人工智能辅助CT阅片系统在综合医院肺结核早发现中的应用

doi:10.19982/j.issn.1000-6621.20250419

中国防痨杂志 ›› 2026, Vol. 48 ›› Issue (5): 586-593.doi: 10.19982/j.issn.1000-6621.20250419

人工智能辅助CT阅片系统在综合医院肺结核早发现中的应用

孙郑¹^,², 刘炯雅²^,⁴, 陈驰², 于全骥¹, 李亚男³, 陈诚¹(), 竺丽梅¹^,²()

¹江苏省疾病预防控制中心慢性传染病防制所, 南京 210009
²南京医科大学公共卫生学院流行病与卫生统计学系, 南京 211166
³南京医科大学第二附属医院公共卫生科, 南京 210000
⁴靖江市疾病预防控制中心慢病科, 靖江 214500

收稿日期:2025-10-30 出版日期:2026-05-10 发布日期:2026-04-27
通信作者: 陈诚,竺丽梅 E-mail:chencheng128@gmail.com;jsjkmck@163.com
基金资助:
江苏省“十四五”流行病学重点学科(ZDXK202250);江苏省老年健康科研项目(LKM2024005)

Study on the application of AI-assisted CT reading system in early detection of pulmonary tuberculosis in general hospitals

Sun Zheng¹^,², Liu Jiongya²^,⁴, Chen Chi², Yu Quanji¹, Li Yanan³, Chen Cheng¹(), Zhu Limei¹^,²()

¹Department of Chronic Infectious Disease Control, Jiangsu Provincial Center for Disease Control and Prevention, Nanjing 210009, China
²Department of Epidemiology and Biostatistics, School of Public Health, Nanjing Medical University, Nanjing 211166, China
³Depaertmen of Public Health, Second Affiliated Hospital of Nanjing Medical University, Nanjing 210000, China
⁴Department of Chronic Disease Control, Center for Disease Control and Prevention of Jingjiang, Jingjiang 214500, China

Received:2025-10-30 Online:2026-05-10 Published:2026-04-27
Contact: Chen Cheng,Zhu Limei E-mail:chencheng128@gmail.com;jsjkmck@163.com
Supported by:
“14th Five-Year” Key Discipline of Epidemiology of Jiangsu Province(ZDXK202250);Jiangsu Province Elderly Health Research Project(LKM2024005)

摘要/Abstract

摘要：

目的: 分析人工智能(artificial intelligence,AI)辅助影像诊断技术用于提高综合医院肺结核发现水平的价值。方法: 本研究采用前瞻性和回顾性相结合的研究方法,于2024年6月至2024年12月,选择江苏省人民医院和南京医科大学第二附属医院两家综合医院门诊拍摄胸部CT的75786例患者进行AI辅助筛查肺结核,随后影像科医生对AI疑似影像进行复核,排除掉非疑似肺结核患者后再由专家组进行复核确诊,最后共确诊259例肺结核患者。同时,利用二元逻辑回归方法分析患者临床特征和CT征象对AI分析效果的影响。结果: 在两家综合医院的75786张CT片中,AI以0.56作为阈值判定疑似肺结核,AI提示疑似肺结核的CT片为8574张,最终确诊肺结核259例,其中临床诊断患者196例,占75.68%(196/259),与同期传统流程报告数量相比发现率提高了102.34%[(259-128)/128]。确诊肺结核AI阈值中位数(四分位数)为0.65(0.59,0.69),被排除患者的阈值中位数(四分位数)为0.62(0.59,0.70),差异无统计学意义(U=1783.500,P=0.738)。钙化和纤维化在非肺结核患者中的比例更高,调整后比值比(aOR)分别为0.52(95%CI:0.32~0.84,P=0.008)、0.52(95%CI:0.31~0.86,P=0.011)。结论: AI阅片系统可显著提高肺结核患者的发现效率,降低漏诊率,但肺结核的诊断仍然要依赖临床综合诊断;钙化和纤维化影像学特征并不有助于肺结核的诊断。因此,AI辅助CT阅片在综合医院中具有较好的应用潜力。

关键词: 结核, 肺, 诊断, 鉴别, 人工智能, 体层摄影术, X线计算机, 医院

Abstract:

