Enhanced Cross-stage-attention U-Net for Esophageal Target Volume Segmentation

Overview

Journal BMC Med Imaging

Publisher Biomed Central

Specialty Radiology

Date 2024 Dec 19

PMID 39696039

Authors

Xiao Lou

Juan Zhu

Jian Yang

Youzhe Zhu

Huazhong Shu

Baosheng Li

Affiliations

Soon will be listed here.

Abstract

Purpose: The segmentation of target volume and organs at risk (OAR) was a significant part of radiotherapy. Specifically, determining the location and scale of the esophagus in simulated computed tomography images was difficult and time-consuming primarily due to its complex structure and low contrast with the surrounding tissues. In this study, an Enhanced Cross-stage-attention U-Net was proposed to solve the segmentation problem for the esophageal gross tumor volume (GTV) and clinical tumor volume (CTV) in CT images.

Methods: First, a module based on principal component analysis theory was constructed to pre-extract the features of the input image. Then, a cross-stage based feature fusion model was designed to replace the skip concatenation of original UNet, which was composed of Wide Range Attention unit, Small-kernel Local Attention unit, and Inverted Bottleneck unit. WRA was employed to capture global attention, whose large convolution kernel was further decomposed to simplify the calculation. SLA was used to complement the local attention to WRA. IBN was structed to fuse the extracted features, where a global frequency response layer was built to redistribute the frequency response of the fused feature maps.

Results: The proposed method was compared with relevant published esophageal segmentation methods. The prediction of the proposed network was MSD = 2.83(1.62, 4.76)mm, HD = 11.79 ± 6.02 mm, DC = 72.45 ± 19.18% in GTV; MSD = 5.26(2.18, 8.82)mm, HD = 16.22 ± 10.01 mm, DC = 71.06 ± 17.72% in CTV.

Conclusion: The reconstruction of the skip concatenation in UNet showed an improvement of performance for esophageal segmentation. The results showed the proposed network had better effect on esophageal GTV and CTV segmentation.

References

You C, Zhao R, Staib L, Duncan J . Momentum Contrastive Voxel-wise Representation Learning for Semi-supervised Volumetric Medical Image Segmentation. Med Image Comput Comput Assist Interv. 2023; 13434:639-652. PMC: 10352821. DOI: 10.1007/978-3-031-16440-8_61. View

Jian M, Tao C, Wu R, Zhang H, Li X, Wang R . HRU-Net: A high-resolution convolutional neural network for esophageal cancer radiotherapy target segmentation. Comput Methods Programs Biomed. 2024; 250:108177. DOI: 10.1016/j.cmpb.2024.108177. View

Yang J, Haas B, Fang R, Beadle B, Garden A, Liao Z . Atlas ranking and selection for automatic segmentation of the esophagus from CT scans. Phys Med Biol. 2017; 62(23):9140-9158. PMC: 6167015. DOI: 10.1088/1361-6560/aa94ba. View

Diniz J, Ferreira J, Diniz P, Silva A, Cardoso de Paiva A . Esophagus segmentation from planning CT images using an atlas-based deep learning approach. Comput Methods Programs Biomed. 2020; 197:105685. DOI: 10.1016/j.cmpb.2020.105685. View

You C, Dai W, Min Y, Staib L, Duncan J . Implicit Anatomical Rendering for Medical Image Segmentation with Stochastic Experts. Med Image Comput Comput Assist Interv. 2024; 14222:561-571. PMC: 11151725. DOI: 10.1007/978-3-031-43898-1_54. View

You C, Dai W, Liu F, Min Y, Dvornek N, Li X . Mine Your Own Anatomy: Revisiting Medical Image Segmentation With Extremely Limited Labels. IEEE Trans Pattern Anal Mach Intell. 2024; PP. PMC: 11903367. DOI: 10.1109/TPAMI.2024.3461321. View

Tran M, Kim S, Yang H, Lee G, Oh I, Kang S . Esophagus Segmentation in CT Images via Spatial Attention Network and STAPLE Algorithm. Sensors (Basel). 2021; 21(13). PMC: 8271959. DOI: 10.3390/s21134556. View

