Diffusion weighted imaging with different b-values for the classification diagnosis of pulmonary tuberculosis and prediction of multidrug-resistance risk

doi:10.19982/j.issn.1000-6621.20240201

Abstract

Abstract:

Objective: To investigate the diagnostic efficacy of diffusion weighted imaging (DWI) with different b-values in classifying tuberculosis (TB) and predicting multidrug-resistance (MDR) risk. Methods: This study enrolled 110 TB patients admitted to Heilongjiang Provincial Infectious Disease Prevention and Control Hospital between October 2020 and May 2023. All patients underwent DWI with various b-values (0, 500, 800, 1000 s/mm²) to compare the diagnostic accuracy for TB classification at different b-values. Based on drug resistance test results, the 110 patients were divided into an MDR-TB group (47 cases) and a drug-sensitive TB group (63 cases). Clinical data and imaging features of the two groups were compared, and logistic regression was used to explore the relationship between apparent diffusion coefficient (ADC) and MDR-TB. Results: At a b-value of 800 s/mm², the rates of missed diagnosis (1.82%, 2/110) and misdiagnosis (3.64%, 4/110) in diagnosing various classifications of TB were significantly lower than those at b-values of 0, 500, and 1000 s/mm² (the missed diagnosis rate and misdiagnosis rate were 11.82% (13/110) and 17.27% (19/110), 9.09% (10/110) and 13.64% (15/110), 8.18% (9/110) and 10.91% (12/110), respectively), the differences were statistically significant (missed diagnosis: χ²=8.657, P=0.003; χ²=5.641, P=0.018; χ²=4.689, P=0.030; misdiagnosis: χ²=10.925, P=0.000; χ²=6.970, P=0.008; χ²=4.314, P=0.038). The area under the curve, accuracy, and Youden index of the receiver operating characteristic curve with a b-value of 800 s/mm² were the highest (0.874, 96.35%, and 0.788, respectively), all higher than those with b-values of 0, 500, and 1000 s/mm² (0.718, 82.73%, 0.612, and 0.735, 86.36%, 0.695, and 0.862, 89.09%, 0.724, respectively). After adjusting confounding factors, ADC was independently correlated with MDR risk, showing a negative correlation (r=-0.719, P<0.05). Conclusion: DWI with a b-value of 800 s/mm² demonstrates superior efficacy in diagnosing different TB classifications, with higher sensitivity and accuracy. ADC was negatively correlated with MDR risk.

Key words: Diffusion magnetic resonance imaging, Tuberculosis, pulmonary, Drug resistance, multiple, Diagnosis

CLC Number:

Liu Xin, Yu Qianhui. Diffusion weighted imaging with different b-values for the classification diagnosis of pulmonary tuberculosis and prediction of multidrug-resistance risk[J]. Chinese Journal of Antituberculosis, 2024, 46(11): 1356-1364. doi: 10.19982/j.issn.1000-6621.20240201