Objective: This study aimed to analyze the value of artificial intelligence (AI)-assisted imaging diagnostic technology in improving the detection rate of pulmonary tuberculosis in general hospitals. Methods: This prospective and retrospective study, conducted from June to December 2024, selected 75786 patients who underwent chest CT scans at the outpatient clinics of two general hospitals, Jiangsu Provincial People’s Hospital and the Second Affiliated Hospital of Nanjing Medical University, for AI-assisted screening for pulmonary tuberculosis. Subsequently, radiologists reviewed the AI-suspected images, excluding those without suspected pulmonary tuberculosis, before a panel of experts confirmed the diagnosis. A total of 259 patients were ultimately diagnosed with pulmonary tuberculosis. Simultaneously, we used binary logistic regression to analyze the impact of clinical characteristics and CT findings on the effectiveness of AI analysis, aiming to demonstrate the effectiveness of AI in detecting pulmonary tuberculosis in actual clinical practice. Results: Among 75786 CT scans from two general hospitals, AI (AI-guided diagnostic tool) used a threshold of 0.56 to identify suspected pulmonary tuberculosis, AI indicated 8574 scans as potentially indicating pulmonary tuberculosis, ultimately confirming 259 cases. Among these, 196 were clinically diagnosed, accounting for 75.68% (196/259), representing a 102.34% ((259-128)/128) increase in detection rate compared to the number of reported cases during the same period (χ²=41.800, P<0.001). The median (interquartile range) AI threshold for confirmed pulmonary tuberculosis was 0.65 (0.59, 0.69), while the median threshold for excluded patients was 0.62 (0.59, 0.70), with no statistically significant difference (U=1783.500, P=0.738). Calcification and fibrosis were more prevalent in non-pulmonary tuberculosis cases, with adjusted odds of 0.52 (95%CI: 0.32-0.84, P=0.008) and 0.52 (95%CI: 0.31-0.86, P=0.011), respectively. Conclusion: AI-assisted CT image reading systems can significantly improve the detection efficiency of pulmonary tuberculosis patients and reduce the missed diagnosis rate; however, the diagnosis of pulmonary tuberculosis still relies on comprehensive clinical diagnosis. Calcification and fibrosis imaging features are not helpful in the diagnosis of pulmonary tuberculosis. Therefore, AI-assisted CT image reading has good application potential in general hospitals.

Key words: Tuberculosis, pulmonary, Diagnosis, differential, Artificial intelligence, Tomography, X-ray computed, Hospitals

中图分类号:

孙郑, 刘炯雅, 陈驰, 于全骥, 李亚男, 陈诚, 竺丽梅. 人工智能辅助CT阅片系统在综合医院肺结核早发现中的应用[J]. 中国防痨杂志, 2026, 48(5): 586-593. doi: 10.19982/j.issn.1000-6621.20250419

Sun Zheng, Liu Jiongya, Chen Chi, Yu Quanji, Li Yanan, Chen Cheng, Zhu Limei. Study on the application of AI-assisted CT reading system in early detection of pulmonary tuberculosis in general hospitals[J]. Chinese Journal of Antituberculosis, 2026, 48(5): 586-593. doi: 10.19982/j.issn.1000-6621.20250419

