切换至 "中华医学电子期刊资源库"

中华损伤与修复杂志(电子版) ›› 2026, Vol. 21 ›› Issue (04) : 304 -308. doi: 10.3877/cma.j.issn.1673-9450.2026.04.011

综述

基于深度学习的创伤性肝损伤智能诊断与决策支持的研究进展
李雪晴, 罗俊, 徐振铎, 徐奇奇, 吴蔚, 高鹏骥()   
  1. 100096 首都医科大学附属北京积水潭医院 北京大学第四临床医学院普通外科
  • 收稿日期:2026-03-17 出版日期:2026-08-01
  • 通信作者: 高鹏骥
  • 基金资助:
    2026年首都卫生发展科研专项

Research progress on deep learning-based intelligent diagnosis and decision support for traumatic liver injury

Xueqing Li, Jun Luo, Zhenduo Xu, Qiqi Xu, Wei Wu, Pengji Gao()   

  1. Department of General Surgery,Beijing Jishuitan Hospital Affiliated to Capital Medical University,the Fourth Clinical Medical College of Peking University,Beijing 100096,China
  • Received:2026-03-17 Published:2026-08-01
  • Corresponding author: Pengji Gao
引用本文:

李雪晴, 罗俊, 徐振铎, 徐奇奇, 吴蔚, 高鹏骥. 基于深度学习的创伤性肝损伤智能诊断与决策支持的研究进展[J/OL]. 中华损伤与修复杂志(电子版), 2026, 21(04): 304-308.

Xueqing Li, Jun Luo, Zhenduo Xu, Qiqi Xu, Wei Wu, Pengji Gao. Research progress on deep learning-based intelligent diagnosis and decision support for traumatic liver injury[J/OL]. Chinese Journal of Injury Repair and Wound Healing(Electronic Edition), 2026, 21(04): 304-308.

创伤性肝损伤(TLI)是腹部钝性创伤中最常见、最危险的脏器损伤之一,早期、准确评估损伤范围和严重程度对制定合理治疗方案、改善患者预后有重要意义。目前,计算机断层扫描(CT)仍是评估肝损伤可靠的影像学方法,但CT图像的解读高度依赖放射科医师经验,存在主观性强、阅片耗时及高负荷下漏诊风险增高等问题。近年来,深度学习技术在医学影像分析领域取得了长足进步。该文探讨了由损伤识别、定量评估及伴随征象识别构成的TLI精准评估体系,以及通过成像优化与多模态整合实现的诊疗决策深度赋能,解构了深度学习模型在数据长尾效应、算法“黑箱”属性及临床工作流适配等方面面临的挑战,并提出了可解释性人工智能(AI)与边缘计算等未来研究方向,以期为AI技术在TLI诊疗中的规范化应用提供参考。

Traumatic liver injury (TLI) is one of the most common and life-threatening visceral injuries in blunt abdominal trauma. Early and accurate assessment of the extent and severity of liver injury is crucial for selecting appropriate treatment strategies and improving patient outcomes. Currently,computed tomography (CT) remains the most reliable imaging modality for evaluating liver injuries. However,CT interpretation is highly dependent on radiologists' expertise and is associated with challenges such as subjectivity,time-consuming image review,and an increased risk of missed diagnoses under heavy workloads. In recent years,deep learning has achieved remarkable progress in medical image analysis. This review explores a precision assessment framework for TLI based on injury detection,quantitative severity assessment,and identification of associated imaging findings. Furthermore,it discusses how diagnostic and therapeutic decision-making can be enhanced through imaging optimization and multimodal data integration. The challenges facing the clinical application of deep learning models,including the long-tail effect of medical data,the “black-box” nature of algorithms,and integration into clinical workflows,are also analyzed. Finally,future research directions,such as explainable artificial intelligence and edge computing,are proposed to promote the standardized application of artificial intelligence technologies in the diagnosis and management of TLI.

