Tubular Structure-Aware Convolutional Neural Networks for Organ at Risks Segmentation in Cervical Cancer Radiotherapy

Xinran Wu, Ming Cui, Yuhua Gao, Deyu Sun, He Ma, Erlei Zhang, Yaoqin Xie, Nazar Zaki, Wenjian Qin

Research output: Chapter in Book/Report/Conference proceedingConference contribution

1 Citation (Scopus)

Abstract

Cervical cancer is the most frequent cancer type among women worldwide and radiotherapy is the major clinical treatment. Organs in the radiation field are called Organ at Risks (OARs), which are prone to irreversible damage during radiotherapy. Therefore, accurate delineation of OARs is a critical step in ensuring radiotherapy dosimetry accuracy. However, currently existing deep learning-based cervical cancer OARs segmentation methods do not make full advantage of anatomical information. In this paper, we develop a novel tubular structure-aware deep convolutional network method integrating the tubular anatomical morphological features into a model for colon, small intestine and rectum in cervical cancer OARs. Firstly, a tubular filter based on variable annular Gaussian kernel and gradient detection was used to produce the tubular feature map. Secondary, tubular feature map concatenated with original image was input into the nnU-Net network for anatomical morphological information learning. Finally, we evaluated our proposed method on the clinical collection datasets with brachytherapy. Compared to the baseline model and state-of-the-art model, DSC and Recall were improved and the relative volume error (RVE) was reduced for the OARs with a tubular shape.

Original languageEnglish
Title of host publicationComputational Mathematics Modeling in Cancer Analysis - 1st International Workshop, CMMCA 2022, Held in Conjunction with MICCAI 2022, Proceedings
EditorsWenjian Qin, Nazar Zaki, Fa Zhang, Jia Wu, Fan Yang
PublisherSpringer Science and Business Media Deutschland GmbH
Pages131-140
Number of pages10
ISBN (Print)9783031172656
DOIs
Publication statusPublished - 2022
Event1st International Workshop on Computational Mathematics Modeling in Cancer Analysis, CMMCA 2022, held in conjunction with the 25th International Conference on Medical Image Computing and Computer Assisted Intervention, MICCAI 2022 - Virtual, Online
Duration: Sept 18 2022Sept 18 2022

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume13574 LNCS
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Conference

Conference1st International Workshop on Computational Mathematics Modeling in Cancer Analysis, CMMCA 2022, held in conjunction with the 25th International Conference on Medical Image Computing and Computer Assisted Intervention, MICCAI 2022
CityVirtual, Online
Period9/18/229/18/22

Keywords

  • Anatomical information
  • Cervical cancer organ at risk
  • Tubular filter

ASJC Scopus subject areas

  • Theoretical Computer Science
  • General Computer Science

Fingerprint

Dive into the research topics of 'Tubular Structure-Aware Convolutional Neural Networks for Organ at Risks Segmentation in Cervical Cancer Radiotherapy'. Together they form a unique fingerprint.

Cite this