• 中国核心期刊数据库收录期刊
  • 中文科技期刊数据库收录期刊
  • 中国期刊全文数据库收录期刊
  • 中国学术期刊综合评价数据库统计源期刊等

快速检索引用检索图表检索高级检索

临床医药

基于AI算法的超声引导胸椎旁阻滞个体化给药策略在腔镜甲状腺手术中的应用研究

  • 梁艳山 ,
  • 陈杰 ,
  • 孙志强 ,
  • 狄飞 ,
  • 吴韬 ,
  • 牟洋 ,
  • 曹兴华
展开
  • 新疆维吾尔自治区中医医院麻醉手术中心1,超声中心2,新疆 乌鲁木齐 830000
梁艳山,男,主治医师,研究方向:手术麻醉

收稿日期: 2025-07-25

  修回日期: 2026-01-14

  录用日期: 2026-07-22

  网络出版日期: 2026-10-09

基金资助

新疆医科大学附属中医医院院级课题(ZYY202013)

Application of AI Algorithm Based Individualized Medication Strategy for Ultrasound-Guided Thoracic Paravertebral Block in Endoscopic Thyroidectomy

Expand
  • Department of Anesthesia and Surgery Center1, Department of Ultrasound Center2, Xinjiang Uygur Autonomous
    Region Hospital of Traditional Chinese Medicine, Xinjiang Urumqi 830000, China

Received date: 2025-07-25

  Revised date: 2026-01-14

  Accepted date: 2026-07-22

  Online published: 2026-10-09

摘要

目的:构建基于人工智能(AI)算法的超声引导胸椎旁神经阻滞(TPVB)个体化给药预测模型,并评估其在腔镜甲状腺手术中的临床应用效果。方法:采用前瞻性随机对照研究设计,纳入2022年6月至2024年3月在我院行腋窝入路腔镜甲状腺手术的患者360例,随机分为AI指导组(n=180)和常规组(n=180)。AI指导组采用本研究组前期基于独立历史数据开发的集成学习预测模型(整合随机森林、支持向量机和深度神经网络算法),根据患者42项特征变量预测最适阻滞药物剂量;常规组采用传统经验给药方案。结果:AI指导组阻滞成功率高于常规组(96.7% vs. 87.2%, P<0.001),术中舒芬太尼用量减少43.2%(P<0.001),术后2、6、12、24 h疼痛评分均降低(P<0.001)。AI指导组局麻药用量减少28.5%(P<0.001),不良反应发生率低于常规组(P<0.05),住院时间缩短0.8 d(P<0.001)。AI指导组的患者满意度评分为(9.2±0.6)分,高于常规组的(8.1±0.9)分(P<0.001)。结论:基于AI算法的超声引导TPVB个体化给药策略能够提高阻滞成功率,改善镇痛效果,减少药物用量和不良反应,为腔镜甲状腺手术围手术期疼痛管理提供了精准化解决方案。


本文引用格式

梁艳山 , 陈杰 , 孙志强 , 狄飞 , 吴韬 , 牟洋 , 曹兴华 .

基于AI算法的超声引导胸椎旁阻滞个体化给药策略在腔镜甲状腺手术中的应用研究

[J]. 中国医药导刊, 2026 , 28(7) : 826 -826-831 . DOI: 10.1009-0959.2026.080001

Abstract

Objective: To develop an artificial intelligence (AI) algorithm based individualized medication prediction model for ultrasound-guided thoracic paravertebral block (TPVB) and evaluate its clinical application effect in endoscopic thyroidectomy.Methods: A prospective randomized controlled study was conducted, and 360 patients who underwent axillary approach endoscopic thyroidectomy in our hospital from June 2022 to March 2024 were included. The patients were randomly divided into the AI-guided group (n=180) and the conventional group (n=180). The AI-guided group utilized a pre-developed ensemble learning prediction model (integrating RF, SVM, and DNN algorithms), which was previously trained on independent historical data by our research team, to predict optimal anesthetic dosage based on 42 feature variables. The conventional group received standard empirical dosing.Results: The AI-guided group showed significantly higher block success rate compared to the conventional group (96.7% vs. 87.2%, P<0.001), with 43.2% reduction in intraoperative sufentanil consumption (P<0.001) and significantly lower pain scores at 2, 6, 12, 24 h after the operation (P<0.001). Compared with the conventional group, the AI-guided group demonstrated 28.5% reduction in local anesthetic consumption (P<0.001), lower adverse reaction incidence rate (P<0.05), and 0.8 d shorter hospital stay (P<0.001). The patient satisfaction score of the AI-guided group was (9.2±0.6) points, which was higher than that of the conventional group [(8.1±0.9) points] (P<0.001).Conclusion: AI algorithm based individualized medication strategy for ultrasound-guided TPVB can improve block success rate, enhance analgesic efficacy, reduce drug consumption and adverse reactions, and provide a precision solution for perioperative pain management in endoscopic thyroidectomy.