[1]
Sakran JVGreer SEWerlin E,et al. Care of the injured worldwide: trauma still the neglected disease of modern society[J] . Scand J Trauma Resusc Emerg Med201220: 64. DOI: 10.1186/1757-7241-20-64.
[2]
Becker CDMentha GTerrier F. Blunt abdominal trauma in adults: role of CT in the diagnosis and management of visceral injuries. Part 1: liver and spleen[J]. Eur Radiol19988(4): 553-562. DOI: 10.1007/s003300050433.
[3]
Wiederkehr HAWiederkehr JWiederkehr BA,et al. Liver trauma management[M]//Tsoulfas GMeshkini M. Trauma and emergency surgery-The role of damage control surgery. Intech Open,2021: 53-57. DOI: 10.5772/intechopen.92351.
[4]
Shanmuganathan K. Multi-detector row CT imaging of blunt abdominal trauma[J]. Semin Ultrasound CT MR200425(2): 180-204. DOI: 10.1016/j.sult.2004.02.002.
[5]
Abujudeh HHBoland GWKaewlai R,et al. Abdominal and pelvic computed tomography (CT) interpretation: discrepancy rates among experienced radiologists[J]. Eur Radiol201020(8): 1952-1957. DOI: 10.1007/s00330-010-1763-1.
[6]
Chen HUnberath MDreizin D. Toward automated interpretable AAST grading for blunt splenic injury[J]. Emerg Radiol202330(1): 41-50. DOI: 10.1007/s10140-022-02099-1.
[7]
Cao KXia Y, Yao J,et al. Large-scale pancreatic cancer detection via non-contrast CT and deep learning[J]. Nat Med202329(12): 3033-3043. DOI: 10.1038/s41591-023-02640-w.
[8]
Hamm CAWang CJSavic LJ,et al. Deep learning for liver tumor diagnosis part Ⅰ: development of a convolutional neural network classifier for multi-phasic MRI[J]. Eur Radiol201929(7): 3338-3347. DOI: 10.1007/s00330-019-06205-9.
[9]
Cheng CTOoyang CHLiao CH,et al. Applications of deep learning in trauma radiology: a narrative review[J]. Biomed J202548(1): 100743. DOI: 10.1016/j.bj.2024.100743.
[10]
Nam DChapiro JParadis V,et al. Artificial intelligence in liver diseases: improving diagnostics,prognostics and response prediction[J]. JHEP Rep Innov Hepatol20224(4): 100443. DOI: 10.1016/j.jhepr.2022.100443.
[11]
Dreizin DChen TLiang Y,et al. Added value of deep learning-based liver parenchymal CT volumetry for predicting major arterial injury after blunt hepatic trauma: a decision tree analysis[J]. Abdom Radiol N Y202146(6): 2556-2566. DOI: 10.1007/s00261-020-02892-x.
[12]
Pomohaci MDGrasu MCBăicoianu-Nițescu ,et al. Systematic review: AI applications in liver imaging with a focus on segmentation and detection[J]. Life Basel Switz202515(2): 258. DOI: 10.3390/life15020258.
[13]
Miotto RWang FWang S,et al. Deep learning for healthcare: review,opportunities and challenges[J]. Brief Bioinform201819(6): 1236-1246. DOI: 10.1093/bib/bbx044.
[14]
Yahaya BSOsman NDKarim NKA,et al. Radiomics and deep learning characterisation of liver malignancies in CT images-a systematic review[J]. Comput Biol Med2025194: 110491. DOI: 10.1016/j.compbiomed.2025.110491.
[15]