图/表 6

图1

表1

表2

表3

表4

表5

参考文献 20

[1]	全国第五次结核病流行病学抽样调查技术指导组, 全国第五次结核病流行病学抽样调查办公室. 2010年全国第五次结核病流行病学抽样调查报告. 中国防痨杂志, 2012, 34(8): 485-508.
[2]	国家呼吸内科医疗质量控制中心, 中华医学会结核病学分会, 中国防痨协会结核病控制专业分会, 等. 综合医疗机构肺结核早期发现临床实践指南. 结核与肺部疾病杂志, 2024, 5(1): 1-14. doi:10.19982/j.issn.1000-6621.20230428.
[3]	Baek JY, Kazmi SZ, Lee H, et al. Impact of Mass Screening Using Chest X-Ray on Mortality Reduction and Treatment Adherence Among Pulmonary Tuberculosis Patients. J Korean Med Sci, 2024, 39(45): e286. doi:10.3346/jkms.2024.39.e286. URL pmid: 39592127
[4]	Hwang EJ, Jeong WG, David PM, et al. AI for Detection of Tuberculosis: Implications for Global Health. Radiol Artif Intell, 2024, 6(2): e230327. doi:10.1148/ryai.230327.
[5]	Mbulayi O, Djungu SJ, Aketi L, et al. Tuberculosis diagnosis using artificial intelligence: current trends and future prospects. Front Med (Lausanne), 2025, 12:1569615. doi:10.3389/fmed.2025.1569615.
[6]	Murphy K, Habib SS, Zaidi SMA, et al. Computer aided detection of tuberculosis on chest radiographs: An evaluation of the CAD4TB v6 system. Sci Rep, 2020, 10(1): 5492. doi:10.1038/s41598-020-62148-y. URL pmid: 32218458
[7]	汪犇, 陈玮, 李进岚, 等. 贵州省毕节市学生肺结核密切接触者筛查结果分析. 中国预防医学杂志, 2023, 24(9): 906-911. doi:10.16506/j.1009-6639.2023.09.005.
[8]	Van’t Hoog A, Viney K, Biermann O, et al. Symptom- and chest-radiography screening for active pulmonary tuberculosis in HIV-negative adults and adults with unknown HIV status. Cochrane Database Syst Rev, 2022, 3(3): Cd010890. doi:10.1002/14651858.CD010890.pub2.
[9]	中华人民共和国国家卫生和计划生育委员会. WS 196—2017结核病分类. 结核与肺部疾病杂志, 2024, 5(4): 379-380. doi:10.19983/j.issn.2096-8493.2024055.
[10]	中华人民共和国国家卫生和计划生育委员会. WS 288—2017肺结核诊断. 结核与肺部疾病杂志, 2024, 5(4): 376-378. doi:10.19983/j.issn.2096-8493.2024022.
[11]	Uplekar M, Weil D, Lonnroth K, et al. WHO’s new end TB strategy. Lancet, 2015, 385(9979): 1799-801. doi:10.1016/s0140-6736(15)60570-0. URL pmid: 25814376
[12]	Xin H, Wang W, Wang X, et al. The performance of computer-aided detection for chest radiography in tuberculosis screening: a population-based retrospective cohort study. Emerg Microbes Infect, 2025, 14(1): 2470998. doi:10.1080/22221751.2025.2470998.
[13]	胡鑫洋, 高静韬. 世界卫生组织《2024年全球结核病报告》解读. 结核与肺部疾病杂志, 2024, 5(6): 500-504. doi:10.19983/j.issn.2096-8493.2024164.
[14]	Innes AL, Martinez A, Gao X, et al. Computer-Aided Detection for Chest Radiography to Improve the Quality of Tuberculosis Diagnosis in Vietnam’s District Health Facilities: An Implementation Study. Trop Med Infect Dis, 2023, 8(11): 488. doi:10.3390/tropicalmed8110488.
[15]	马依迪丽·尼加提, 阿里木江·阿卜杜凯尤木, 米日古丽·达毛拉, 等. 基于人工智能肺结核筛查技术在基层医院影像诊断中的应用价值. 新发传染病电子杂志, 2021, 6(2): 138-142. doi:10.19871/j.cnki.xfcrbzz.2021.02.014.
[16]	倪晗, 于全骥, 邵燕, 等. 2014—2024年江苏省学生肺结核流行病学特征分析. 江苏预防医学, 2025, 36(2): 160-163. doi:10.13668/j.issn.1006-9070.2025.02.007.
[17]	Shen L, Zhang Y, Xin H, et al. Association of inactive pulmonary tuberculosis lesions with the risk of active disease development in the elderly: a population-based retrospective study. ERJ Open Res, 2025, 11(4): 01231-2024. doi:10.1183/23120541.01231-2024.
[18]	Bagcchi S. WHO’s Global Tuberculosis Report 2022. Lancet Microbe, 2023, 4(1): e20. doi:10.1016/s2666-5247(22)00359-7.
[19]	Targeted tuberculin testing and treatment of latent tuberculosis infection. This official statement of the American Thoracic Society was adopted by the ATS Board of Directors, July 1999. This is a Joint Statement of the American Thoracic Society (ATS) and the Centers for Disease Control and Prevention (CDC). This statement was endorsed by the Council of the Infectious Diseases Society of America. (IDSA), September 1999, and the sections of this statement. Am J Respir Crit Care Med, 2000, 161(4 Pt 2): S221-S247. doi:10.1164/ajrccm.161.supplement_3.ats600.
[20]	Innes AL, Martinez A, Hoang GL, et al. Active case finding to detect symptomatic and subclinical pulmonary tuberculosis disease: implementation of computer-aided detection for chest radiography in Viet Nam. Western Pac Surveill Response J, 2024, 15(4): 1-12. doi:10.5365/wpsar.2024.15.4.1118. URL pmid: 39416596

