切换至 "中华医学电子期刊资源库"

中华消化病与影像杂志(电子版) ›› 2022, Vol. 12 ›› Issue (06) : 342 -347. doi: 10.3877/cma.j.issn.2095-2015.2022.06.004

论著

基于高分辨T2WI深度学习诊断直肠癌直肠系膜内淋巴结状态的研究
贾玉萍1, 宋歌声1,(), 李爱银1   
  1. 1. 250014 济南,山东第一医科大学第一附属医院(山东省千佛山医院)放射科
  • 收稿日期:2022-04-20 出版日期:2022-12-01
  • 通信作者: 宋歌声
  • 基金资助:
    山东省科技发展计划(2014GSF118086); 济南市科技计划(201907034)

Using deep learning to diagnose the status of mesenteric lymph nodes in rectal cancer basing on high-resolution T2WI

Yuping Jia1, Gesheng Song1,(), Aiyin Li1   

  1. 1. Department of Radiology, First Affiliated Hospital of Shandong First Medical University & Shandong Provincial Qianfoshan Hospital, Jinan 250014, China
  • Received:2022-04-20 Published:2022-12-01
  • Corresponding author: Gesheng Song
引用本文:

贾玉萍, 宋歌声, 李爱银. 基于高分辨T2WI深度学习诊断直肠癌直肠系膜内淋巴结状态的研究[J/OL]. 中华消化病与影像杂志(电子版), 2022, 12(06): 342-347.

Yuping Jia, Gesheng Song, Aiyin Li. Using deep learning to diagnose the status of mesenteric lymph nodes in rectal cancer basing on high-resolution T2WI[J/OL]. Chinese Journal of Digestion and Medical Imageology(Electronic Edition), 2022, 12(06): 342-347.

目的

利用基于高清T2WI的深度学习模型来鉴别直肠癌直肠系膜内转移淋巴结和非转移淋巴结。

方法

收集2016年6月至2021年4月山东省千佛山医院就诊的直肠占位性病变患者166例,治疗前均接受3.0T磁共振扫描,包括高分辨T2WI(HR-T2WI)序列,并在2周内接受全直肠系膜切除术(TME)。术后由影像科和病理科医生共同对直肠系膜内的淋巴结进行定位和定性。将入组患者的淋巴结按照7∶3的比例随机分为训练组和测试组。提取入组患者的HR-T2WI图像,并对直肠系膜内淋巴结进行感兴趣区(ROI)标记。应用随机翻转和添加随机噪声两种数据增强方式来增加鲁棒性。建立包含5个卷积模块和2个全连接层的卷积神经网络(CNN)模型,并分别进行训练和测试。评价指标包括ROC曲线、AUC、准确性、敏感性和特异性。

结果

最终获得604个淋巴结(298个良性和306个恶性)。分别将训练集淋巴结(215恶性+205良性)和测试集淋巴结(91恶性+93良性)进行逆行深度学习CNN模型训练后,得到训练集和测试集的AUC分别为0.910、0.820。测试集准确率、敏感性和特异性分别为0.725、0.698和0.752。

结论

基于HR-T2WI序列深度学习方法可以用来鉴别直肠癌直肠系膜内淋巴结状态。

Objective

To distinguish metastatic and non-metastatic mesenteric lymph nodes(LNs)in rectal cancer using deep learning based on high-resolution T2WI.

Methods

A total of 166 patients with rectal space-occupying lesions were enrolled in Shandong Provincial Qianfoshan Hospital from June 2016 to April 2021.All patients underwent 3.0T magnetic resonance scanning, including high-resolution T2WI(HR-T2WI)sequence, and underwent total mesorectal excision(TME)within two weeks.The LNs within the mesentery were located and characterized by radiologist and pathologist.The LNs were randomly divided into training set and test set according to the ratio of 7∶3.HR-T2WI images were extracted from the enrolled patients, and the mesorectal LNs were labeled to get region of interest(ROI). Random flipping and adding random noise were used to increase the robustness.A convolutional neural network(CNN)model consisting of five convolution modules and two fully connected layers was established, and was trained and tested respectively.The evaluation indexes included ROC curve, AUC, accuracy, sensitivity and specificity.