Figures/Tables 6

临床资料	耐多药组(47例)	药物敏感组(63例)	统计检验值	P值
年龄(岁, $x ˉ$ ±s)	48.23±2.49	47.96±2.54	t=0.556	0.579
BMI( $x ˉ$ ±s)	21.87±3.52	23.85±3.46	t=2.947	0.004
平均收缩压(mmHg, $x ˉ$ ±s)	129.54±11.75	129.43±10.98	t=0.050	0.960
平均舒张压(mmHg, $x ˉ$ ±s)	84.72±8.94	85.02±9.24	t=0.171	0.865
心率(次/min, $x ˉ$ ±s)	105.86±9.85	106.14±10.23	t=0.144	0.886
性别[例(构成比,%)]			χ²=8.019	0.005
男性	40(85.11)	38(60.32)
女性	7(14.89)	25(39.68)
婚姻情况[例(构成比,%)]			χ²=0.014	0.905
已婚	37(78.72)	49(77.78)
未婚	2(4.26)	3(4.76)
离异/丧偶	8(17.02)	11(17.46)
居住方式[例(构成比,%)]			χ²=0.026	0.872
独居	8(17.02)	10(15.87)
与家人居住	39(82.98)	53(84.13)
受教育水平[例(构成比,%)]			χ²=0.052	0.820
本科及以上	9(19.15)	11(17.46)
高中/专科	24(51.06)	32(50.79)
初中及以下	14(29.79)	20(31.75)
家庭月收入[例(构成比,%)]			χ²=9.660	0.002
<2000元	20(42.55)	10(15.87)
2000~5000元	23(48.94)	39(61.91)
>5000元	4(8.51)	14(22.22)
吸烟史[例(构成比,%)]			χ²=0.001	0.978
是	23(48.94)	31(49.21)
否	24(51.06)	32(50.79)
饮酒史[例(构成比,%)]			χ²=0.001	0.980
是	26(55.32)	35(55.56)
否	21(44.68)	28(44.44)
高血压史[例(构成比,%)]			χ²=0.006	0.937
是	19(40.43)	25(39.68)
否	28(59.57)	38(60.32)
合并糖尿病[例(构成比,%)]			χ²=3.996	0.046
是	16(34.04)	11(17.46)
否	31(65.96)	52(82.54)
冠心病史[例(构成比,%)]			χ²=0.026	0.872
是	8(17.02)	10(15.87)
否	39(82.98)	53(84.13)
高血脂史[例(构成比,%)]			χ²=0.000	0.987
是	12(25.53)	16(25.40)
否	35(74.47)	47(74.60)
慢性咳嗽史[例(构成比,%)]			χ²=5.988	0.014
是	17(36.17)	10(15.87)
否	30(63.83)	53(84.13)
不规则治疗[例(构成比,%)]			χ²=12.644	<0.001
是	20(42.55)	8(12.70)
否	27(57.45)	55(87.30)
肺结核患者密切接触史[例(构成比,%)]			χ²=0.016	0.900
是	2(4.26)	3(4.76)
否	45(95.74)	60(95.24)
肺结核分型[例(构成比,%)]			χ²=0.405	0.525
Ⅰ型	1(2.13)	3(4.76)
