现代仪器与医疗

面向LR-M肝癌精准诊断的超声造影时空感知深度学习系统开发

PDF下载

林雅丹 ¹ , 阮思敏 ² , 陈立达 ² , 何云 ¹^,^*

现代仪器与医疗 | 医学工程 2025,31(3): 1-9

收起

现代仪器与医疗 |医学工程 2025 , 31 (3) : 1 -9

面向LR-M肝癌精准诊断的超声造影时空感知深度学习系统开发

全屏

林雅丹¹, 阮思敏², 陈立达², 何云¹^,^*

作者信息

^1.广西医科大学第一附属医院超声医学科，南宁　530021

^2.中山大学附属第一医院超声医学科，广州　510080

通讯作者:

何云，女，硕士研究生，主任医师，研究方向：医学影像与人工智能

作者简介:

林雅丹，女，硕士研究生，研究方向：医学影像与人工智能

Development of a spatiotemporal perception deep learning system based on contrast-enhanced ultrasound for precise diagnosis of LR-M hepatocellular carcinoma

Yadan Lin¹, Simin Ruan², Lida Chen², Yun He¹^,^*

Affiliations

^1.Department of Medical Ultrasound, the First Affiliated Hospital of Guangxi Medical University, Nanning 530021, China

^2.Department of Medical Ultrasound, the First Affiliated Hospital of Sun Yat-sen University, Guangzhou 510080, China

出版时间: 2025-06-10 doi: 10.11876/mimt202503001

文章导航

摘要

收起

目的

利用深度学习（deep learning，DL）结合超声造影（contrast-enhanced ultrasound，CEUS）视频开发一个时空混合感知的DL模型，从肝脏影像报告与数据系统（liver imaging reporting and data system，LI-RADS）LR-M类肝结节中无创诊断肝细胞癌（hepatocellular carcinoma，HCC）。

方法

对来自中山大学附属第一医院和广西医科大学第一附属医院的经CEUS诊断为LR-M类的患者进行回顾性分析，以病理和综合诊断为标准，分为HCC组和非HCC组。使用R（2+1）D-18卷积神经网络对CEUS视频进行多尺度时空分时相并行特征提取，构建HCC诊断模型，并通过帧采样策略（15帧、30帧、45帧）平衡信息完整性与计算效率，辅以梯度加权分类激活映射技术（gradient-weighted class activation mapping，Grad-CAM）可视化解释决策依据。使用准确率、敏感度、特异度、阳性预测值、阴性预测值和受试者工作特征（receiver operating characteristic，ROC）曲线来评估模型的诊断效能。

结果

来自中山大学附属第一医院的176例LR-M患者按7∶3划分训练集和验证集；来自广西医科大学第一附属医院的46例LR-M患者为测试集。当以45帧视频作为输入时，模型诊断效能最佳，在验证集和测试集中，敏感度为91.67%和84.21%，特异度为96.43%和81.48%，曲线下面积（area under curve，AUC）为0.940和0.828。与甲胎蛋白（alpha-fetoprotein，AFP）指标相比，模型对HCC的检出率更高（验证集：50.00%，测试集：31.58%；P值均<0.05），联合使用DL模型与AFP可以优化整体诊断效能。

结论

基于CEUS时空感知的深度学习模型可以精准诊断LR-M类中的HCC，诊断效能显著优于AFP，两者联合诊断有望减少不必要的肝穿刺活检。

关键词

超声造影 / 肝脏影像报告与数据系统 / 肝细胞癌 / 深度学习 / 诊断效能1

Abstract

收起

Objective

To develop a spatiotemporal hybrid perception deep learning (DL) model using contrast-enhanced ultrasound (CEUS) videos for diagnosis of hepatocellular carcinoma (HCC) in liver imaging reporting and data system (LI-RADS) category LR-M liver nodules.

Methods

This study conducted a retrospective analysis of patients diagnosed with LR-M nodules by CEUS from the First Affiliated Hospital of Sun Yat-sen University and the First Affiliated Hospital of Guangxi Medical University. Using pathological and comprehensive diagnostic results as the gold standard, the patients were categorized into HCC and non-HCC. A multi-scale spatiotemporal split-phase parallel feature extraction of CEUS video using R(2+1)D-18 convolutional neural network is used to construct a diagnostic model for HCC, and the information integrity and computational efficiency are balanced by a frame sampling strategy (15/30/45 frames), supplemented by gradient-weighted class activation mapping (Grad-CAM) visualisation to explain the decision basis. The diagnostic efficacy of the model is evaluated using accuracy, sensitivity, specificity, positive predictive value, negative predictive value and receiver operating characteristic (ROC) curve.

