODS
METHODS
We reviewed 80 cHL, 29 PMBCL and 8 GZL adult patients with mediastinal bulky disease and histopathological diagnoses who underwent FDG-PET pre-treatment. Volumetric and radiomic parameters were measured using FDG-PET both for bulky lesions (BL) and for all lesions (AL) using LIFEx software (threshold SUV ≥ 2.5). Binary and multiclass classifications were performed with various machine learning techniques fed by a relevant subset of radiomic features.
RESULTS
RESULTS
The analysis showed significant differences between the lymphoma groups in terms of SUVmax, SUVmean, MTV, TLG and several textural features of both first- and second-order grey level. Among machine learning classifiers, the tree-based ensembles achieved the best performance both for binary and multiclass classifications in histological differentiation.
CONCLUSIONS
CONCLUSIONS
Our results support the value of metabolic heterogeneity as an imaging biomarker, and the use of radiomic features for early characterization of mediastinal bulky lymphoma.
Identifiants
pubmed: 37046592
pii: cancers15071931
doi: 10.3390/cancers15071931
pmc: PMC10093023
pii:
doi:
Types de publication
Journal Article
Langues
eng
Références
Med Phys. 2017 Sep;44(9):4608-4619
pubmed: 28513853
J Nucl Med. 2011 Mar;52(3):369-78
pubmed: 21321270
Pathobiology. 2018;85(1-2):130-145
pubmed: 28719907
Adv Anat Pathol. 2016 Sep;23(5):285-309
pubmed: 27441757
Q J Nucl Med Mol Imaging. 2021 Mar;65(1):72-78
pubmed: 31140234
J Nucl Med. 2015 Jan;56(1):38-44
pubmed: 25500829
J Comput Assist Tomogr. 2021 Mar-Apr 01;45(2):191-202
pubmed: 33273161
Br J Radiol. 2018 Nov;91(1091):20170926
pubmed: 29947266
Cancer Sci. 2015 Feb;106(2):186-93
pubmed: 25495273
Ann Hematol. 2010 Oct;89(10):985-91
pubmed: 20428871
Ann Oncol. 2018 Jan 1;29(1):264-270
pubmed: 29045536
J Nucl Med. 2017 Mar;58(3):387-392
pubmed: 27754906
PLoS One. 2015 Oct 16;10(10):e0140830
pubmed: 26473950
J Nucl Med. 2000 Jan;41(1):85-92
pubmed: 10647609
Eur J Nucl Med Mol Imaging. 2013 Jan;40(1):133-40
pubmed: 23064544
Eur J Nucl Med Mol Imaging. 2022 Apr;49(5):1535-1543
pubmed: 34850248
Ann Oncol. 2018 Oct 1;29(Suppl 4):iv19-iv29
pubmed: 29796651
J Clin Oncol. 1988 Jun;6(6):931-3
pubmed: 3373263
BMC Cancer. 2021 Nov 24;21(1):1268
pubmed: 34819043
Ann Hematol. 2012 May;91(5):697-703
pubmed: 22071570
J Nucl Med. 2016 Nov;57(11):1823-1828
pubmed: 27261515
Insights Imaging. 2020 Aug 12;11(1):91
pubmed: 32785796
Clin Cancer Res. 2007 Jun 15;13(12):3449-59
pubmed: 17575207
Semin Nucl Med. 2018 Jan;48(1):76-81
pubmed: 29195620
Comput Methods Programs Biomed. 2020 Mar;185:105153
pubmed: 31678792
Annu Rev Biomed Eng. 2017 Jun 21;19:221-248
pubmed: 28301734
J Clin Oncol. 2014 Sep 20;32(27):3059-68
pubmed: 25113753
Cancer. 1988 May 1;61(9):1776-81
pubmed: 3355975
Eur Radiol. 2017 Oct;27(10):4082-4090
pubmed: 28374077
Chin Med J (Engl). 2020 Nov 18;134(4):439-447
pubmed: 33230019
Ann Oncol. 2009 Sep;20(9):1543-1547
pubmed: 19474116
Maedica (Bucur). 2019 Jun;14(2):126-130
pubmed: 31523292
PLoS One. 2015 Dec 15;10(12):e0145063
pubmed: 26669541
Int J Radiat Oncol Biol Phys. 2013 Apr 1;85(5):1375-82
pubmed: 23219566
Cancer Manag Res. 2017 Dec 05;9:761-768
pubmed: 29263700
Clin Cancer Res. 2008 Feb 1;14(3):685-91
pubmed: 18245527
J Digit Imaging. 2019 Aug;32(4):582-596
pubmed: 31144149
Blood. 2016 May 19;127(20):2375-90
pubmed: 26980727
Commun Biol. 2021 Nov 12;4(1):1286
pubmed: 34773070
Cancer Res. 2018 Aug 15;78(16):4786-4789
pubmed: 29959149
Proc Natl Acad Sci U S A. 2008 Apr 1;105(13):5213-8
pubmed: 18362333
Nucl Med Mol Imaging. 2014 Dec;48(4):278-86
pubmed: 26396632
Med Image Anal. 2017 Dec;42:60-88
pubmed: 28778026
J Cancer Res Ther. 2012 Jan-Mar;8(1):96-102
pubmed: 22531522
Br J Radiol. 2020 Feb 01;93(1106):20180781
pubmed: 31860336
Cancer. 2003 Feb 15;97(4):1015-24
pubmed: 12569601
Eur J Nucl Med Mol Imaging. 2014 Nov;41(11):2017-22
pubmed: 24902639