Utility of the deep learning technique for the diagnosis of orbital invasion on CT in patients with a nasal or sinonasal tumor.
Deep learning
Head and neck
Nasal or sinonasal tumor
Orbital invasion
Periorbita
Transfer learning
Journal
Cancer imaging : the official publication of the International Cancer Imaging Society
ISSN: 1470-7330
Titre abrégé: Cancer Imaging
Pays: England
ID NLM: 101172931
Informations de publication
Date de publication:
22 Sep 2022
22 Sep 2022
Historique:
received:
20
04
2022
accepted:
14
09
2022
entrez:
22
9
2022
pubmed:
23
9
2022
medline:
28
9
2022
Statut:
epublish
Résumé
In nasal or sinonasal tumors, orbital invasion beyond periorbita by the tumor is one of the important criteria in the selection of the surgical procedure. We investigated the usefulness of the convolutional neural network (CNN)-based deep learning technique for the diagnosis of orbital invasion, using computed tomography (CT) images. A total of 168 lesions with malignant nasal or sinonasal tumors were divided into a training dataset (n = 119) and a test dataset (n = 49). The final diagnosis (invasion-positive or -negative) was determined by experienced radiologists who carefully reviewed all of the CT images. In a CNN-based deep learning analysis, a slice of the square target region that included the orbital bone wall was extracted and fed into a deep-learning training session to create a diagnostic model using transfer learning with the Visual Geometry Group 16 (VGG16) model. The test dataset was subsequently tested in CNN-based diagnostic models and by two other radiologists who were not specialized in head and neck radiology. At approx. 2 months after the first reading session, two radiologists again reviewed all of the images in the test dataset, referring to the diagnoses provided by the trained CNN-based diagnostic model. The diagnostic accuracy was 0.92 by the CNN-based diagnostic models, whereas the diagnostic accuracies by the two radiologists at the first reading session were 0.49 and 0.45, respectively. In the second reading session by two radiologists (diagnosing with the assistance by the CNN-based diagnostic model), marked elevations of the diagnostic accuracy were observed (0.94 and 1.00, respectively). The CNN-based deep learning technique can be a useful support tool in assessing the presence of orbital invasion on CT images, especially for non-specialized radiologists.
Sections du résumé
BACKGROUND
BACKGROUND
In nasal or sinonasal tumors, orbital invasion beyond periorbita by the tumor is one of the important criteria in the selection of the surgical procedure. We investigated the usefulness of the convolutional neural network (CNN)-based deep learning technique for the diagnosis of orbital invasion, using computed tomography (CT) images.
METHODS
METHODS
A total of 168 lesions with malignant nasal or sinonasal tumors were divided into a training dataset (n = 119) and a test dataset (n = 49). The final diagnosis (invasion-positive or -negative) was determined by experienced radiologists who carefully reviewed all of the CT images. In a CNN-based deep learning analysis, a slice of the square target region that included the orbital bone wall was extracted and fed into a deep-learning training session to create a diagnostic model using transfer learning with the Visual Geometry Group 16 (VGG16) model. The test dataset was subsequently tested in CNN-based diagnostic models and by two other radiologists who were not specialized in head and neck radiology. At approx. 2 months after the first reading session, two radiologists again reviewed all of the images in the test dataset, referring to the diagnoses provided by the trained CNN-based diagnostic model.
RESULTS
RESULTS
The diagnostic accuracy was 0.92 by the CNN-based diagnostic models, whereas the diagnostic accuracies by the two radiologists at the first reading session were 0.49 and 0.45, respectively. In the second reading session by two radiologists (diagnosing with the assistance by the CNN-based diagnostic model), marked elevations of the diagnostic accuracy were observed (0.94 and 1.00, respectively).
CONCLUSION
CONCLUSIONS
The CNN-based deep learning technique can be a useful support tool in assessing the presence of orbital invasion on CT images, especially for non-specialized radiologists.
Identifiants
pubmed: 36138422
doi: 10.1186/s40644-022-00492-0
pii: 10.1186/s40644-022-00492-0
pmc: PMC9502604
doi:
Types de publication
Journal Article
Langues
eng
Sous-ensembles de citation
IM
Pagination
52Informations de copyright
© 2022. The Author(s).
Références
Head Neck. 2020 Aug;42(8):2002-2012
pubmed: 32124507
Diagn Interv Imaging. 2019 Apr;100(4):251-257
pubmed: 30819638
Laryngoscope. 1999 Feb;109(2 Pt 1):230-5
pubmed: 10890772
Cancers (Basel). 2021 Oct 01;13(19):
pubmed: 34638447
Radiology. 2017 Dec;285(3):923-931
pubmed: 28678669
Diagn Interv Imaging. 2017 May;98(5):369-371
pubmed: 28477770
Eur Radiol. 2021 May;31(5):2866-2876
pubmed: 33125556
Head Neck. 2005 Jul;27(7):575-84
pubmed: 15825201
J Clin Oncol. 2020 Apr 20;38(12):1304-1311
pubmed: 31815574
Paediatr Perinat Epidemiol. 1988 Jan;2(1):43-58
pubmed: 3070483
Head Neck. 2000 Aug;22(5):456-62
pubmed: 10897104
Radiology. 2021 Jul;300(1):120-129
pubmed: 33944629
J Neurol Surg B Skull Base. 2020 Aug;81(4):369-375
pubmed: 33072478
Arch Otolaryngol Head Neck Surg. 1988 Jun;114(6):632-4
pubmed: 2835065
Head Neck. 2008 Feb;30(2):242-50
pubmed: 18023029
Front Oncol. 2021 Jun 11;11:654685
pubmed: 34178641
PLoS One. 2014 Aug 27;9(8):e105682
pubmed: 25162396
Facial Plast Surg. 2014 Oct;30(5):487-508
pubmed: 25397705
Radiographics. 2017 Nov-Dec;37(7):2113-2131
pubmed: 29131760
EJNMMI Res. 2021 Jan 25;11(1):10
pubmed: 33492478