The Squeeze &amp; Excitation Normalization Based nnU-Net for Segmenting Head &amp; Neck Tumors

XIE Juanying; PENG Ying; WANG Mingzhao

doi:10.23919/cje.2022.00.306

Volume 33 Issue 3

May 2024

Turn off MathJax

Article Contents

Article Navigation > Chinese Journal of Electronics > 2024 > 33(3): 766-775

Juanying XIE, Ying PENG, and Mingzhao WANG, “The Squeeze & Excitation Normalization Based nnU-Net for Segmenting Head & Neck Tumors,” Chinese Journal of Electronics, vol. 33, no. 3, pp. 766–775, 2024 doi: 10.23919/cje.2022.00.306

Citation:

Juanying XIE, Ying PENG, and Mingzhao WANG, “The Squeeze & Excitation Normalization Based nnU-Net for Segmenting Head & Neck Tumors,” Chinese Journal of Electronics, vol. 33, no. 3, pp. 766–775, 2024 doi: 10.23919/cje.2022.00.306

Citation:

PDF( 7014 KB)

The Squeeze & Excitation Normalization Based nnU-Net for Segmenting Head & Neck Tumors

doi: 10.23919/cje.2022.00.306

1.
School of Computer Science, Shaanxi Normal University, Xi’an 710119, China

More Information

Author Bio:
Juanying XIE is a Professor and a Ph.D. student Supervisor of the School of Computer Science of Shaanxi Normal University, Xi’an, China. She received the Ph.D. and M.S. degree from Xidian University in 2012 and 2004, respectively. She received the B.S. degree from Shaanxi Normal University in 1993. Her research interests include machine learning, data mining, and biomedical data analysis. Her research is highly cited, with one article in the top 1% of ESI and one is the top 3 hotspot article of SCIENTIA SINICA Informationis and 3 articles was included in F5000. She is an Associate Editor of Health Information Science and Systems, and an editor board member of the journal of Shaanxi Normal University (Natural Science Edition). (Email: xiejuany@snnu.edu.cn)

Ying PENG received the M.S. degree in the application technology of computer science and B.S. degree in computer science from Shaanxi Normal University, Xi’an, China, in 2022 and 2019, respectively. Her research interests include deep learning and biomedical image segmentation. (Email: py183248@snnu.edu.cn)

Mingzhao WANG is a Post Doctor supervised by Professor Juanying Xie at the School of Computer Science of Shaanxi Normal University, Xi’an, China. He received the Ph.D. degree in bioinformatics and M.S. degree in the application technology of computer science from Shaanxi Normal University in 2021 and 2017, respectively. He received the B.S. degree in computer science from Shanxi Normal University, Linfen, China, in 2014. His research interests include machine learning and bioinformatics. (Email: wangmz2017@snnu.edu.cn)
Corresponding author: Email: xiejuany@snnu.edu.cn
Received Date: 2022-09-09
Accepted Date: 2022-12-13

Available Online: 2023-02-18

Publish Date: 2024-05-05

Abstract

Abstract

Head and neck cancer is one of the most common malignancies in the world. We propose SE-nnU-Net by adapting SE (squeeze and excitation) normalization into nnU-Net, so as to segment head and neck tumors in PET/CT images by combining advantages of SE capturing features of interest regions and nnU-Net configuring itself for a specific task. The basic module referred to convolution-ReLU-SE is designed for SE-nnU-Net. In the encoder it is combined with residual structure while in the decoder without residual structure. The loss function combines Dice loss and Focal loss. The specific data preprocessing and augmentation techniques are developed, and specific network architecture is designed. Moreover, the deep supervised mechanism is introduced to calculate the loss function using the last four layers of the decoder of SE-nnU-Net. This SE-nnU-net is applied to HECKTOR 2020 and HECKTOR 2021 challenges, respectively, using different experimental design. The experimental results show that SE-nnU-Net for HECKTOR 2020 obtained 0.745, 0.821, and 0.725 in terms of Dice, Precision, and Recall, respectively, while the SE-nnU-Net for HECKTOR 2021 obtains 0.778 and 3.088 in terms of Dice and median HD95, respectively. This SE-nnU-Net for segmenting head and neck tumors can provide auxiliary opinions for doctors’ diagnoses.
- Head and neck tumors,
- PET/CT images,
- Image segmentation,
- SE-nnU-Net,
- Squeeze and excitation normalization

