volume 95 issue 1134

The efficacy of 18F-FDG-PET-based radiomic and deep-learning features using a machine-learning approach to predict the pathological risk subtypes of thymic epithelial tumors

Masatoyo Nakajo 1
Aya Takeda 2
Akie Katsuki 3
Megumi Jinguji 1
Kazuyuki Ohmura 3
Atsushi Tani 1
Masami Sato 2
Takashi Yoshiura 1
Publication typeJournal Article
Publication date2022-03-21
scimago Q1
wos Q1
SJR0.875
CiteScore5.6
Impact factor3.4
ISSN00071285, 1748880X
PubMed ID:  35312337
General Medicine
Radiology, Nuclear Medicine and imaging
Abstract
Objective:

To examine whether the machine-learning approach using 18-fludeoxyglucose positron emission tomography (18F-FDG-PET)-based radiomic and deep-learning features is useful for predicting the pathological risk subtypes of thymic epithelial tumors (TETs).

Methods:

This retrospective study included 79 TET [27 low-risk thymomas (types A, AB and B1), 31 high-risk thymomas (types B2 and B3) and 21 thymic carcinomas] patients who underwent pre-therapeutic 18F-FDG-PET/CT. High-risk TETs (high-risk thymomas and thymic carcinomas) were 52 patients. The 107 PET-based radiomic features, including SUV-related parameters [maximum SUV (SUVmax), metabolic tumor volume (MTV), and total lesion glycolysis (TLG)] and 1024 deep-learning features extracted from the convolutional neural network were used to predict the pathological risk subtypes of TETs using six different machine-learning algorithms. The area under the curves (AUCs) were calculated to compare the predictive performances.

Results:

SUV-related parameters yielded the following AUCs for predicting thymic carcinomas: SUVmax 0.713, MTV 0.442, and TLG 0.479 or high-risk TETs: SUVmax 0.673, MTV 0.533, and TLG 0.539. The best-performing algorithm was the logistic regression model for predicting thymic carcinomas (AUC 0.900, accuracy 81.0%), and the random forest (RF) model for high-risk TETs (AUC 0.744, accuracy 72.2%). The AUC was significantly higher in the logistic regression model than three SUV-related parameters for predicting thymic carcinomas, and in the RF model than MTV and TLG for predicting high-risk TETs (each; p < 0.05).

Conclusion:

18F-FDG-PET-based radiomic analysis using a machine-learning approach may be useful for predicting the pathological risk subtypes of TETs.

Advances in knowledge:

Machine-learning approach using 18F-FDG-PET-based radiomic features has the potential to predict the pathological risk subtypes of TETs.

Found 
Found 

Top-30

Journals

1
2
Diagnostics
2 publications, 10.53%
BMC Cancer
2 publications, 10.53%
Frontiers in Oncology
2 publications, 10.53%
Mediastinum
2 publications, 10.53%
Seminars in Cancer Biology
1 publication, 5.26%
European Radiology
1 publication, 5.26%
Japanese Journal of Radiology
1 publication, 5.26%
BMC Medical Imaging
1 publication, 5.26%
Nuclear Medicine Communications
1 publication, 5.26%
Journal of Biophotonics
1 publication, 5.26%
Molecular Diagnosis and Therapy
1 publication, 5.26%
BMC Medicine
1 publication, 5.26%
Bioengineering
1 publication, 5.26%
European Journal of Medical Research
1 publication, 5.26%
1
2

Publishers

1
2
3
4
5
6
7
8
Springer Nature
8 publications, 42.11%
MDPI
3 publications, 15.79%
Frontiers Media S.A.
2 publications, 10.53%
AME Publishing Company
2 publications, 10.53%
Elsevier
1 publication, 5.26%
Ovid Technologies (Wolters Kluwer Health)
1 publication, 5.26%
Wiley
1 publication, 5.26%
1
2
3
4
5
6
7
8
  • We do not take into account publications without a DOI.
  • Statistics recalculated weekly.

Are you a researcher?

Create a profile to get free access to personal recommendations for colleagues and new articles.
Metrics
19
Share
Cite this
GOST |
Cite this
GOST Copy
Nakajo M. et al. The efficacy of 18F-FDG-PET-based radiomic and deep-learning features using a machine-learning approach to predict the pathological risk subtypes of thymic epithelial tumors // British Journal of Radiology. 2022. Vol. 95. No. 1134.
GOST all authors (up to 50) Copy
Nakajo M., Takeda A., Katsuki A., Jinguji M., Ohmura K., Tani A., Sato M., Yoshiura T. The efficacy of 18F-FDG-PET-based radiomic and deep-learning features using a machine-learning approach to predict the pathological risk subtypes of thymic epithelial tumors // British Journal of Radiology. 2022. Vol. 95. No. 1134.
RIS |
Cite this
RIS Copy
TY - JOUR
DO - 10.1259/bjr.20211050
UR - https://doi.org/10.1259/bjr.20211050
TI - The efficacy of 18F-FDG-PET-based radiomic and deep-learning features using a machine-learning approach to predict the pathological risk subtypes of thymic epithelial tumors
T2 - British Journal of Radiology
AU - Nakajo, Masatoyo
AU - Takeda, Aya
AU - Katsuki, Akie
AU - Jinguji, Megumi
AU - Ohmura, Kazuyuki
AU - Tani, Atsushi
AU - Sato, Masami
AU - Yoshiura, Takashi
PY - 2022
DA - 2022/03/21
PB - British Institute of Radiology
IS - 1134
VL - 95
PMID - 35312337
SN - 0007-1285
SN - 1748-880X
ER -
BibTex
Cite this
BibTex (up to 50 authors) Copy
@article{2022_Nakajo,
author = {Masatoyo Nakajo and Aya Takeda and Akie Katsuki and Megumi Jinguji and Kazuyuki Ohmura and Atsushi Tani and Masami Sato and Takashi Yoshiura},
title = {The efficacy of 18F-FDG-PET-based radiomic and deep-learning features using a machine-learning approach to predict the pathological risk subtypes of thymic epithelial tumors},
journal = {British Journal of Radiology},
year = {2022},
volume = {95},
publisher = {British Institute of Radiology},
month = {mar},
url = {https://doi.org/10.1259/bjr.20211050},
number = {1134},
doi = {10.1259/bjr.20211050}
}