基于Vgg16-Unet模型的MRI图像下前列腺分区智能分割研究

呼延若曦; 吴哲; 许杉杉; 崔慧林; 曹锡梅; 马晋锋; 吴毅; 顾玮

doi:10.16016/j.2097-0927.202211045

您当前的位置：

首页 >

文章列表页 >

基于Vgg16-Unet模型的MRI图像下前列腺分区智能分割研究

生物医学工程 | 更新时间：2026-01-09

- 基于Vgg16-Unet模型的MRI图像下前列腺分区智能分割研究
- Intelligent segmentation of prostate zones in MR images based on Vgg16-Unet
- 陆军军医大学学报 2023年45卷第13期页码：1441-1449
- 作者机构：
  
  1.400044 重庆，重庆大学：医学院
  2.400044 重庆，重庆大学：生物工程学院
  3.400038 重庆，陆军军医大学(第三军医大学)生物医学工程与影像医学系数字医学教研室
  4.030001 太原，山西医科大学组织胚胎学教研室
  5.030013 太原，山西省肿瘤医院普外科
- 作者简介：
  
  [吴毅，E-mail：wuy1979@tmmu.edu.cn WU Yi, E-mail: wuy1979@tmmu.edu.cn
  [顾玮，E-mail：guwei@cqu.edu.cn GU Wei, E-mail: guwei@cqu.edu.cn
- 基金信息：
  
  国家自然科学基金(31971113);重庆英才青年拔尖人才项目(CQYC201905037)
- DOI：10.16016/j.2097-0927.202211045
  中图分类号： R312;R445.2;R737.25
- 收稿：2022-11-08，
  
  修回：2022-12-18，
  
  纸质出版：2023-07-15
- 稿件说明：
移动端阅览
呼延若曦, 吴哲, 许杉杉, 等. 基于Vgg16-Unet模型的MRI图像下前列腺分区智能分割研究[J]. 陆军军医大学学报, 2023,45(13):1441-1449.

Ruoxi HUYAN, Zhe WU, Shanshan XU, et al. Intelligent segmentation of prostate zones in MR images based on Vgg16-Unet[J]. Journal of Army Medical University, 2023, 45(13): 1441-1449.
呼延若曦, 吴哲, 许杉杉, 等. 基于Vgg16-Unet模型的MRI图像下前列腺分区智能分割研究[J]. 陆军军医大学学报, 2023,45(13):1441-1449. DOI： 10.16016/j.2097-0927.202211045.

Ruoxi HUYAN, Zhe WU, Shanshan XU, et al. Intelligent segmentation of prostate zones in MR images based on Vgg16-Unet[J]. Journal of Army Medical University, 2023, 45(13): 1441-1449. DOI： 10.16016/j.2097-0927.202211045.

摘要

目的

基于前列腺核磁共振图像(MRI)，构建深度学习智能分割模型，展开MRI图像下前列腺分区的智能分割研究。

方法

收集山西省肿瘤医院2018年1月至2020年10月33例经MRI扫描的前列腺癌患者T2WI本地数据，包括T1期6例，T2期15例，T3期9例，T4期3例。选取荷兰奈梅亨拉德堡德大学医学中心提供的前列腺MRI公开数据集中前列腺癌患者的T2WI序列数据，共379例数据，所有数据按照7 ∶1 ∶2的比例随机划分为训练集265例、验证集38例和测试集76例。在Unet模型基础上以Vgg16模型为编码器，使用多层卷积层的同时利用迁移学习策略，构建Vgg16-Unet模型，以医师手工勾画和标注的前列腺分区(前纤维基质带、中央带、外周带、移形带)为金标准，在测试集上，采用Dice相似系数(Dice similarity coefficient，DSC)、95%豪斯多夫距离(95% Hausdorff surface Distance，HD95)评估模型对前列腺分区的分割精度。

结果

模型在测试集上对前纤维基质带、中央带、外周带、移形带实现了较为准确的分割，其平均DSC分别为56.95%、47.28%、80.78%、90.63%，平均HD95分别为20.84、20.02、15.39、11.20 mm。模型智能分割与手工标注测量的体积一致性较好，其差值均位于95%一致性区间内。

结论

构建的Vgg16-Unet模型分割精度优于Unet、Unet++、ResUnet++ 3个Unet经典变种网络，能够显著提高前列腺癌MRI图像分割效率，减轻医师工作量。

Abstract

Objective

To create an intelligent segmentation model of deep learning based on the magnetic resonance (MR) images of the prostate zones.