年龄组(岁)	男性(例)	构成比(%)	女性(例)	构成比(%)	合计	构成比(%)
15~24	1043	2.92	1080	2.72	2123	2.80
25~34	2487	6.97	3497	8.81	5985	7.90
35~44	4375	12.27	5941	14.97	10316	13.61
45~54	5421	15.20	7371	18.57	12792	16.88
55~64	9176	25.73	10236	25.79	19412	25.62
65~74	8469	23.75	7652	19.28	16335	21.55
75~84	3770	10.57	3044	7.66	7026	9.27
≥85	923	2.59	874	2.20	1797	2.37
合计	35664	100.00	39695	100.00	75786	100.00

指标	A医院	B医院	合计
AI阅片系统辅助CT阅片(例)	44247	31539	75786
AI阅片系统初筛疑似肺结核(例)	5225	3349	8574
影像医生复核疑似肺结核(例)	257	105	362
确诊肺结核(例)	190	69	259
临床诊断	136	60	196
病原学诊断	54	9	63
2024年同期传统流程报告的肺结核患者(例)	67	61	128
肺结核患者发现增长率(%)	183.53	13.11	102.34

特征	非肺结核患者(103例)	肺结核患者(259例)	统计检验值	P值
年龄(岁,$\overline{x}$±s)	61.27±14.55	61.76±16.32	t=-0.262	0.793
性别[例(构成比,%)]			χ²=0.116	0.734
男性	66(64.08)	161(62.16)
女性	37(35.92)	98(37.84)
吸烟[例(构成比,%)]			χ²=1.175	0.278
否	84(81.55)	223(86.10)
是	19(18.45)	36(13.90)
饮酒[例(构成比,%)]			χ²=0.601	0.438
否	92(89.32)	238(91.89)
是	11(10.68)	21(8.11)
结核病史[例(构成比,%)]			χ²=0.730	0.393
无	99(96.12)	243(93.82)
有	4(3.88)	16(6.18)
糖尿病[例(构成比,%)]			χ²<0.001	0.987
无	93(90.29)	234(90.35)
有	10(9.71)	25(9.65)
高血压[例(构成比,%)]			χ²=0.583	0.445
无	79(76.70)	208(80.31)
有	24(23.30)	51(19.69)
肺炎[例(构成比,%)]			χ²=1.631	0.202
否	101(98.06)	246(94.98)
是	2(1.94)	13(5.02)
慢性阻塞性肺疾病[例(构成比,%)]			χ²=1.075	0.300
否	101(98.06)	248(95.75)
是	2(1.94)	11(4.25)
咳嗽[例(构成比,%)]			χ²=2.103	0.147
否	82(79.61)	187(72.20)
是	21(20.39)	72(27.80)
咳痰[例(构成比,%)]			χ²=3.240	0.072
否	91(88.35)	208(80.31)
是	12(11.65)	51(19.69)
发热[例(构成比,%)]			χ²=0.723	0.395
否	98(95.15)	240(92.66)
是	5(4.85)	19(7.34)

CT征象	非肺结核患者(103例)	肺结核患者(259例)	χ²值	P值
胸腔积液			0.023	0.881
否	85(82.52)	212(81.85)
是	18(17.48)	47(18.15)
钙化			10.062	0.002
否	49(47.57)	170(65.64)
是	54(52.43)	89(34.36)
空洞			0.001	0.971
否	92(89.32)	231(89.19)
是	11(10.68)	28(10.81)
结节			0.681	0.409
否	22(21.36)	66(25.48)
是	81(78.64)	193(74.52)
胸膜角钝化			0.800	0.371
否	103(100.00)	257(99.23)
是	0(0.00)	2(0.77)
心脏增大			0.024	0.878
否	102(99.03)	256(98.84)
是	1(0.97)	3(1.16)
磨玻璃样改变			0.361	0.548
否	81(78.64)	196(75.68)
是	22(21.36)	63(24.32)
胸膜增厚			1.356	0.244
否	77(74.76)	208(80.31)
是	26(25.24)	51(19.69)
纤维化			7.700	0.006
否	62(60.19)	194(74.90)
是	41(39.81)	65(25.10)
斑片影			0.062	0.804
否	69(66.99)	177(68.34)
是	34(33.01)	82(31.66)