FullText(HTML)

References(40)

References

[1]	F. Bray, J. Ferlay, I. Soerjomataram, et al., “Global cancer statistics 2018: Globocan estimates of incidence and mortality worldwide for 36 cancers in 185 countries,” CA:A Cancer Journal for Clinicians, vol. 68, no. 6, pp. 394–424, 2018. doi: 10.3322/caac.21492
[2]	S. L. Breen, J. Publicover, S. De Silva, et al., “Intraobserver and interobserver variability in GTV delineation on FDG-PET-CT images of head and neck cancers,” International Journal of Radiation Oncology, Biology, Physics, vol. 68, no. 3, pp. 763–770, 2007. doi: 10.1016/j.ijrobp.2006.12.039
[3]	D. Kawahara, M. Tsuneda, S. Ozawa, et al., “Stepwise deep neural network (stepwise-net) for head and neck auto-segmentation on CT images,” Computers in Biology and Medicine, vol. 143, article no. 105295, 2022. doi: 10.1016/j.compbiomed.2022.105295
[4]	B. X. Wu, P. L. Khong, and T. Chan, “Automatic detection and classification of nasopharyngeal carcinoma on PET/CT with support vector machine,” International Journal of Computer Assisted Radiology and Surgery, vol. 7, no. 4, pp. 635–646, 2012. doi: 10.1007/s11548-011-0669-y
[5]	Q. Song, J. J. Bai, D. F. Han, et al., “Optimal co-segmentation of tumor in PET-CT images with context information,” IEEE Transactions on Medical Imaging, vol. 32, no. 9, pp. 1685–1697, 2013. doi: 10.1109/TMI.2013.2263388
[6]	Y. M. Moe, A. R. Groendahl, M. Mulstad, et al., Deep learning for automatic tumour segmentation in PET/CT images of patients with head and neck cancers, arXiv preprint arXiv: 1908.00841, 2019, doi: 10.48550/arXiv.1908.00841.
[7]	V. Andrearczyk, V. Oreiller, M. Vallières, et al., “Automatic segmentation of head and neck tumors and nodal metastases in PET-CT scans,” in Proceedings of the International Conference on Medical Imaging with Deep Learning, Montréal, Canada, pp.33–43, 2020.
[8]	Y. D. Yuan, “Automatic head and neck tumor segmentation in PET/CT with scale attention network,” in Proceedings of First 3D Head and Neck Tumor Segmentation in PET/CT Challenge, Lima, Peru, pp.44–52, 2021.
[9]	J. Ma and X. P. Yang, “Combining CNN and hybrid active contours for head and neck tumor segmentation in CT and pet images,” in Proceedings of First 3D Head and Neck Tumor Segmentation in PET/CT Challenge, Lima, Peru, pp.59–64, 2021.
[10]	H. Chen, H. B. Chen, and L. S. Wang, “Iteratively refine the segmentation of head and neck tumor in FDG-PET and CT images,” in Proceedings of First 3D Head and Neck Tumor Segmentation in PET/CT Challenge, Lima, Peru, pp.53–58, 2021.
[11]	F. Isensee, P. F. Jaeger, S. A. A. Kohl, et al., “nnU-Net: A self-configuring method for deep learning-based biomedical image segmentation,” Nature Methods, vol. 18, no. 2, pp. 203–211, 2021. doi: 10.1038/s41592-020-01008-z
[12]	T. Falk, D. Mai, R. Bensch, et al., “U-net: Deep learning for cell counting, detection, and morphometry,” Nature Methods, vol. 16, no. 1, pp. 67–70, 2019. doi: 10.1038/s41592-018-0261-2
[13]	H. El-Hariri, L. A. S. M. Neto, P. Cimflova, et al., “Evaluating nnU-Net for early ischemic change segmentation on non-contrast computed tomography in patients with acute ischemic stroke,” Computers in Biology and Medicine, vol. 141, article no. 105033, 2022. doi: 10.1016/j.compbiomed.2021.105033
[14]	G. B. Zhang, Z. Y. Yang, B. Huo, et al., “Multiorgan segmentation from partially labeled datasets with conditional nnU-Net,” Computers in Biology and Medicine, vol. 136, article no. 104658, 2021. doi: 10.1016/j.compbiomed.2021.104658