Ⅱ型	5(10.64)	7(11.11)
Ⅲ型	29(61.70)	39(61.90)
Ⅳ型	4(8.51)	6(9.52)
混合型	8(17.02)	8(12.71)
累及肺叶数[例(构成比,%)]			χ²=0.000	0.987
1个	12(25.53)	16(25.40)
2个	10(21.28)	13(20.63)
3个	9(19.15)	12(19.05)
4个	6(12.77)	8(12.70)
5个	10(21.27)	14(22.22)
磨玻璃影[例(构成比,%)]	7(14.89)	9(14.29)	χ²=0.008	0.929
实变[例(构成比,%)]	23(48.94)	31(49.21)	χ²=0.001	0.978
树芽征[例(构成比,%)]	39(82.98)	37(58.73)	χ²=7.412	0.006
空洞[例(构成比,%)]			χ²=7.021	0.008
0个	20(42.55)	32(50.79)
1个	10(21.28)	14(22.22)
2个	5(10.64)	7(11.11)
3个	4(8.51)	6(9.52)
4个	3(6.38)	4(6.36)
≥5个	5(10.64)	0(0.00)
结核球[例(构成比,%)]	10(21.28)	13(20.63)	χ²=0.007	0.935
纤维化[例(构成比,%)]	35(74.47)	47(74.60)	χ²=0.000	0.987
钙化[例(构成比,%)]	16(34.04)	21(33.33)	χ²=0.006	0.938
肺毁损[例(构成比,%)]	11(23.40)	14(22.22)	χ²=0.021	0.884
支气管扩张[例(构成比,%)]	18(38.30)	23(36.51)	χ²=0.037	0.848
肺气肿[例(构成比,%)]	12(25.53)	16(25.40)	χ²=0.000	0.987
淋巴结肿大[例(构成比,%)]	15(31.91)	20(31.75)	χ²=0.000	0.985
淋巴结钙化[例(构成比,%)]	7(14.89)	9(14.29)	χ²=0.008	0.929
胸膜增厚粘连[例(构成比,%)]	40(85.11)	54(85.71)	χ²=0.008	0.929
胸腔积液[例(构成比,%)]	10(21.28)	12(19.05)	χ²=0.084	0.772
壁形态不均匀[例(构成比,%)]	35(74.47)	46(73.02)	χ²=0.029	0.864
最大层面长径(mm, $x ˉ$ ±s)	30.29±4.23	26.78±5.34	t=1.631	0.106
最大层面短径(mm, $x ˉ$ ±s)	20.47±2.76	21.77±4.98	t=1.613	0.110
最大层面长短径比值( $x ˉ$ ±s)	1.50±0.35	1.28±0.22	t=4.037	<0.001
内壁长径(mm, $x ˉ$ ±s)	20.64±4.32	19.93±3.87	t=0.906	0.367
内壁短径(mm, $x ˉ$ ±s)	13.45±2.96	13.21±2.65	t=0.447	0.656
内壁长短径比值( $x ˉ$ ±s)	1.53±0.43	1.49±0.39	t=0.509	0.612
MRI信号强度
T₁WI	0.98±0.22	1.00±0.24	t=0.448	0.655
T₂WI	1.42±0.36	1.39±0.31	t=0.469	0.640
ADC(×10^-3mm²/s, $x ˉ$ ±s)	0.73±0.12	1.17±0.25	t=11.137	<0.001