Results

A total of 176 LR-M patients from the First Affiliated Hospital of Sun Yat-sen University were divided into training and validation sets in a 7:3 ratio, while 46 LR-M patients from the First Affiliated Hospital of Guangxi Medical University served as the test set. When using 45-frame videos as input, with sensitivities of 91.67% and 84.21%, specificities of 96.43% and 81.48%, and area under curves (AUCs) of 0.940 and 0.828 in the validation and test sets. The model had a higher detection rate of HCC compared with the alpha-fetoprotein (AFP) index (validation set: 50.00%, test set: 31.58%; both P value < 0.05), and the combined use of the DL model with AFP could optimize the overall diagnostic efficacy.

Conclusion

The DL model with CEUS spatiotemporal perception enables accurate diagnosis of HCC in LR-M category liver lesions, demonstrating significantly superior performance over AFP. Moreover, the combined diagnosis of the both can reduce unnecessary liver biopsy.

Key words

Contrast-enhanced ultrasound / Liver imaging reporting and data system / Hepatocellular carcinoma / Deep learning / Diagnostic performance

引用本文

林雅丹, 阮思敏, 陈立达, 何云. 面向LR-M肝癌精准诊断的超声造影时空感知深度学习系统开发. 现代仪器与医疗, 2025 , 31 (3) : 1 -9 . DOI: 10.11876/mimt202503001

Yadan Lin, Simin Ruan, Lida Chen, Yun He. Development of a spatiotemporal perception deep learning system based on contrast-enhanced ultrasound for precise diagnosis of LR-M hepatocellular carcinoma[J]. Modern Instruments & Medical Treatment, 2025 , 31 (3) : 1 -9 . DOI: 10.11876/mimt202503001

基金

收起

国家自然科学基金(82272076)

参考文献

收起

文献

收起

参考文献引证文献

排序方式：

[1]

Bray F, Laversanne M, Sung H, et al. Global cancer statistics 2022: GLOBOCAN estimates of incidence and mortality worldwide for 36 cancers in 185 countries[J]. CA Cancer J Clin, 2024, 74(3): 229-263.

[2]

Xie D, Shi J, Zhou J, et al. Clinical practice guidelines and real-life practice in hepatocellular carcinoma: a Chinese perspective[J]. Clinical and Molecular Hepatology, 2023, 29(2): 206-216.

[3]

Vitale A, Trevisani F, Farinati F, et al. Treatment of hepatocellular carcinoma in the precision medicine era: from treatment stage migration to therapeutic hierarchy[J]. Hepatology, 2020, 72(6): 2206-2218.

[4]

中华人民共和国国家卫生健康委员会医政司. 原发性肝癌诊疗指南(2024年版)[J]. 中国实用外科杂志, 2024, 44(4): 361-386.

[5]

Chernyak V, Fowler K J, Kamaya A, et al. Liver imaging reporting and data system (LI-RADS) version 2018: imaging of hepatocellular carcinoma in at-risk patients[J]. Radiology, 2018, 289(3): 816-830.

[6]

Huang J Y, Li J W, Lu Q, et al. Diagnostic accuracy of CEUS LI-RADS for the characterization of liver nodules 20 mm or smaller in patients at risk for hepatocellular carcinoma[J]. Radiology, 2020, 294(2): 329-339.

[7]

Terzi E, Iavarone M, Pompili M, et al. Contrast ultrasound LIRADS LR-5 identifies hepatocellular carcinoma in cirrhosis in a multicenter restropective study of 1,006 nodules[J]. J Hepatol, 2018, 68(3): 485-492.

[8]

Chan S L, Mo F, Johnson P J, et al. Performance of serum alpha-fetoprotein levels in the diagnosis of hepatocellular carcinoma in patients with a hepatic mass[J]. HPB (Oxford), 2014, 16(4): 366-372.

[9]

Li C Q, Huang H, Ruan S M, et al. An assessment of liver lesions using a combination of CEUS LI-RADS and AFP[J]. Abdominal Radiology (New York), 2022, 47(4): 1311-1320.

[10]

Huang W, Wen R, Wu Y, et al. Can modifications of LR-M criteria improve the diagnostic performance of contrast-enhanced ultrasound LI-RADS for small hepatic lesions up to 3 cm?[J]. J Ultrasound Med, 2023, 42(10): 2403-2413.