Results

A total of 604 LNs(298 benign and 306 malignant)were finally obtained.After retrograde deep learning CNN model training of the LNs in the training set(215 malignant+ 205 benign)and the LNs in the test set(91 malignant+ 93 benign), the AUCs of the training set and the test set were 0.910 and 0.820 respectively.The accuracy, sensitivity and specificity of the test set were 0.725, 0.698 and 0.752 respectively.

Conclusion

The deep learning method based on HR-T2WI sequence can be used to identify the status of mesorectal LNs in rectal cancer.

表1 矢状位、轴位HR-T2WI、冠状位T2WI、轴位T1WI序列扫描参数
图1 深度学习ROI设置示意图注:1A中蓝色区域为标注的ROI区域,实际上采用的ROI在蓝色区域基础上进行外扩。1B为实际标注示意图。黑色边框对应图A中蓝色区域,白色区域为蓝色区域加外扩部分
图2 数据处理与数据增强过程注:其中Resize是必须操作,flip(翻转)和add noise(添加噪声)是随机组合进行
图3 模型结构设计示意图注:本研究中模型结构包括5个卷积模块(conv block)和2个全连接层
图4 模型结构过程图
表2 166例患者的临床指标
图5 深度学习训练集和测试集ROC曲线AUC分别为0.91和0.81
1
Siegel RLMiller KDJemal A.Cancer statistics,2020[J].CA Cancer J Clin202070(1):7-30.
2
Siegel RLMiller KDGoding Sauer A,et al.Colorectal cancer statistics,2020[J].CA Cancer J Clin202070(3):145-164.
3
Zhao LLiang MYang Y,et al.Value of multiple models of diffusion-weighted imaging for improving the nodal staging of preoperatively node-negative rectal cancer[J].J Abdom Radiol(NY)202146(10):4548-4555.
4
Xu HZhao WGuo W,et al.Prediction Model Combining Clinical and MR Data for Diagnosis of Lymph Node Metastasis in Patients with Rectal Cancer[J].J Magn Reson Imaging202153(3):874-883.
5
Kim HJChoi GS.Clinical Implications of Lymph Node Metastasis in Colorectal Cancer:Current Status and Future Perspectives[J].J Ann Coloproctol201935(3):109-117.
6
Deng Y.Rectal Cancer in Asian vs.Western Countries:Why the Variation in Incidence? [J].J Curr Treat Options Oncol201718(10):64.
7
Chan Heang-PingSamala Ravi KHadjiiski Lubomir M,et al.Deep Learning in Medical Image Analysis[J].J Adv Exp Med Biol202012(13):3-21.
8
Lee Jeong HoonHa Eun JuKim Ju Han.Application of deep learning to the diagnosis of cervical lymph node metastasis from thyroid cancer with CT[J].J Eur Radiol201929(10):5452-5457.
9
Ariji YoshikoSugita YoshihikoNagao Toru,et al.CT evaluation of extra nodal extension of cervical lymph node metastases in patients with oral squamous cell carcinoma using deep learning classification[J].Oral Radiol202036(2):148-155.
10
Sun QiuchangLin XiaonaZhao Yuanshen,et al.Deep Learning vs.Radiomics for Predicting Axillary Lymph Node Metastasis of Breast Cancer Using Ultrasound Images:Don′t Forget the Peritumoral Region[J].Front Oncol202010(53):1-12.
