Methodological framework for radiomics applications in Hodgkin's lymphoma.
Feature selection
Lymphoma
Outcome prediction
PET/CT
Radiomics
Response prediction
Silhouette
Similarity
Journal
European journal of hybrid imaging
ISSN: 2510-3636
Titre abrégé: Eur J Hybrid Imaging
Pays: England
ID NLM: 101724113
Informations de publication
Date de publication:
01 Jun 2020
01 Jun 2020
Historique:
received:
22
03
2020
accepted:
06
05
2020
entrez:
30
6
2021
pubmed:
1
7
2021
medline:
1
7
2021
Statut:
epublish
Résumé
According to published data, radiomics features differ between lesions of refractory/relapsing HL patients from those of long-term responders. However, several methodological aspects have not been elucidated yet. The study aimed at setting up a methodological framework in radiomics applications in Hodgkin's lymphoma (HL), especially at (a) developing a novel feature selection approach, (b) evaluating radiomic intra-patient lesions' similarity, and (c) classifying relapsing refractory (R/R) vs non-(R/R) patients. We retrospectively included 85 patients (male:female = 52:33; median age 35 years, range 19-74). LIFEx (www.lifexsoft.org) was used for [ HL fingerprints included up to 15 features. Intra-patient lesion similarity analysis resulted in mean/median silhouette values below 0.5 (low similarity especially in the non-R/R group). In the test set, the fingerprint_One classification accuracy was 62% (78% sensitivity and 53% specificity); the classification by RUBTE using fingerprint_All resulted in 82% accuracy (70% sensitivity and 88% specificity). Lesion similarity analysis was developed, and it allowed to demonstrate that HL lesions were not homogeneous within patients in terms of radiomics signature. Therefore, a random target lesion selection should not be adopted for radiomics applications. Moreover, the classifier to predict R/R vs non-R/R performed the best when all the lesions were used.
Sections du résumé
BACKGROUND
BACKGROUND
According to published data, radiomics features differ between lesions of refractory/relapsing HL patients from those of long-term responders. However, several methodological aspects have not been elucidated yet.
PURPOSE
OBJECTIVE
The study aimed at setting up a methodological framework in radiomics applications in Hodgkin's lymphoma (HL), especially at (a) developing a novel feature selection approach, (b) evaluating radiomic intra-patient lesions' similarity, and (c) classifying relapsing refractory (R/R) vs non-(R/R) patients.
METHODS
METHODS
We retrospectively included 85 patients (male:female = 52:33; median age 35 years, range 19-74). LIFEx (www.lifexsoft.org) was used for [
RESULTS
RESULTS
HL fingerprints included up to 15 features. Intra-patient lesion similarity analysis resulted in mean/median silhouette values below 0.5 (low similarity especially in the non-R/R group). In the test set, the fingerprint_One classification accuracy was 62% (78% sensitivity and 53% specificity); the classification by RUBTE using fingerprint_All resulted in 82% accuracy (70% sensitivity and 88% specificity).
CONCLUSIONS
CONCLUSIONS
Lesion similarity analysis was developed, and it allowed to demonstrate that HL lesions were not homogeneous within patients in terms of radiomics signature. Therefore, a random target lesion selection should not be adopted for radiomics applications. Moreover, the classifier to predict R/R vs non-R/R performed the best when all the lesions were used.
Identifiants
pubmed: 34191173
doi: 10.1186/s41824-020-00078-8
pii: 10.1186/s41824-020-00078-8
pmc: PMC8218114
doi:
Types de publication
Journal Article
Langues
eng
Pagination
9Références
Hematol Oncol. 2019 Jun;37 Suppl 1:87-91
pubmed: 31187532
Blood Lymphat Cancer. 2017;7:37-52
pubmed: 28701859
Eur Radiol. 2020 Jan;30(1):523-536
pubmed: 31350588
Clin Nucl Med. 2019 Oct;44(10):e559-e565
pubmed: 31306204
Int J Clin Oncol. 2019 Oct;24(10):1292-1300
pubmed: 31165310
J Nucl Med. 2015 Jan;56(1):38-44
pubmed: 25500829
Ann Nucl Med. 2018 Jul;32(6):410-416
pubmed: 29754276
Cancer Res. 2018 Aug 15;78(16):4786-4789
pubmed: 29959149
Br J Haematol. 2014 Feb;164(3):315-28
pubmed: 24131306
Eur J Nucl Med Mol Imaging. 2019 Dec;46(13):2760-2769
pubmed: 31286200
Int J Mol Sci. 2019 Nov 05;20(21):
pubmed: 31694167
Med Phys. 2017 Sep;44(9):4608-4619
pubmed: 28513853
J Med Imaging (Bellingham). 2015 Oct;2(4):041002
pubmed: 26251842
Eur Radiol. 2017 Mar;27(3):1012-1020
pubmed: 27380902
Sci Rep. 2017 Mar 23;7(1):358
pubmed: 28336974
Blood. 2018 Apr 12;131(15):1654-1665
pubmed: 29500175
Sci Rep. 2019 Feb 4;9(1):1322
pubmed: 30718585
Eur J Nucl Med Mol Imaging. 2019 Dec;46(13):2737-2745
pubmed: 31690962
Med Phys. 2014 Dec;41(12):121904
pubmed: 25471964
Eur J Nucl Med Mol Imaging. 2019 Dec;46(13):2656-2672
pubmed: 31214791
Blood. 2018 May 31;131(22):2413-2425
pubmed: 29449275
Leuk Lymphoma. 2020 Feb;61(2):251-254
pubmed: 31942807
Am J Hematol. 2018 May;93(5):704-715
pubmed: 29634090
Adv Hematol. 2011;2011:439456
pubmed: 21318045
Phys Med Biol. 2017 Jan 21;62(2):652-668
pubmed: 28033121
J Intern Med. 2018 Oct;284(4):358-376
pubmed: 29989234
Oncol Res. 2009;17(7):331-7
pubmed: 19408578
Eur J Nucl Med Mol Imaging. 2018 Feb;45(2):207-217
pubmed: 28944403
J Nucl Med. 2018 Aug;59(8):1321-1328
pubmed: 29301932