[11]

Georgevici A I, Terblanche M. Neural networks and deep learning: a brief introduction[J]. Intensive Care Medicine, 2019, 45(5): 712-714.

[12]

Hosny A, Parmar C, Quackenbush J, et al. Artificial intelligence in radiology[J]. Nature Reviews Cancer, 2018, 18(8): 500-510.

[13]

Hu H T, Wang W, Chen L D, et al. Artificial intelligence assists identifying malignant versus benign liver lesions using contrast-enhanced ultrasound[J]. Journal of Gastroenterology and Hepatology, 2021, 36(10): 2875-2883.

[14]

Liu F, Liu D, Wang K, et al. Deep learning radiomics based on contrast-enhanced ultrasound might optimize curative treatments for very-early or early-stage hepatocellular carcinoma patients[J]. Liver Cancer, 2020, 9(4): 397-413.

[15]

Ji S, Yang M, Yu K. 3D convolutional neural networks for human action recognition[J]. IEEE Trans Pattern Anal Mach Intell, 2013, 35(1): 221-231.

[16]

Tran D, Wang H, Torresani L, et al. A closer look at spatiotemporal convolutions for action recognition[C]//2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition, June 18-23, 2018, Salt Lake City, UT, USA. IEEE, 2018: 6450-6459.

[17]

He K, Zhang X, Ren S, et al. Deep residual learning for image recognition[C]//2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR), June 27-30, 2016, Las Vegas, NV, USA. IEEE, 2016: 770-778

[18]

Selvaraju R R, Cogswell M, Das A, et al. Grad-CAM: visual explanations from deep networks via gradient-based localization[C]//2017 IEEE International Conference on Computer Vision (ICCV), 2017, Venice, Italy. IEEE, 2017: 618-626.

[19]

Ding J, Qin Z, Zhou Y, et al. Impact of revision of the LR-M criteria on the diagnostic performance of contrast-enhanced ultrasound LI-RADS[J]. Ultrasound Med Biol, 2021, 47(12): 3403-3410.

[20]

Shao Y, Dang Y, Cheng Y, et al. Predicting the efficacy of neoadjuvant chemotherapy for pancreatic cancer using deep learning of contrast-enhanced ultrasound videos[J]. Diagnostics (Basel), 2023, 13(13): 2183.

[21]

Reyes M, Meier R, Pereira S, et al. On the interpretability of artificial intelligence in radiology: challenges and opportunities[J]. Radiol Artif Intell, 2020, 2(3): e190043.

[22]

Gong W, Wu J, Wei H, et al. Combining serum AFP and CEUS LI-RADS for better diagnostic performance in Chinese high-risk patients[J]. La Radiologia medica, 2023, 128(4): 393-401.

[23]

Wen R, Peng Y, Liang Y, et al. CEUS LI-RADS in combination with the serum biomarker-based ASAP model improves the diagnostic performance of HCC in high-risk patients[J]. Ultrasound Med Biol, 2024, 50(11): 1739-1744.

2025年第31卷第3期

PDF下载

引用本文

BibTeX

文章信息

doi: 10.11876/mimt202503001

首发时间：2026-10-23
出版时间：2025-06-10

补充材料

相关文章

文章信息

作者

出版历史

基金

国家自然科学基金(82272076)

作者信息

^1.广西医科大学第一附属医院超声医学科，南宁　530021

^2.中山大学附属第一医院超声医学科，广州　510080

通讯作者:

何云，女，硕士研究生，主任医师，研究方向：医学影像与人工智能

参考文献

分享链接

https://castjournals.cast.org.cn/joweb/xdyqyyl/CN/10.11876/mimt202503001

分享至

全文二维码

扫描看全文

引用本文

BibTeX

本文的引用情况

2种不同金属材料的力学参数

科 Family	属数 Number of genus	种数 Number of species	占总种数比例 Percentage of total species (%)	属 Genus	种数 Number of species	占总种数比例 Percentage of total species (%)
鹅膏菌科Amanitaceae	2	11	5.26	鹅膏菌属 Amanita	10	4.78
小菇科 Mycenaceae	2	12	5.74	丝盖伞属 Inocybe	5	2.39
多孔菌科 Polyporaceae	8	14	6.70	蜡蘑属 Laccaria	5	2.39
红菇科 Russulaceae	3	23	11.00	小皮伞属 Marasmius	6	2.87
				小菇属 Mycena	11	5.26
				光柄菇属 Pluteus	5	2.39
				红菇属 Russula	17	8.13
				栓菌属 Trametes	5	2.39

关闭全屏

BibTeX
EndNote
RefWorks
TxT