Shen XZhou YShi X,et al. The application of deep learning in abdominal trauma diagnosis by CT imaging[J]. World J Emerg Surg202419(1): 17. DOI: 10.1186/s13017-024-00546-7.
[16]
Cheng CTLin HHHsu CP,et al. Deep learning for automated detection and localization of traumatic abdominal solid organ injuries on CT scans[J]. J Imaging Inform Med202437(3): 1113-1123. DOI: 10.1007/s10278-024-01038-5.
[17]
Hermans SHu ZBall RL,et al. RSNA 2023 abdominal trauma AI challenge: review and outcomes[J]. Radiol Artif Intell20257(1): e240334. DOI: 10.1148/ryai.240334.
[18]
Brunese MCAvella PCappuccio M,et al. Future perspectives on radiomics in acute liver injury and liver trauma[J]. J Pers Med202414(6): 572. DOI: 10.3390/jpm14060572.
[19]
Isensee FJaeger PFKohl SAA,et al. nnU-Net: a self-configuring method for deep learning-based biomedical image segmentation[J]. Nat Methods202118(2): 203-211. DOI: 10.1038/s41592-020-01008-z.
[20]
Yu CAnakwenze CPZhao Y,et al. Multi-organ segmentation of abdominal structures from non-contrast and contrast enhanced CT images[J]. Sci Rep202212(1): 19093. DOI: 10.1038/s41598-022-21206-3.
[21]
Farzaneh NStein EBSoroushmehr R,et al. A deep learning framework for automated detection and quantitative assessment of liver trauma[J]. BMC Med Imaging202222(1): 39. DOI: 10.1186/s12880-022-00759-9.
[22]
Huang SZhou ZQian X,et al. Automated quantitative assessment of pediatric blunt hepatic trauma by deep learning-based CT volumetry[J]. Eur J Med Res202227(1): 305. DOI: 10.1186/s40001-022-00943-1.
[23]
Dreizin DZhou YFu S,et al. A multiscale deep learning method for quantitative visualization of traumatic hemoperitoneum at CT: assessment of feasibility and comparison with subjective categorical estimation[J]. Radiol Artif Intell20202(6): e190220. DOI: 10.1148/ryai.2020190220.
[24]
Dreizin DZhou YZhang Y,et al. Performance of a deep learning algorithm for automated segmentation and quantification of traumatic pelvic hematomas on CT[J]. J Digit Imaging202033(1): 243-251. DOI: 10.1007/s10278-019-00207-1.
[25]
Azarfar GKo SBAdams SJ,et al. Applications of deep learning to reduce the need for iodinated contrast media for CT imaging: a systematic review[J]. Int J Comput Assist Radiol Surg202318(10): 1903-1914. DOI: 10.1007/s11548-023-02862-w.
[26]
Liu JTian YDuzgol C,et al. Virtual contrast enhancement for CT scans of abdomen and pelvis[J]. Comput Med Imaging Graph2022100: 102094. DOI: 10.1016/j.compmedimag.2022.102094.
[27]
Choi JWCho YJHa JY,et al. Generating synthetic contrast enhancement from non-contrast chest computed tomography using a generative adversarial network[J]. Sci Rep202111(1): 20403. DOI: 10.1038/s41598-021-00058-3.
[28]
Haubold JHosch RUmutlu L,et al. Contrast agent dose reduction in computed tomography with deep learning using a conditional generative adversarial network[J]. Eur Radiol202131(8): 6087-6095. DOI: 10.1007/s00330-021-07714-2.