[15]	J. Hu, L. Shen, and G. Sun, “Squeeze-and-excitation networks,” in Proceedings of the 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition, Salt Lake City, UT, USA, pp.7132–7141, 2018.
[16]	L. Rundo, C. Han, Y. Nagano, et al., “Use-net: Incorporating squeeze-and-excitation blocks into U-Net for prostate zonal segmentation of multi-institutional MRI datasets,” Neurocomputing, vol. 365, pp. 31–43, 2019. doi: 10.1016/j.neucom.2019.07.006
[17]	B. H. Zhang, S. L. Qi, Y. N. Wu, et al., “Multi-scale segmentation squeeze-and-excitation UNet with conditional random field for segmenting lung tumor from CT images,” Computer Methods and Programs in Biomedicine, vol. 222, article no. 106946, 2022. doi: 10.1016/j.cmpb.2022.106946
[18]	A. Iantsen, V. Jaouen, D. Visvikis, et al., “Squeeze-and-excitation normalization for brain tumor segmentation,” in Proceedings of 6th International MICCAI Brainlesion Workshop, Lima, Peru, pp.366–373, 2021.
[19]	V. Andrearczyk, V. Oreiller, M. Jreige, et al., “Overview of the HECKTOR challenge at MICCAI 2020: Automatic head and neck tumor segmentation in PET/CT,” in Proceedings of First 3D Head and Neck Tumor Segmentation in PET/CT Challenge, Lima, Peru, pp.1–21, 2021.
[20]	A. Iantsen, D. Visvikis, and M. Hatt, “Squeeze-and-excitation normalization for automated delineation of head and neck primary tumors in combined PET and CT images,” in Proceedings of First 3D Head and Neck Tumor Segmentation in PET/CT Challenge, Lima, Peru, pp.37–43, 2021.
[21]	L. R. Dice, “Measures of the amount of ecologic association between species,” Ecology, vol. 26, no. 3, pp. 297–302, 1945. doi: 10.2307/1932409
[22]	D. P. Huttenlocher, G. A. Klanderman, and W. J. Rucklidge, “Comparing images using the Hausdorff distance,” IEEE Transactions on Pattern Analysis and Machine Intelligence, vol. 15, no. 9, pp. 850–863, 1993. doi: 10.1109/34.232073
[23]	L. C. Chen, G. Papandreou, I. Kokkinos, et al., “DeepLab: Semantic image segmentation with deep convolutional nets, atrous convolution, and fully connected CRFs,” IEEE Transactions on Pattern Analysis and Machine Intelligence, vol. 40, no. 4, pp. 834–848, 2018. doi: 10.1109/TPAMI.2017.2699184
[24]	J. Y. Xie and Y. Peng, “The head and neck tumor segmentation using nnu-net with spatial and channel ‘squeeze & excitation’ blocks,” in Proceedings of First 3D Head and Neck Tumor Segmentation in PET/CT Challenge, Lima, Peru, pp.28–36, 2021.
[25]	K. Ghimire, Q. Chen, and X. Feng, “Patch-based 3D UNet for head and neck tumor segmentation with an ensemble of conventional and dilated convolutions,” in Proceedings of First 3D Head and Neck Tumor Segmentation in PET/CT Challenge, Lima, Peru, pp.78–84, 2021.
[26]	F. Yousefirizi and A. Rahmim, “GAN-based bi-modal segmentation using mumford-shah loss: Application to head and neck tumors in PET-CT images,” in Proceedings of First 3D Head and Neck Tumor Segmentation in PET/CT Challenge, Lima, Peru, pp.99–108, 2021.
[27]	S. M. Zhu, Z. Z. Dai, and N. Wen, “Two-stage approach for segmenting gross tumor volume in head and neck cancer with CT and PET imaging,” in Proceedings of First 3D Head and Neck Tumor Segmentation in PET/CT Challenge, Lima, Peru, pp.22–27, 2021.