Liu X, Liu Q, Zhang Y, Wang M, Tang J . TSSK-Net: Weakly supervised biomarker localization and segmentation with image-level annotation in retinal OCT images. Comput Biol Med. 2022; 153:106467. DOI: 10.1016/j.compbiomed.2022.106467. View

Verellen D, De Ridder M, Linthout N, Tournel K, Soete G, Storme G . Innovations in image-guided radiotherapy. Nat Rev Cancer. 2007; 7(12):949-60. DOI: 10.1038/nrc2288. View

10.

Dietz B, Yip E, Yun J, Fallone B, Wachowicz K . Real-time dynamic MR image reconstruction using compressed sensing and principal component analysis (CS-PCA): Demonstration in lung tumor tracking. Med Phys. 2017; 44(8):3978-3989. DOI: 10.1002/mp.12354. View

11.

You C, Dai W, Min Y, Liu F, Clifton D, Kevin Zhou S . Rethinking Semi-Supervised Medical Image Segmentation: A Variance-Reduction Perspective. Adv Neural Inf Process Syst. 2024; 36():9984-10021. PMC: 11136570. View

12.

Trullo R, Petitjean C, Nie D, Shen D, Ruan S . Fully automated esophagus segmentation with a hierarchical deep learning approach. Conf Proc IEEE Int Conf Signal Image Process Appl. 2018; 2017:503-506. PMC: 6193464. DOI: 10.1109/ICSIPA.2017.8120664. View

13.

Dawson L, Sharpe M . Image-guided radiotherapy: rationale, benefits, and limitations. Lancet Oncol. 2006; 7(10):848-58. DOI: 10.1016/S1470-2045(06)70904-4. View

14.

Alam S, Zhang P, Zhang S, Chen I, Rimner A, Tyagi N . Early Prediction of Acute Esophagitis for Adaptive Radiation Therapy. Int J Radiat Oncol Biol Phys. 2021; 110(3):883-892. PMC: 8180486. DOI: 10.1016/j.ijrobp.2021.01.007. View

15.

CAMPBELL F, Robson J . Application of Fourier analysis to the visibility of gratings. J Physiol. 1968; 197(3):551-66. PMC: 1351748. DOI: 10.1113/jphysiol.1968.sp008574. View

16.

You C, Xiang J, Su K, Zhang X, Dong S, Onofrey J . Incremental Learning Meets Transfer Learning: Application to Multi-site Prostate MRI Segmentation. Distrib Collab Fed Learn Afford AI Healthc Resour Div Glob Health (2022). 2023; 13573:3-16. PMC: 10323962. DOI: 10.1007/978-3-031-18523-6_1. View

17.

Lagendijk J, Raaymakers B, van den Berg C, Moerland M, Philippens M, van Vulpen M . MR guidance in radiotherapy. Phys Med Biol. 2014; 59(21):R349-69. DOI: 10.1088/0031-9155/59/21/R349. View

18.

Huang T, Zhang G, Guerrero T, Starkschall G, Lin K, Forster K . Semi-automated CT segmentation using optic flow and Fourier interpolation techniques. Comput Methods Programs Biomed. 2006; 84(2-3):124-34. DOI: 10.1016/j.cmpb.2006.09.003. View

19.

Kurugol S, Bas E, Erdogmus D, Dy J, Sharp G, Brooks D . Centerline extraction with principal curve tracing to improve 3D level set esophagus segmentation in CT images. Annu Int Conf IEEE Eng Med Biol Soc. 2012; 2011:3403-6. PMC: 3349355. DOI: 10.1109/IEMBS.2011.6090921. View

20.

Jin D, Guo D, Ho T, Harrison A, Xiao J, Tseng C . DeepTarget: Gross tumor and clinical target volume segmentation in esophageal cancer radiotherapy. Med Image Anal. 2020; 68:101909. DOI: 10.1016/j.media.2020.101909. View