References 19

[1]	Kumbar AS, Muhammed Faseed CH, Harsha DS. Pattern of chest radiographic abnormalities and co-morbidities in newly detected sputum positive pulmonary tuberculosis cases. IP Indian J Immunol Resp Med, 2021, 6(1): 10-17. doi:10.18231/J.IJIRM.2021.003.
[2]	Manjula CM, Majra JP. Adverse drug reactions among new sputum positive pulmonary tuberculosis patients on daily treatment regimen: a prospective study from Haryana, India. Int J Community Med Public Health, 2021, 8(2): 775-780. doi:10.18203/2394-6040.ijcmph20210237.
[3]	舒薇, 刘宇红. 世界卫生组织《2023年全球结核病报告》解读. 结核与肺部疾病杂志, 2024, 5(1): 15-19. doi:10.19983/j.issn.2096-8493.2024006.
[4]	Yang F, Yu H, Kantipudi K, et al. Differentiating between drug-sensitive and drug-resistant tuberculosis with machine learning for clinical and radiological features. Quant Imaging Med Surg, 2022, 12(1): 675-687. doi:10.21037/qims-21-290.
[5]	Zeng Y, Zhai XL, Wáng YXJ, et al. Illustration of a number of atypical computed tomography manifestations of active pulmonary tuberculosis. Quant Imaging Med Surg, 2021, 11(4): 1651-1667. doi:10.21037/qims-20-1323.
[6]	Li CH, Fan X, Lv SX, et al. Clinical and Computed Tomography Features Associated with Multidrug-Resistant Pulmonary Tuberculosis: A Retrospective Study in China. Infect Drug Resist, 2023, 16: 651-659. doi:10.2147/IDR.S394071.
[7]	Song QS, Zheng CJ, Wang KP, et al. Differences in pulmonary nodular consolidation and pulmonary cavity among drug-sensitive, rifampicin-resistant and multi-drug resistant tuberculosis patients: a computerized tomography study with history length matched cases. J Thorac Dis, 2022, 14(7): 2522-2531. doi:10.21037/jtd-22-145.
[8]	中国研究型医院学会感染与炎症放射专业委员会, 中华医学会放射学分会传染病学组, 中国医师协会放射医师分会感染影像专业委员会, 等. 非结核分枝杆菌肺病影像诊断专家共识. 中国研究型医院, 2021, 8(3): 1-6. doi:10.19450/j.cnki.jcrh.2021.03.001.
[9]	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. 2017-11-09.
[10]	贾红彦, 董静, 张宗德, 等. 结核分枝杆菌感染的免疫学检测技术研究进展及临床应用现状. 中国防痨杂志, 2022, 44(7): 720-726. doi:10.19982/j.issn.1000-6621.20220103.
[11]	Hailemichael Y, Johnson DH, Abdel-Wahab N, et al. Interleukin-6 blockade abrogates immunotherapy toxicity and promotes tumor immunity. Cancer Cell, 2022, 40(5): 509-523. doi:10.1016/j.ccell.2022.04.004. pmid: 35537412
[12]	汤慧中, 栗俊, 马隆佰, 等. MDCT诊断薄壁空洞性肺结核和薄壁囊腔型肺癌的临床价值探析. 中国CT和MRI杂志, 2020, 18(7): 74-77. doi:10.3969/j.issn.1672-5131.2020.07.024.
[13]	Blanc FX, Badje AD, Bonnet M, et al. Systematic or Test-Guided Treatment for Tuberculosis in HIV-Infected Adults. N Engl J Med, 2020, 382(25): 2397-2410. doi:10.1056/NEJMoa1910708.
[14]	段慧, 韩丹, 康绍磊, 等. 基于AI数据分析对不典型肺结核误诊为周围型肺癌的临床研究. 临床放射学杂志, 2021, 40(7): 1312-1315.
[15]	Feng Y, Wang J, Shao Z, et al. Predicting related factors of immunological response to hepatitis B vaccine in hemodialysis patients based on integration of decision tree classification and logistic regression. Hum Vaccin Immunother, 2021, 17(9): 3214-3220. doi:10.1080/21645515.2021.1895603.
[16]	王钰, 付际游, 张少言, 等. 耐多药肺结核患者中医证素与临床特征的初步研究. 中国防痨杂志, 2020, 42(2): 115-120. doi:10.3969/j.issn.1000-6621.2020.02.007.
[17]	孙金昊, 陶磊, 贺向红, 等. GeneXpert MTB/RIF系统、MGIT-960快速培养系统及药敏检测肺结核患者痰标本、支气管肺泡灌洗液标本阳性率比较及耐药分析. 临床军医杂志, 2022, 50(12): 1275-1277, 1281. doi:10.16680/j.1671-3826.2022.12.19.
[18]	龚文平, 米洁, 吴雪琼. 免疫活性物质: 结核病和非结核分枝杆菌病治疗的新选择. 中国防痨杂志, 2022, 44(11): 1107-1121. doi:10.19982/j.issn.1000-6621.20220239.
[19]	韩婷婷, 刘桂珍, 陈秋奇, 等. 世界卫生组织《应对结核病及其共病合作行动框架》解读. 中国防痨杂志, 2023, 45(1): 25-30. doi:10.19982/j.issn.1000-6621.20220464.

类别	确诊	0s/mm²	500s/mm²	800s/mm²	1000s/mm²
Ⅰ型肺结核	4(3.64)	1(0.91)	2(1.82)	3(2.73)	3(2.73)
Ⅱ型肺结核	12(10.91)	7(6.36)	9(8.18)	12(10.91)	10(9.09)
Ⅲ型肺结核	68(61.82)	75(68.18)	72(65.45)	70(63.64)	72(65.45)
Ⅳ型肺结核	10(9.09)	16(14.55)	13(11.82)	10(9.09)	11(10.00)
混合型肺结核	16(14.55)	5(4.55)	6(5.45)	13(11.82)	11(10.00)
误诊	-	13(11.82)	10(9.09)	2(1.82)	9(8.18)
漏诊	-	19(17.27)	15(13.64)	4(3.64)	12(10.91)

