Skip to main content

Lund University Publications

LUND UNIVERSITY LIBRARIES

Pelvic U-Net : multi-label semantic segmentation of pelvic organs at risk for radiation therapy anal cancer patients using a deeply supervised shuffle attention convolutional neural network

Lempart, Michael LU ; Nilsson, Martin P. LU ; Scherman, Jonas ; Gustafsson, Christian Jamtheim LU ; Nilsson, Mikael LU ; Alkner, Sara LU ; Engleson, Jens LU ; Adrian, Gabriel LU orcid ; Munck af Rosenschöld, Per LU orcid and Olsson, Lars E. LU orcid (2022) In Radiation Oncology 17(1).
Abstract

Background: Delineation of organs at risk (OAR) for anal cancer radiation therapy treatment planning is a manual and time-consuming process. Deep learning-based methods can accelerate and partially automate this task. The aim of this study was to develop and evaluate a deep learning model for automated and improved segmentations of OAR in the pelvic region. Methods: A 3D, deeply supervised U-Net architecture with shuffle attention, referred to as Pelvic U-Net, was trained on 143 computed tomography (CT) volumes, to segment OAR in the pelvic region, such as total bone marrow, rectum, bladder, and bowel structures. Model predictions were evaluated on an independent test dataset (n = 15) using the Dice similarity coefficient (DSC), the... (More)

Background: Delineation of organs at risk (OAR) for anal cancer radiation therapy treatment planning is a manual and time-consuming process. Deep learning-based methods can accelerate and partially automate this task. The aim of this study was to develop and evaluate a deep learning model for automated and improved segmentations of OAR in the pelvic region. Methods: A 3D, deeply supervised U-Net architecture with shuffle attention, referred to as Pelvic U-Net, was trained on 143 computed tomography (CT) volumes, to segment OAR in the pelvic region, such as total bone marrow, rectum, bladder, and bowel structures. Model predictions were evaluated on an independent test dataset (n = 15) using the Dice similarity coefficient (DSC), the 95th percentile of the Hausdorff distance (HD95), and the mean surface distance (MSD). In addition, three experienced radiation oncologists rated model predictions on a scale between 1–4 (excellent, good, acceptable, not acceptable). Model performance was also evaluated with respect to segmentation time, by comparing complete manual delineation time against model prediction time without and with manual correction of the predictions. Furthermore, dosimetric implications to treatment plans were evaluated using different dose-volume histogram (DVH) indices. Results: Without any manual corrections, mean DSC values of 97%, 87% and 94% were found for total bone marrow, rectum, and bladder. Mean DSC values for bowel cavity, all bowel, small bowel, and large bowel were 95%, 91%, 87% and 81%, respectively. Total bone marrow, bladder, and bowel cavity segmentations derived from our model were rated excellent (89%, 93%, 42%), good (9%, 5%, 42%), or acceptable (2%, 2%, 16%) on average. For almost all the evaluated DVH indices, no significant difference between model predictions and manual delineations was found. Delineation time per patient could be reduced from 40 to 12 min, including manual corrections of model predictions, and to 4 min without corrections. Conclusions: Our Pelvic U-Net led to credible and clinically applicable OAR segmentations and showed improved performance compared to previous studies. Even though manual adjustments were needed for some predicted structures, segmentation time could be reduced by 70% on average. This allows for an accelerated radiation therapy treatment planning workflow for anal cancer patients.

(Less)
Please use this url to cite or link to this publication:
author
; ; ; ; ; ; ; ; and
organization
publishing date
type
Contribution to journal
publication status
published
subject
keywords
Anal cancer, Deep learning, Organs at risk, Radiation therapy, Semantic segmentation
in
Radiation Oncology
volume
17
issue
1
article number
114
publisher
BioMed Central (BMC)
external identifiers
  • scopus:85132970884
  • pmid:35765038
ISSN
1748-717X
DOI
10.1186/s13014-022-02088-1
language
English
LU publication?
yes
id
a3288c56-c33e-4260-9da8-52d75a795e12
date added to LUP
2022-09-23 15:29:41
date last changed
2024-06-11 01:27:42
@article{a3288c56-c33e-4260-9da8-52d75a795e12,
  abstract     = {{<p>Background: Delineation of organs at risk (OAR) for anal cancer radiation therapy treatment planning is a manual and time-consuming process. Deep learning-based methods can accelerate and partially automate this task. The aim of this study was to develop and evaluate a deep learning model for automated and improved segmentations of OAR in the pelvic region. Methods: A 3D, deeply supervised U-Net architecture with shuffle attention, referred to as Pelvic U-Net, was trained on 143 computed tomography (CT) volumes, to segment OAR in the pelvic region, such as total bone marrow, rectum, bladder, and bowel structures. Model predictions were evaluated on an independent test dataset (n = 15) using the Dice similarity coefficient (DSC), the 95th percentile of the Hausdorff distance (HD<sub>95</sub>), and the mean surface distance (MSD). In addition, three experienced radiation oncologists rated model predictions on a scale between 1–4 (excellent, good, acceptable, not acceptable). Model performance was also evaluated with respect to segmentation time, by comparing complete manual delineation time against model prediction time without and with manual correction of the predictions. Furthermore, dosimetric implications to treatment plans were evaluated using different dose-volume histogram (DVH) indices. Results: Without any manual corrections, mean DSC values of 97%, 87% and 94% were found for total bone marrow, rectum, and bladder. Mean DSC values for bowel cavity, all bowel, small bowel, and large bowel were 95%, 91%, 87% and 81%, respectively. Total bone marrow, bladder, and bowel cavity segmentations derived from our model were rated excellent (89%, 93%, 42%), good (9%, 5%, 42%), or acceptable (2%, 2%, 16%) on average. For almost all the evaluated DVH indices, no significant difference between model predictions and manual delineations was found. Delineation time per patient could be reduced from 40 to 12 min, including manual corrections of model predictions, and to 4 min without corrections. Conclusions: Our Pelvic U-Net led to credible and clinically applicable OAR segmentations and showed improved performance compared to previous studies. Even though manual adjustments were needed for some predicted structures, segmentation time could be reduced by 70% on average. This allows for an accelerated radiation therapy treatment planning workflow for anal cancer patients.</p>}},
  author       = {{Lempart, Michael and Nilsson, Martin P. and Scherman, Jonas and Gustafsson, Christian Jamtheim and Nilsson, Mikael and Alkner, Sara and Engleson, Jens and Adrian, Gabriel and Munck af Rosenschöld, Per and Olsson, Lars E.}},
  issn         = {{1748-717X}},
  keywords     = {{Anal cancer; Deep learning; Organs at risk; Radiation therapy; Semantic segmentation}},
  language     = {{eng}},
  number       = {{1}},
  publisher    = {{BioMed Central (BMC)}},
  series       = {{Radiation Oncology}},
  title        = {{Pelvic U-Net : multi-label semantic segmentation of pelvic organs at risk for radiation therapy anal cancer patients using a deeply supervised shuffle attention convolutional neural network}},
  url          = {{http://dx.doi.org/10.1186/s13014-022-02088-1}},
  doi          = {{10.1186/s13014-022-02088-1}},
  volume       = {{17}},
  year         = {{2022}},
}