Clinical Validation of a Deep-Learning Segmentation Software in Head and Neck: An Early Analysis in a Developing Radiation Oncology Center.
auto-contouring
deep-learning
head and neck
radiotherapy artificial intelligence
Journal
International journal of environmental research and public health
ISSN: 1660-4601
Titre abrégé: Int J Environ Res Public Health
Pays: Switzerland
ID NLM: 101238455
Informations de publication
Date de publication:
25 07 2022
25 07 2022
Historique:
received:
11
04
2022
revised:
12
07
2022
accepted:
20
07
2022
entrez:
28
7
2022
pubmed:
29
7
2022
medline:
30
7
2022
Statut:
epublish
Résumé
Organs at risk (OARs) delineation is a crucial step of radiotherapy (RT) treatment planning workflow. Time-consuming and inter-observer variability are main issues in manual OAR delineation, mainly in the head and neck (H & N) district. Deep-learning based auto-segmentation is a promising strategy to improve OARs contouring in radiotherapy departments. A comparison of deep-learning-generated auto-contours (AC) with manual contours (MC) was performed by three expert radiation oncologists from a single center. Planning computed tomography (CT) scans of patients undergoing RT treatments for H&N cancers were considered. CT scans were processed by Limbus Contour auto-segmentation software, a commercial deep-learning auto-segmentation based software to generate AC. H&N protocol was used to perform AC, with the structure set consisting of bilateral brachial plexus, brain, brainstem, bilateral cochlea, pharyngeal constrictors, eye globes, bilateral lens, mandible, optic chiasm, bilateral optic nerves, oral cavity, bilateral parotids, spinal cord, bilateral submandibular glands, lips and thyroid. Manual revision of OARs was performed according to international consensus guidelines. The AC and MC were compared using the Dice similarity coefficient (DSC) and 95% Hausdorff distance transform (DT). A total of 274 contours obtained by processing CT scans were included in the analysis. The highest values of DSC were obtained for the brain (DSC 1.00), left and right eye globes and the mandible (DSC 0.98). The structures with greater MC editing were optic chiasm, optic nerves and cochleae. In this preliminary analysis, deep-learning auto-segmentation seems to provide acceptable H&N OAR delineations. For less accurate organs, AC could be considered a starting point for review and manual adjustment. Our results suggest that AC could become a useful time-saving tool to optimize workload and resources in RT departments.
Sections du résumé
BACKGROUND
Organs at risk (OARs) delineation is a crucial step of radiotherapy (RT) treatment planning workflow. Time-consuming and inter-observer variability are main issues in manual OAR delineation, mainly in the head and neck (H & N) district. Deep-learning based auto-segmentation is a promising strategy to improve OARs contouring in radiotherapy departments. A comparison of deep-learning-generated auto-contours (AC) with manual contours (MC) was performed by three expert radiation oncologists from a single center.
METHODS
Planning computed tomography (CT) scans of patients undergoing RT treatments for H&N cancers were considered. CT scans were processed by Limbus Contour auto-segmentation software, a commercial deep-learning auto-segmentation based software to generate AC. H&N protocol was used to perform AC, with the structure set consisting of bilateral brachial plexus, brain, brainstem, bilateral cochlea, pharyngeal constrictors, eye globes, bilateral lens, mandible, optic chiasm, bilateral optic nerves, oral cavity, bilateral parotids, spinal cord, bilateral submandibular glands, lips and thyroid. Manual revision of OARs was performed according to international consensus guidelines. The AC and MC were compared using the Dice similarity coefficient (DSC) and 95% Hausdorff distance transform (DT).
RESULTS
A total of 274 contours obtained by processing CT scans were included in the analysis. The highest values of DSC were obtained for the brain (DSC 1.00), left and right eye globes and the mandible (DSC 0.98). The structures with greater MC editing were optic chiasm, optic nerves and cochleae.
CONCLUSIONS
In this preliminary analysis, deep-learning auto-segmentation seems to provide acceptable H&N OAR delineations. For less accurate organs, AC could be considered a starting point for review and manual adjustment. Our results suggest that AC could become a useful time-saving tool to optimize workload and resources in RT departments.
Identifiants
pubmed: 35897425
pii: ijerph19159057
doi: 10.3390/ijerph19159057
pmc: PMC9329735
pii:
doi:
Types de publication
Journal Article
Langues
eng
Sous-ensembles de citation
IM
Références
Phys Med. 2021 May;85:175-191
pubmed: 34022660
J Med Imaging Radiat Oncol. 2016 Jun;60(3):393-406
pubmed: 27170216
Radiat Oncol. 2012 Mar 13;7:32
pubmed: 22414264
Radiother Oncol. 2015 Feb;114(2):230-8
pubmed: 25701297
Radiother Oncol. 2018 Jul;128(1):37-43
pubmed: 29548560
J Med Imaging (Bellingham). 2018 Jan;5(1):015006
pubmed: 29487883
Front Oncol. 2021 Nov 12;11:761393
pubmed: 34868976
Radiother Oncol. 2005 Oct;77(1):25-31
pubmed: 15919126
Ann Oncol. 2020 Nov;31(11):1462-1475
pubmed: 33239190
Mol Oncol. 2020 Jul;14(7):1470-1491
pubmed: 32536001
Radiat Oncol. 2012 Sep 18;7:160
pubmed: 22989046
Front Oncol. 2021 Jun 07;11:626499
pubmed: 34164335
J Med Internet Res. 2021 Jul 12;23(7):e26151
pubmed: 34255661
Radiother Oncol. 2020 Mar;144:152-158
pubmed: 31812930
J Contemp Brachytherapy. 2020 Oct;12(5):497-500
pubmed: 33299440
Psychooncology. 2022 Mar;31(3):478-485
pubmed: 34591369
Tech Innov Patient Support Radiat Oncol. 2017 Jul 21;3-4:1-6
pubmed: 32095559
Radiat Oncol J. 2019 Jun;37(2):134-142
pubmed: 31266293
Med Phys. 2017 Feb;44(2):547-557
pubmed: 28205307
Radiother Oncol. 2020 Dec;153:205-212
pubmed: 33075394
Cancer Manag Res. 2017 Aug 18;9:363-371
pubmed: 28860859
Phys Imaging Radiat Oncol. 2020 Nov 30;16:144-148
pubmed: 33458358
J Clin Oncol. 2007 Nov 1;25(31):4873-9
pubmed: 17971582
Radiother Oncol. 2020 Jan;142:115-123
pubmed: 31653573
Radiother Oncol. 2015 Oct;117(1):83-90
pubmed: 26277855
Radiat Oncol. 2021 Jun 8;16(1):101
pubmed: 34103062
Head Neck. 2019 Apr;41(4):959-966
pubmed: 30620435
J Natl Compr Canc Netw. 2020 Jul;18(7):873-898
pubmed: 32634781
Med Phys. 2020 Sep;47(9):e929-e950
pubmed: 32510603
Acta Oncol. 2021 Feb;60(2):229-236
pubmed: 32988249
Acta Oncol. 2016 Jul;55(7):799-806
pubmed: 27248772