Dose-volume-based evaluation of convolutional neural network-based auto-segmentation of thoracic organs at risk
Background and purpose: The geometrical accuracy of auto-segmentation using convolutional neural networks (CNNs) has been demonstrated. This study aimed to investigate the dose-volume impact of differences between automatic and manual OARs for locally advanced (LA) and peripherally located early-sta...
Main Authors: | , , , , , , , |
---|---|
Format: | Article |
Language: | English |
Published: |
Elsevier
2022-07-01
|
Series: | Physics and Imaging in Radiation Oncology |
Subjects: | |
Online Access: | http://www.sciencedirect.com/science/article/pii/S2405631622000677 |
_version_ | 1811277042272960512 |
---|---|
author | Noémie Johnston Jeffrey De Rycke Yolande Lievens Marc van Eijkeren Jan Aelterman Eva Vandersmissen Stephan Ponte Barbara Vanderstraeten |
author_facet | Noémie Johnston Jeffrey De Rycke Yolande Lievens Marc van Eijkeren Jan Aelterman Eva Vandersmissen Stephan Ponte Barbara Vanderstraeten |
author_sort | Noémie Johnston |
collection | DOAJ |
description | Background and purpose: The geometrical accuracy of auto-segmentation using convolutional neural networks (CNNs) has been demonstrated. This study aimed to investigate the dose-volume impact of differences between automatic and manual OARs for locally advanced (LA) and peripherally located early-stage (ES) non-small cell lung cancer (NSCLC). Material and methods: A single CNN was created for automatic delineation of the heart, lungs, main left and right bronchus, esophagus, spinal cord and trachea using 55/10/40 patients for training/validation/testing. Dice score coefficient (DSC) and 95th percentile Hausdorff distance (HD95) were used for geometrical analysis. A new treatment plan based on the auto-segmented OARs was created for each test patient using 3D for ES-NSCLC (SBRT, 3–8 fractions) and IMRT for LA-NSCLC (24–35 fractions). The correlation between geometrical metrics and dose-volume differences was investigated. Results: The average (±1 SD) DSC and HD95 were 0.82 ± 0.07 and 16.2 ± 22.4 mm, while the average dose-volume differences were 0.5 ± 1.5 Gy (ES) and 1.5 ± 2.8 Gy (LA). The geometrical metrics did not correlate with the observed dose-volume differences (average Pearson for DSC: −0.27 ± 0.18 (ES) and −0.09 ± 0.12 (LA); HD95: 0.1 ± 0.3 mm (ES) and 0.2 ± 0.2 mm (LA)). Conclusions: After post-processing, manual adjustments of automatic contours are only needed for clinically relevant OARs situated close to the tumor or within an entry or exit beam e.g., the heart and the esophagus for LA-NSCLC and the bronchi for ES-NSCLC. The lungs do not need to be checked further in detail. |
first_indexed | 2024-04-13T00:08:36Z |
format | Article |
id | doaj.art-e7bfeb5bed11459dba361fc568572472 |
institution | Directory Open Access Journal |
issn | 2405-6316 |
language | English |
last_indexed | 2024-04-13T00:08:36Z |
publishDate | 2022-07-01 |
publisher | Elsevier |
record_format | Article |
series | Physics and Imaging in Radiation Oncology |
spelling | doaj.art-e7bfeb5bed11459dba361fc5685724722022-12-22T03:11:10ZengElsevierPhysics and Imaging in Radiation Oncology2405-63162022-07-0123109117Dose-volume-based evaluation of convolutional neural network-based auto-segmentation of thoracic organs at riskNoémie Johnston0Jeffrey De Rycke1Yolande Lievens2Marc van Eijkeren3Jan Aelterman4Eva Vandersmissen5Stephan Ponte6Barbara Vanderstraeten7Centre Hospitalier Universitaire de Liège, Service de Radiothérapie, Liège, BelgiumGhent University, Faculty of Medicine and Health Sciences, Department of Human Structure and Repair, Gent, BelgiumGhent University, Faculty of Medicine and Health Sciences, Department of Human Structure and Repair, Gent, Belgium; Ghent University Hospital, Department of Radiotherapy-Oncology, Gent, BelgiumGhent University, Faculty of Medicine and Health Sciences, Department of Human Structure and Repair, Gent, Belgium; Ghent University Hospital, Department of Radiotherapy-Oncology, Gent, BelgiumGhent University, Department of Physics and Astronomy, Ghent University Centre for X-ray Tomography, Gent, Belgium; Ghent University, Department TELIN / IMEC, Image Processing Interpretation Group, Gent, BelgiumAgfa NV, Radiology Solutions R&D, Mortsel, BelgiumCentre Hospitalier Universitaire de