因素	β值	s${}_{\overline{x}}$值	Wald χ²值	aOR(95%CI)值^a	P值
年龄	0.033	0.252	0.017	1.03(0.63~1.70)	0.895
女性	0.005	0.008	0.471	1.01(0.99~1.02)	0.493
咳痰	0.636	0.353	3.220	1.89(0.94~3.77)	0.073
钙化	-0.654	0.247	7.028	0.52(0.32~0.84)	0.008
纤维化	-0.658	0.258	6.488	0.52(0.31~0.86)	0.011

人工智能辅助CT阅片系统在综合医院肺结核早发现中的应用

Study on the application of AI-assisted CT reading system in early detection of pulmonary tuberculosis in general hospitals

RichHTML

PDF (PC)

可视化

摘要/Abstract

引用本文

使用本文

图/表 6

参考文献 20

相关文章 15

编辑推荐

Metrics

本文评价

[1]	刘立阳. 医院消毒供应室护理质量控制对院内呼吸道感染发生率的影响研究[J]. 中国防痨杂志, 2025, 47(S2): 208-211.
[2]	宋清, 李丽伟. 6S分诊管理模式在医院门诊护理服务中的应用效果分析[J]. 中国防痨杂志, 2025, 47(S1): 234-236.
[3]	中国防痨协会影像专业分会, 中华医学会结核病学分会, 中国防痨协会标准化专业分会, 中国防痨协会结核病控制专业分会. 胸部影像学检查人工智能辅助读片技术在肺结核患者发现中的应用专家共识[J]. 中国防痨杂志, 2025, 47(9): 1093-1104.
[4]	王琳, 屈妍. 多重耐药菌医院感染防控研究进展[J]. 中国防痨杂志, 2025, 47(9): 1196-1203.
[5]	阿布都热苏力·吐尔孙, 阿布都克尤木江·阿布力孜, 帕提曼·买买提, 黄陈翠, 沈玲燕, 马依迪丽·尼加提. 基于纵向胸部CT的影像组学深度学习预测肺结核患者疗效的研究[J]. 中国防痨杂志, 2025, 47(8): 1044-1052.
[6]	焦家欢, 孙长峰, 吴刚, 黄富礼, 盛云建. 基于机器学习算法的诊断模型对结核性胸腔积液的应用价值[J]. 中国防痨杂志, 2025, 47(8): 1053-1061.
[7]	朱庆东, 赵春艳, 谢周华, 宋树林, 宋畅. 基于人工智能的CT影像组学在结核病诊断和治疗反应监测中应用的研究进展[J]. 中国防痨杂志, 2025, 47(8): 1068-1076.
[8]	李婷, 夏岚, 刘双, 王丹霞, 逯嘉, 於一帆, 程钎钎. 四川省强化省、市两级结核病定点医疗机构患者登记管理干预的短期效果分析[J]. 中国防痨杂志, 2025, 47(5): 647-652.
[9]	高丽, 陈盛玉, 张文倩, 李丹, 张国钦. 天津市公立医院四类高危人群结核病监测能力现况调查[J]. 中国防痨杂志, 2025, 47(2): 195-200.
[10]	李婷婷, 刘欢庆, 雷倩, 尤著宏, 赵国连. 基于机器学习算法的诊断模型对结核性胸膜炎的应用价值[J]. 中国防痨杂志, 2025, 47(11): 1508-1514.
[11]	蔡可欣. 大型公立医院“一院多区”模式下的档案管理创新策略[J]. 中国防痨杂志, 2024, 46(S2): 4-6.
[12]	姚耀. DRG医保付费模式下医院成本管理改革路径探究[J]. 中国防痨杂志, 2024, 46(S2): 7-9.
[13]	徐梓毓. 公立医院开展肺结核科普的现状及改善途径[J]. 中国防痨杂志, 2024, 46(S2): 536-538.
[14]	金沈洁. 基于行为转变理论护理在多重耐药菌医院感染防控中的效果观察[J]. 中国防痨杂志, 2024, 46(S1): 166-169.
[15]	梅春林, 杨澄清, 杜荣辉, 曹探赜, 冯伟, 陈淑芳, 刘修平, 欧佳莉. 综合医院GeneXpert MTB/RIF检测支气管肺泡灌洗液极低结核分枝杆菌诊断肺结核的准确性[J]. 中国防痨杂志, 2024, 46(9): 1037-1041.