参考文献

        [1] 郑传铭,毛晓春,王佳峰,等.无充气腋窝入路完全腔镜下甲状腺癌根治术效果初步评价[J].中国肿瘤临床,2018,45(1):27-32.

         [2]  Li Xiaosong, Ding Wei, Zhang Hong. Surgical outcomes of endoscopic thyroidectomy approaches for thyroid cancer: a systematic review and network meta-analysis[J].Front Endocrinol (Lausanne), 2023,14:1256209.

         [3]  储勤军,陈向东,彭云水,等.成人术后疼痛管理临床实践指南(2024版)计划书[J].中华麻醉学杂志,2024,44(9):1069-1074.

         [4]  Forero M, Adhikary SD, Lopez H, et al. The erector spinae plane block: a novel analgesic technique in thoracic neuropathic pain[J].Reg Anesth Pain Med, 2016,41(5):621-627.

         [5]  Elsharkawy H, Pawa A, Mariano ER. Interfascial plane blocks: back to basics[J].Reg Anesth Pain Med, 2018,43(4):341-346.

         [6]  Hashimoto DA, Witkowski E, Gao L, et al. Artificial intelligence in anesthesiology: current techniques, clinical applications, and limitations[J].Anesthesiology, 2020,132(2):379-394.

         [7]  Hashimoto DA, Rosman G, Rus D, et al. Artificial intelligence in surgery: promises and perils[J].Ann Surg, 2018,268(1):70-76.

         [8]  Topol EJ. High-performance medicine: the convergence of human and artificial intelligence[J].Nat Med, 2019,25(1):44-56.

         [9]  Terkawi AS, Tsang S, Sessler DI, et al. Improving analgesic efficacy and safety of thoracic paravertebral block for breast surgery: a mixed-effects meta-analysis[J].Pain Physician, 2015,18(5):E757-E780.

         [10] Krediet AC, Moayeri N, van Geffen GJ, et al. Different approaches to ultrasound-guided thoracic paravertebral block: an illustrated review[J].Anesthesiology, 2015,123(2):459-474.

         [11] Hu YJ, Ku TH, Jan Ronghong, et al. Decision tree-based learning to predict patient controlled analgesia consumption and readjustment[J].BMC Med Inform Decis Mak, 2012,12:131.

         [12] Tighe PJ, Harle CA, Hurley RW, et al. Teaching a machine to feel postoperative pain: combining high-dimensional clinical data with machine learning algorithms to forecast acute postoperative pain[J].Pain Med, 2015,16(7):1386-1401.

         [13] Wei Changna, Wang Liying, Chang Xiangyang, et al. A prediction model using machine-learning algorithm for assessing intrathecal hyperbaric bupivacaine dose during cesarean section[J].BMC Anesthesiol, 2021,21(1):116.

         [14] Richardson J, Lönnqvist PA, Naja Z. Bilateral thoracic paravertebral block: potential and practice[J].Br J Anaesth, 2011,106(2):164-171.

         [15] Neal JM, Barrington MJ, Fettiplace MR, et al. The third American Society of Regional Anesthesia and Pain Medicine Practice Advisory on local anesthetic systemic toxicity: executive summary 2017[J].Reg Anesth Pain Med, 2018,43(2):113-123.

         [16] Zhang Zeyu, Xia Fada, Li Xinying. Ambulatory endoscopic thyroidectomy via a chest-breast approach has an acceptable safety profile for thyroid nodule[J].Front Endocrinol (Lausanne), 2021, 12:795627.

         [17] Bomberg H, Wetjen L, Wagenpfeil S, et al. Risks and benefits of ultrasound, nerve stimulation, and their combination for guiding peripheral nerve blocks: a retrospective registry analysis[J].Anesth Analg, 2018,127(4):1035-1043.

         [18] Shillan D, Sterne JA, Champneys A, et al. Use of machine learning to analyse routinely collected intensive care unit data: a systematic review[J].Crit Care, 2019,23(1):284.

         [19] Fritz BA, Maybrier HR, Avidan MS. Intraoperative electroencephalogram suppression at lower volatile anaesthetic concentrations predicts postoperative delirium occurring in the intensive care unit[J].Br J Anaesth, 2018,121(1):241-248.

         [20] Lopes S, Rocha G, Guimarães-Pereira L. Artificial intelligence and its clinical application in anesthesiology: a systematic review[J].J Clin Monit Comput, 2024,38(2):247-259.

         [21] Connor CW. Artificial intelligence and machine learning in anesthesiology[J].Anesthesiology, 2019,131(6):1346-1359.

文章导航

/

〈 〉