Liège, Service de Radiothérapie, Liège, BelgiumGhent University, Faculty of Medicine and Health Sciences, Department of Human Structure and Repair, Gent, Belgium; Ghent University Hospital, Department of Radiotherapy-Oncology, Gent, Belgium; Corresponding author at: Ghent University Hospital, Department of Radiotherapy-Oncology, RTP Ingang 98, Corneel Heymanslaan 10, B-9000 Gent, Belgium.Background and purpose: The geometrical accuracy of auto-segmentation using convolutional neural networks (CNNs) has been demonstrated. This study aimed to investigate the dose-volume impact of differences between automatic and manual OARs for locally advanced (LA) and peripherally located early-stage (ES) non-small cell lung cancer (NSCLC). Material and methods: A single CNN was created for automatic delineation of the heart, lungs, main left and right bronchus, esophagus, spinal cord and trachea using 55/10/40 patients for training/validation/testing. Dice score coefficient (DSC) and 95th percentile Hausdorff distance (HD95) were used for geometrical analysis. A new treatment plan based on the auto-segmented OARs was created for each test patient using 3D for ES-NSCLC (SBRT, 3–8 fractions) and IMRT for LA-NSCLC (24–35 fractions). The correlation between geometrical metrics and dose-volume differences was investigated. Results: The average (±1 SD) DSC and HD95 were 0.82 ± 0.07 and 16.2 ± 22.4 mm, while the average dose-volume differences were 0.5 ± 1.5 Gy (ES) and 1.5 ± 2.8 Gy (LA). The geometrical metrics did not correlate with the observed dose-volume differences (average Pearson for DSC: −0.27 ± 0.18 (ES) and −0.09 ± 0.12 (LA); HD95: 0.1 ± 0.3 mm (ES) and 0.2 ± 0.2 mm (LA)). Conclusions: After post-processing, manual adjustments of automatic contours are only needed for clinically relevant OARs situated close to the tumor or within an entry or exit beam e.g., the heart and the esophagus for LA-NSCLC and the bronchi for ES-NSCLC. The lungs do not need to be checked further in detail.http://www.sciencedirect.com/science/article/pii/S2405631622000677Lung cancerRadiotherapyTreatment planningDoseVolumeDice |
spellingShingle | Noémie Johnston Jeffrey De Rycke Yolande Lievens Marc van Eijkeren Jan Aelterman Eva Vandersmissen Stephan Ponte Barbara Vanderstraeten Dose-volume-based evaluation of convolutional neural network-based auto-segmentation of thoracic organs at risk Physics and Imaging in Radiation Oncology Lung cancer Radiotherapy Treatment planning Dose Volume Dice |
title | Dose-volume-based evaluation of convolutional neural network-based auto-segmentation of thoracic organs at risk |
title_full | Dose-volume-based evaluation of convolutional neural network-based auto-segmentation of thoracic organs at risk |
title_fullStr | Dose-volume-based evaluation of convolutional neural network-based auto-segmentation of thoracic organs at risk |
title_full_unstemmed | Dose-volume-based evaluation of convolutional neural network-based auto-segmentation of thoracic organs at risk |
title_short | Dose-volume-based evaluation of convolutional neural network-based auto-segmentation of thoracic organs at risk |
title_sort | dose volume based evaluation of convolutional neural network based auto segmentation of thoracic organs at risk |
topic | Lung cancer Radiotherapy Treatment planning Dose Volume Dice |
url | http://www.sciencedirect.com/science/article/pii/S2405631622000677 |
work_keys_str_mv | AT noemiejohnston dosevolumebasedevaluationofconvolutionalneuralnetworkbasedautosegmentationofthoracicorgansatrisk AT jeffreyderycke dosevolumebasedevaluationofconvolutionalneuralnetworkbasedautosegmentationofthoracicorgansatrisk AT yolandelievens dosevolumebasedevaluationofconvolutionalneuralnetworkbasedautosegmentationofthoracicorgansatrisk AT marcvaneijkeren dosevolumebasedevaluationofconvolutionalneuralnetworkbasedautosegmentationofthoracicorgansatrisk AT janaelterman dosevolumebasedevaluationofconvolutionalneuralnetworkbasedautosegmentationofthoracicorgansatrisk AT evavandersmissen dosevolumebasedevaluationofconvolutionalneuralnetworkbasedautosegmentationofthoracicorgansatrisk AT stephanponte dosevolumebasedevaluationofconvolutionalneuralnetworkbasedautosegmentationofthoracicorgansatrisk AT barbaravanderstraeten dosevolumebasedevaluationofconvolutionalneuralnetworkbasedautosegmentationofthoracicorgansatrisk |