[28]	M. A. Naser, L. V. Van Dijk, R. J. He, et al., “Tumor segmentation in patients with head and neck cancers using deep learning based-on multi-modality PET/CT images,” in Proceedings of First 3D Head and Neck Tumor Segmentation in PET/CT Challenge, Lima, Peru, pp.85–98, 2021.
[29]	C. Rao, S. Pai, I. Hadzic, et al., “Oropharyngeal tumour segmentation using ensemble 3D PET-CT fusion networks for the HECKTOR challenge,” in Proceedings of First 3D Head and Neck Tumor Segmentation in PET/CT Challenge, Lima, Peru, pp.65–77, 2021.
[30]	V. Andrearczyk, V. Oreiller, S. Boughdad, et al., “Overview of the HECKTOR challenge at MICCAI 2021: Automatic head and neck tumor segmentation and outcome prediction in PET/CT images,” in Proceedings of Second 3D Head and Neck Tumor Segmentation in PET/CT Challenge, Strasbourg, France, pp.1–37, 2022.
[31]	C. Y. An, H. Chen, and L. S. Wang, “A coarse-to-fine framework for head and neck tumor segmentation in CT and PET images,” in Proceedings of Second 3D Head and Neck Tumor Segmentation in PET/CT Challenge, Strasbourg, France, pp.50–57, 2022.
[32]	J. S. Lu, W. H. Lei, R. Gu, et al., “Priori and posteriori attention for generalizing head and neck tumors segmentation,” in Proceedings of Second 3D Head and Neck Tumor Segmentation in PET/CT Challenge, Strasbourg, France, pp.134–140, 2022.
[33]	F. Yousefirizi, I. Janzen, N. Dubljevic, et al., “Segmentation and risk score prediction of head and neck cancers in PET/CT volumes with 3D U-Net and cox proportional hazard neural networks,” in Proceedings of Second 3D Head and Neck Tumor Segmentation in PET/CT Challenge, Strasbourg, France, pp.236–247, 2022.
[34]	J. T. Ren, B. N. Huynh, A. R. Groendahl, et al., “Pet normalizations to improve deep learning auto-segmentation of head and neck tumors in 3D PET/CT,” in Proceedings of Second 3D Head and Neck Tumor Segmentation in PET/CT Challenge, Strasbourg, France, pp.83–91, 2022.
[35]	M. A. Naser, K. A. Wahid, L. V. Van Dijk, et al., “Head and neck cancer primary tumor auto segmentation using model ensembling of deep learning in PET/CT images,” in Proceedings of Second 3D Head and Neck Tumor Segmentation in PET/CT Challenge, Strasbourg, France, pp.121–133, 2022.
[36]	A. De Biase, W. Tang, N. Sourlos, et al., “Skip-SCSE multi-scale attention and co-learning method for oropharyngeal tumor segmentation on multi-modal PET-CT images,” in Proceedings of Second 3D Head and Neck Tumor Segmentation in PET/CT Challenge, Strasbourg, France, pp.109–120, 2022.
[37]	G. S. Wang, Z. Y. Huang, H. Shen, et al., “The head and neck tumor segmentation in PET/CT based on multi-channel attention network,” in Proceedings of Second 3D Head and Neck Tumor Segmentation in PET/CT Challenge, Strasbourg, France, pp.68–74, 2022.
[38]	M. Cho, Y. Choi, D. Hwang, et al., “Multimodal spatial attention network for automatic head and neck tumor segmentation in FDG-PET and CT images,” in Proceedings of Second 3D Head and Neck Tumor Segmentation in PET/CT Challenge, Strasbourg, France, pp.75–82, 2022.
[39]	M. Y. Meng, Y. G. Peng, L. Bi, et al., “Multi-task deep learning for joint tumor segmentation and outcome prediction in head and neck cancer,” in Proceedings of Second 3D Head and Neck Tumor Segmentation in PET/CT Challenge, Strasbourg, France, pp.160–167, 2022.
[40]	J. Y. Xie and Y. Peng, “The head and neck tumor segmentation based on 3D U-Net,” in Proceedings of Second 3D Head and Neck Tumor Segmentation in PET/CT Challenge, Strasbourg, France, pp.92–98, 2022.