[29]
Jensen CTGupta SSaleh MM,et al. Reduced-dose deep learning reconstruction for abdominal CT of liver metastases[J]. Radiology2022303(1): 90-98. DOI: 10.1148/radiol.211838.
[30]
Zhen SHCheng MTao YB,et al. Deep learning for accurate diagnosis of liver tumor based on magnetic resonance imaging and clinical data[J]. Front Oncol202010: 680. DOI: 10.3389/fonc.2020.00680.
[31]
Wang YZhang JLi M,et al. SMART: development and application of a multimodal multi-organ trauma screening model for abdominal injuries in emergency settings[J]. Acad Radiol202532(5): 2655-2666. DOI: 10.1016/j.acra.2024.11.056.
[32]
Khader FMüller-Franzes GWang T,et al. Multimodal deep learning for integrating chest radiographs and clinical parameters: a case for transformers[J] . Radiology2023309(1): e230806. DOI: 10.1148/radiol.230806.
[33]
Salahuddin ZWoodruff HCChatterjee A,et al. Transparency of deep neural networks for medical image analysis: a review of interpretability methods[J]. Comput Biol Med2022140: 105111. DOI: 10.1016/j.compbiomed.2021.105111.
[34]
Agrawal AKhatri GDKhurana B,et al. A survey of ASER members on artificial intelligence in emergency radiology: trends,perceptions,and expectations[J]. Emerg Radiol202330(3): 267-277. DOI: 10.1007/s10140-023-02121-0.
[1] 李牧言, 张树庚, 刘连新. 腹腔镜全尾状叶切除术[J/OL]. 中华普外科手术学杂志(电子版), 2026, 20(04): 323-323.
[2] 刘荣. 机器人右半肝切除术[J/OL]. 中华普外科手术学杂志(电子版), 2026, 20(04): 324-324.
[3] 马禧圆, 母德安, 曾少杰, 张志远, 张伟. 超微创腹腔镜IV段肝肿瘤切除术[J/OL]. 中华普外科手术学杂志(电子版), 2026, 20(04): 325-325.
[4] 孟凡征, 王阳清, 张珅瑜, 刘连新. 人工智能在肝癌微创外科中的发展[J/OL]. 中华普外科手术学杂志(电子版), 2026, 20(04): 307-310.
[5] 王葵, 林宇轩, 沈皓, 沈锋. 人工智能化在肝癌微创外科的实践[J/OL]. 中华普外科手术学杂志(电子版), 2026, 20(04): 311-314.
[6] 叶子兴, 李红运, 李英杰, 王娟, 孙玉姣, 王惠珍, 贺蕾, 吕建阳, 张洪波, 张彤彤, 陈雯雯, 杨晓宇, 申川, 纪志刚. 深度学习模型在膀胱镜下膀胱肿瘤分割中的应用[J/OL]. 中华腔镜泌尿外科杂志(电子版), 2026, 20(04): 399-406.
[7] 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(2026年版)[J/OL]. 中华肝脏外科手术学电子杂志, 2026, 15(04): 443-507.
[8] 李海波, 李洋, 曾涛, 冯程, 李华. 基于3期RCT中国人群数据的肝细胞癌一线治疗方案网状Meta分析[J/OL]. 中华肝脏外科手术学电子杂志, 2026, 15(04): 558-567.
[9] 苟亚博, 肖晋昌, 神斌, 黄乾鑫, 魏宁, 刘洪涛, 徐浩, 祖茂衡, 张庆桥. 不同类型肝静脉型布加综合征介入治疗疗效分析[J/OL]. 中华肝脏外科手术学电子杂志, 2026, 15(04): 591-599.
[10] 汪祖宇, 刘笑安, 王汉譞, 吕少诚, 郎韧. 终末期良性肝病肝移植术后顽固性脾亢影响因素及预测模型构建[J/OL]. 中华肝脏外科手术学电子杂志, 2026, 15(04): 606-611.
[11] 林婉琪, 任芳菲, 钟伟航, 杨卿, 马玲, 陈海丽, 向莎, 陈孟娇, 朱丽萍, 何心念, 张献玲. 基于多维度评估的前瞻性干预策略在肝癌肝移植围手术期的应用及效果[J/OL]. 中华肝脏外科手术学电子杂志, 2026, 15(04): 612-618.
[12] 丁育, 林倩, 王志, 吴文睿, 张克林, 赵海璇. 肝移植术后早期感染风险简易评分模型构建及分级护理策略[J/OL]. 中华肝脏外科手术学电子杂志, 2026, 15(04): 619-626.
[13] 全宗宗, 张进, 秦一雨. 蛋白乳酸化修饰及其在肝胆胰肿瘤中的研究进展[J/OL]. 中华肝脏外科手术学电子杂志, 2026, 15(04): 648-652.
[14] 罗希, 汤永生, 鲁旭, 李华. 免疫时代中晚期肝癌治疗新格局:联合策略、转化突破与精准预测[J/OL]. 中华肝脏外科手术学电子杂志, 2026, 15(04): 508-514.
[15] 王孟龙. 门静脉高压症的治疗选择策略[J/OL]. 中华肝脏外科手术学电子杂志, 2026, 15(04): 523-528.
阅读次数
全文


摘要


AI


AI小编
你好!我是《中华医学电子期刊资源库》AI小编,有什么可以帮您的吗?