Methods

The T2WI sequence data of 33 patients with prostate cancer

including 6 at T1 stage

15 at T2 stage

9 at T3 stage

and 3 at T4 stage

receiving MRI scanning in Shanxi Provincial Cancer Hospital from January 2018 to October 2020 were collected. While

the T2WI sequence data of 346 patients with prostate cancer from the Prostate MRI Public Dataset provided by Radboud University Nijmegen Medical Centre

Netherlands were also collected. Then these totally 379 cases were randomly divided into training sets

validation sets and test sets at a ratio of 7 ∶1 ∶2 (265

38 and 76 cases

respectively). Based on Unet model

Vgg16 was used as the encoder

and multiple layers of convolution were used with the transfer learning strategy at the same time to build a Vgg16-Unet model. Then with a gold standard of the prostate zones (anterior fibrous matrix zone

central zone

peripheral zone

and transition zone) manually segmented by doctors

Dice similarity coefficient (DSC) and 95% Hausdorff surface distance (HD95) were employed to evaluate the segmentation accuracy of the prostate zones on the test set.

Results

The model achieved relatively higher accuracy of segmentation for anterior fibrous matrix zone

central zone

peripheral zone and transition zone on the test set. The average DSC was 56.95%

47.28%

80.78%

and 90.63%

and the average HD95 value was 20.84

20.02

15.39 and 11.20 mm

respectively. The volume predicted by the model was consistent with the volume measured by manual segmentation

and their differences were basically within the 95% consistency interval.

Conclusion

The segmentation accuracy of our Vgg16-Unet model is better than that of 3 classical variants

Unet

Unet++ and ResUnet++. The model can significantly improve the segmentation efficiency of prostate cancer MRI images and reduce the workload of clinicians.

关键词

Keywords

references

XIA C F, DONG X S, LI H, et al . Cancer statistics in China and United States, 2022: profiles, trends, and determinants[J ] . Chin Med J (Engl), 2022, 135(5): 584-590. DOI: 10.1097/CM9.0000000000002108.

SIEGELR L, MILLER K D, JEMAL A. Cancer statistics, 2015[J]. CA A Cancer J Clin, 2015, 65(1): 5-29. DOI: 10.3322/caac.21254.

胡学刚, 杨洪光. 基于全卷积DenseNet的前列腺MRI分割新方法[J]. 计算机应用研究, 2020, 37(6): 1886-1889, 1894. DOI: 10.19734/j.issn.1001-3695.2018.11.0909.

HU X G, YANG H G. Novelprostate MRI segmentation method based on full convolutionDenseNet[J]. Appl Res Comput, 2020, 37(6): 1886-1889, 1894. DOI: 10.19734/j.issn.1001-3695.2018.11.0909.

KURATA Y, NISHIO M, KIDO A, et al . Automatic segmentation of the uterus on MRI using a convolutional neural network[J ] . Comput Biol Med, 2019, 114: 103438. DOI: 10.1016/j.compbiomed.2019.103438.

LEE B, YAMANAKKANAVAR N, CHOIJ Y. Automatic segmentation of brain MRI using a novel patch-wise U-net deep architecture[J]. PLoS One, 2020, 15(8): e0236493. DOI: 10.1371/journal.pone.0236493.

AFZALI A, BABAPOUR MOFRAD F, POULADIAN M. 2D statistical lung shape analysis using chest radiographs: modelling and segmentation[J]. J Digit Imaging, 2021, 34(3): 523-540. DOI: 10.1007/s10278-021-00440-7.

ZABIHOLLAHY F, SCHIEDA N, KRISHNA JEYARAJ S, et al . Automated segmentation of prostate zonal anatomy on T2-weighted (T2W) and apparent diffusion coefficient (ADC) map MR images using U-Nets[J ] . Med Phys, 2019, 46(7): 3078-3090. DOI: 10.1002/mp.13550.

LITJENSG, DEBATS O, BARENTSZ J, et al . Computer-aided detection of prostate cancer in MRI[J ] . IEEE Trans Med Imaging, 2014, 33(5): 1083-1092. DOI: 10.1109/tmi.2014.2303821.

GRIMES D A, SCHULZ K F. An overview of clinical research: the lay of the land[J]. Lancet, 2002, 359(9300): 57-61. DOI: 10.1016/S0140-6736(02)07283-5.

刘燕茹, 罗利霞, 刘明芳. Windows系统上DICOM图像开窗显示方法实现[J]. 通讯世界, 2020, 27(1): 150. DOI: 10.3969/j.issn.1006-4222.2020.01.094.

LIU Y R, LUO L X, LIU M F. Realization of window display method of DICOM image on Windows system[J]. Telecom World, 2020, 27(1): 150. DOI: 10.3969/j.issn.1006-4222.2020.01.094.

ALSHAMMARI A. Construction of VGG16 convolution neural Network (VGG16_CNN) classifier with NestNet-based segmentation paradigm for brain metastasis classification[J]. Sensors (Basel), 2022, 22(20): 8076. DOI: 10.3390/s22208076.