11
Zhao XingyuXie PeiyiWang Mengmeng,et al.Deep learning-based fully automated detection and segmentation of lymph nodes on multiparametric-mri for rectal cancer:A multicentre study[J].EBioMedicine202056:102780.
12
Abe TYasui MImamura H,et al.Combination of extramural venous invasion and lateral lymph node size detected with magnetic resonance imaging is a reliable biomarker for lateral lymph node metastasis in patients with rectal cancer[J].World J Surg Oncol202220(1):5.
13
Inra JASyngal S.Colorectal cancer in young adults[J].Dig Dis Sci201560(3):722-33.
14
Benson ABVenook APAl-Hawary MM,et al.NCCN Guidelines Insights:Rectal Cancer,Version 6.2020[J].J Natl Compr Canc Netw202018(7):806-815.
15
Fernandes MCGollub MJBrown G.The importance of MRI for rectal cancer evaluation[J].Surg Oncol202243:101739.
16
Long LZhang HHe X,et al.Value of intravoxel incoherent motion magnetic resonance imaging for differentiating metastatic from nonmetastatic mesorectal lymph nodes with different short-axis diameters in rectal cancer[J].Cancer Res Ther201915(7):1508-1515.
17
Ge YXHu SDWang Z,et al.Feasibility and reproducibility of T2 mapping and DWI for identifying malignant lymph nodes in rectal cancer[J].Eur Radiol202131(5):3347-3354.
18
许梅海,尹家瑜,申炜,等.单、双指数模型扩散加权成像对直肠癌转移性淋巴结诊断价值.放射学实践杂志202035(8):1042-1047.
19
Almlöv KWoisetschläger MLoftås P,et al.MRI Lymph Node Evaluation for Prediction of Metastases in Rectal Cancer[J].Anticancer Res202040(5):2757-2763.
20
Sofic AHusic-Selimovic AEfendic A,et al.MRI Evaluation of Extramural Venous Invasion(EMVI)with Rectal Carcinoma Using High Resolution T2 and Combination of High Resolution T2 and Contrast Enhanced T1 Weighted Imaging[J].Acta Inform Med202129(2):113-117.
21
Gormly KL.High-Resolution T2-Weighted MRI to Evaluate Rectal Cancer:Why Variations Matter[J].Korean J Radiol202122(9):1475-1480.
22
Chang KBai HXZhou H,et al.Residual convolutional neural network for the determination of IDH status in low-and high-grade gliomas from MR imaging[J].Clin Cancer Res201824(5):1073-1081.
23
汪豪,吉邦宁,何刚,等.一种提高直肠癌诊断精度的基于U型网络和残差块的电子计算机断层扫描图像分割算法[J].生物医学工程学杂志2022(1):166-174,184.
24
Esteva AKuprel BNovoa RA,et al.Dermatologist-level classification of skin cancer with deep neural networks[J].Nature2017542:115-118.
25