DENG J, DONG W, SOCHER R, et al . ImageNet: a large-scale hierarchical image database[C ] //2009 IEEE Conference on computer vision and pattern recognition, 2009: 248-255. DOI: 10.1109/CVPR.2009.5206848 http://dx.doi.org/10.1109/CVPR.2009.5206848 ..

HEK M, GIRSHICK R, DOLLÁR P. Rethinking ImageNet pre-training[EB/OL ] . [2018-11-21 ] . https://arxiv.org/abs/1811.08883 https://arxiv.org/abs/1811.08883 .

YU J, CAI L, CHEN C, et al. Cascade Path Augmentation Unet for bladder cancer segmentation in MRI[J ] . Med Phys, 2022, 49(7): 4622-4631. DOI: 10.1002/mp.15646.

KIRKLAND E J. Bilinear interpolation[M ] //Advanced computing in electron microscopy. Boston: Springer US, 2010: 261-263. DOI: 10.1007/978-1-4419-6533-2_12 http://dx.doi.org/10.1007/978-1-4419-6533-2_12 .

VUM H, GRIMBERGEN G, NYHOLM T, et al . Evaluation of multislice inputs to convolutional neural networks for medical image segmentation[J ] . Med Phys, 2020, 47(12): 6216-6231. DOI: 10.1002/mp.14391.

吴毅, 张小勤. 人工智能在医学图像处理中的研究进展与展望[J]. 第三军医大学学报, 2021, 43(18): 1707-1712. DOI: 10.16016/j.1000-5404.202106194.

WU Y, ZHANG X Q. Artificial intelligence in medical image processing: progress andprospect[J]. J Third Mil Med Univ, 2021, 43(18): 1707-1712. DOI: 10.16016/j.1000-5404.202106194.

REN X H, XIANG L, NIE D, et al . Interleaved 3D-CNNs for joint segmentation of small-volume structures in head and neck CT images[J ] . Med Phys, 2018, 45(5): 2063-2075. DOI: 10.1002/mp.12837.

吴哲, 庞亚, 明智, 等. 人工智能技术在鼻咽癌放疗危及器官自动勾画中的应用研究[J]. 实用肿瘤学杂志, 2021, 35(2): 137-141. DOI: 10.11904/j.issn.1002-3070.2021.02.007.

WU Z, PANG Y, MING Z, et al . Application of artificial intelligence technology in automatic delineation of organs at risk in nasopharyngeal carcinoma radiotherapy[J ] . Pract Oncol J, 2021, 35(2): 137-141. DOI: 10.11904/j.issn.1002-3070.2021.02.007.

刘想, 高歌, 韩超, 等. 基于3D U-Net模型实现T2WI图像中前列腺的区域分割及临床验证[J]. 中国临床医学影像杂志, 2022, 33(1): 33-38, 42. DOI: 10.12117/jccmi.2022.01.008.

LIU X, GAO G, HAN C, et al . Regional segmentation of prostate and clinical evaluation in T2WI image based on 3D U-Net model[J ] . J China Clin Med Imaging, 2022, 33(1): 33-38, 42. DOI: 10.12117/jccmi.2022.01.008.

KHAN Z, YAHYA N, ALSAIH K, et al . Evaluation of deep neural networks for semantic segmentation of prostate in T2W MRI[J ] . Sensors (Basel), 2020, 20(11): 3183. DOI: 10.3390/s20113183.

ZAVALA-ROMERO O, BRETOA L, XU I R, et al . Segmentation of prostate and prostate zones using deep learning: a multi-MRI vendor analysis[J ] . Al, 2020, 196(10): 932-942. DOI: 10.1007/s00066-020-01607-x.

BARDIS M, HOUSHYAR R, CHANTADULY C, et al . Segmentation of the prostate transition zone and peripheral zone on MR images with deep learning[J ] . Radiol Imaging Cancer, 2021, 3(3): e200024. DOI: 10.1148/rycan.2021200024.

CUOCOLO R, COMELLI A, STEFANO A, et al . Deep learning whole-gland and zonal prostate segmentation on a public MRI dataset[J ] . J Magn Reson Imaging, 2021, 54(2): 452-459. DOI: 10.1002/jmri.27585.

LI Y C, LIN C, ZHANG Y, et al . Automatic segmentation of prostate MRI based on 3D pyramid pooling Unet[J ] . Med Phys, 2023, 50(2): 906-921. DOI: 10.1002/mp.15895.

浏览量

下载量

CSCD

文章被引用时，请邮件提醒。

提交

工具集

关联资源

融合深度学习、临床与影像组学特征的综合模型提升非小细胞肺癌EGFR突变预测效能

基于CT影像的3D ResNet18模型在肺结核耐药性识别中的预测性能研究

深度学习和磁共振图像辅助脑胶质瘤基因分型的研究进展

基于小波变换和CNN-Transformer的超声甲状腺结节分割算法研究

基于深度学习的MRI影像女性尿控解剖元件及相关结构的自动分割