Paul RHawkins SHBalagurunathan Y,et al.Deep feature transfer learning in combination with traditional features predicts survival among patients with lung adenocarcinoma[J].Tomography20162:388-395.
26
Chen CMChou YHHan KC,et al.Breast lesions on sonograms:computer-aided diagnosis with nearly setting-independent features and artificial neural networks[J].Radiology2003226:504-514.
27
Wang JYang XCai H,et al.Discrimination of breast cancer with microcalcifications on mammography by deep learning[J].Sci Rep20166:27327.
28
Reda IShalaby AElmogy M,et al.A comprehensive non-invasive framework for diagnosing prostate cancer[J].Comput Biol Med201781:148-158.
29
周云朋.基于深度神经网络的高分辨MRI直肠淋巴结辅助诊断系统的临床应用价值研究[D].青岛:青岛大学,2019.
30
Li JinZhou YangWang Peng,et al.Deep transfer learning based on magnetic resonance imaging can improve the diagnosis of lymph node metastasis in patients with rectal cancer[J].Quant Imaging Med Surg202111(6):2477-2485.
[1] 王佳佳, 詹韵韵, 姜凡, 孙碧云, 毕玉, 李如冰, 彭梅. Peyton四步教学法在超声住院医师规范化培训颈部淋巴结分区中的应用[J/OL]. 中华医学超声杂志(电子版), 2024, 21(08): 814-818.
[2] 李乐平, 张荣华, 商亮. 腹腔镜食管胃结合部腺癌根治淋巴结清扫策略[J/OL]. 中华普外科手术学杂志(电子版), 2025, 19(01): 9-12.
[3] 赵丽霞, 王春霞, 陈一锋, 胡东平, 张维胜, 王涛, 张洪来. 内脏型肥胖对腹腔镜直肠癌根治术后早期并发症的影响[J/OL]. 中华普外科手术学杂志(电子版), 2025, 19(01): 35-39.
[4] 吴晖, 佴永军, 施雪松, 魏晓为. 两种解剖入路下行直肠癌侧方淋巴结清扫的效果比较[J/OL]. 中华普外科手术学杂志(电子版), 2025, 19(01): 40-43.
[5] 周世振, 朱兴亚, 袁庆港, 刘理想, 王凯, 缪骥, 丁超, 汪灏, 管文贤. 吲哚菁绿荧光成像技术在腹腔镜直肠癌侧方淋巴结清扫中的应用效果分析[J/OL]. 中华普外科手术学杂志(电子版), 2025, 19(01): 44-47.
[6] 高杰红, 黎平平, 齐婧, 代引海. ETFA和CD34在乳腺癌中的表达及与临床病理参数和预后的关系研究[J/OL]. 中华普外科手术学杂志(电子版), 2025, 19(01): 64-67.
[7] 徐逸男. 不同术式治疗梗阻性左半结直肠癌的疗效观察[J/OL]. 中华普外科手术学杂志(电子版), 2025, 19(01): 72-75.
[8] 李代勤, 刘佩杰. 动态增强磁共振评估中晚期低位直肠癌同步放化疗后疗效及预后的价值[J/OL]. 中华普外科手术学杂志(电子版), 2025, 19(01): 100-103.
[9] 刘敏思, 李荣, 李媚. 基于GGT与Plt比值的模型在HBV相关肝细胞癌诊断中的作用[J/OL]. 中华肝脏外科手术学电子杂志, 2024, 13(06): 831-835.
[10] 陆镜明, 韩大为, 任耀星, 黄天笑, 向俊西, 张谞丰, 吕毅, 王傅民. 基于术前影像组学的肝内胆管细胞癌淋巴结转移预测的系统性分析[J/OL]. 中华肝脏外科手术学电子杂志, 2024, 13(06): 852-858.
[11] 张润锦, 阳盼, 林燕斯, 刘尊龙, 刘建平, 金小岩. EB病毒相关胆管癌伴多发转移一例及国内文献复习[J/OL]. 中华肝脏外科手术学电子杂志, 2024, 13(06): 865-869.
[12] 刘郁, 段绍斌, 丁志翔, 史志涛. miR-34a-5p 在结肠癌患者的表达及其与临床特征及预后的相关性研究[J/OL]. 中华消化病与影像杂志(电子版), 2024, 14(06): 485-490.
[13] 张立俊, 孙存杰, 胡春峰, 孟冲, 张辉. MSCT、DCE-MRI 评估术前胃癌TNM 分期的准确性研究[J/OL]. 中华消化病与影像杂志(电子版), 2024, 14(06): 519-523.
[14] 孙铭远, 褚恒, 徐海滨, 张哲. 人工智能应用于多发性肺结节诊断的研究进展[J/OL]. 中华临床医师杂志(电子版), 2024, 18(08): 785-790.
[15] 崔军威, 蔡华丽, 胡艺冰, 胡慧. 亚甲蓝联合金属定位夹及定位钩针标记在乳腺癌辅助化疗后评估腋窝转移淋巴结的临床应用价值探究[J/OL]. 中华临床医师杂志(电子版), 2024, 18(07): 625-632.
阅读次数
全文


摘要