肺结核分型	0s/mm²	500s/mm²	800s/mm²	1000s/mm²	800s/mm² vs. 0s/mm²		800s/mm² vs. 500s/mm²		1000s/mm² vs. 0s/mm²		1000s/mm² vs. 500s/mm²
肺结核分型	0s/mm²	500s/mm²	800s/mm²	1000s/mm²	t值	P值	t值	P值	t值	P值	t值	P值
Ⅰ型肺结核(4例)	1.25±0.26	1.20±0.20	1.12±0.15	1.13±0.14	4.512	<0.001	3.261	0.004	4.435	<0.001	3.198	0.005
Ⅱ型肺结核(12例)	1.02±0.22	1.04±0.15	0.91±0.12	0.90±0.12	6.198	<0.001	6.034	<0.001	6.043	<0.001	5.892	<0.001
Ⅲ型肺结核(68例)	1.23±0.31	1.19±0.24	1.15±0.13	1.13±0.10	3.123	0.042	3.012	0.044	3.245	0.041	3.029	0.043
Ⅳ型肺结核(10例)	1.25±0.28	1.22±0.25	1.18±0.12	1.15±0.09	3.123	0.043	3.226	0.044	3.498	0.039	3.502	0.031
混合型肺结核(16例)	0.93±0.25	0.82±0.16	0.70±0.11	0.69±0.12	6.891	<0.001	5.019	<0.001	7.023	<0.001	5.223	<0.001
Ⅰ型vs. Ⅱ型
t值	2.343	1.710	3.818	3.755
P值	0.033	0.108	0.002	0.002
Ⅰ型vs. Ⅲ型
t值	0.126	0.082	0.594	0.762
P值	0.900	0.935	0.554	0.449
Ⅰ型vs. Ⅳ型
t值	0.000	0.142	0.132	0.323
P值	1.000	0.889	0.897	0.752
Ⅰ型vs.混合型
t值	2.274	4.062	7.453	6.949
P值	0.035	0.001	<0.001	<0.001
Ⅱ型vs. Ⅲ型
t值	3.098	2.088	5.959	7.128
P值	0.003	0.040	<0.001	<0.001
Ⅱ型vs. Ⅳ型
t值	2.910	2.089	5.255	5.429
P值	0.008	0.049	<0.001	<0.001
Ⅱ型vs.混合型
t值	0.110	3.697	6.392	4.583
P值	0.913	0.001	<0.001	<0.001
Ⅲ型vs. Ⅳ型
t值	0.193	0.367	0.687	0.597
P值	0.848	0.714	0.494	0.552
Ⅲ型vs.混合型
t值	3.600	5.854	12.795	15.234
P值	0.001	<0.001	<0.001	<0.001
Ⅳ型vs.混合型
t值	3.034	4.997	10.459	10.401
P值	0.006	<0.001	<0.001	<0.001

b值(s/mm²)	AUC	95%CI值	敏感度(%)	特异度(%)	准确率(%)	约登指数
0	0.718	0.703~0.736	81.30	87.50	82.73	0.612
500	0.735	0.714~0.758	77.42	95.00	86.36	0.695
800	0.874	0.841~0.907	91.70	89.50	96.35	0.788
1000	0.862	0.832~0.897	89.25	89.70	89.09	0.724

变量	未校准模型	模型1	模型2	模型3	模型4	模型5
ADC(×10^-3mm²/s,95%CI)	0.57 (0.32~0.76)	0.55 (0.33~0.75)	0.54 (0.34~0.78)	0.53 (0.33~0.77)	0.53 (0.35~0.79)	0.52 (0.35~0.74)
ADC二分类(×10^-3mm²/s,95%CI)^a
≤1.00	1	1	1	1	1	1
>1.00	0.55 (0.31~0.74)	0.47 (0.26~0.68)	0.46 (0.28~0.69)	0.46 (0.25~0.60)	0.45 (0.27~0.59)	0.45 (0.26~0.60)

影响因素	弥散加权成像
影响因素	β(95%CI)值	P值
体质量指数	-0.637(-0.865~1.345)	0.095
男性	-0.829(-0.974~2.013)	0.147
家庭月收入	-0.698(-0.927~1.184)	0.086
合并糖尿病	-0.982(-1.089~2.018)	0.123
慢性咳嗽史	-0.087(-0.142~0.546)	0.215
不规则治疗	-0.134(-0.902~0.361)	0.139
树芽征	-0.645(-0.985~-0.260)	0.027
空洞个数	-0.523(-0.851~-0.317)	0.035
最大层面长短径比值	-0.882(-0.903~-0.029)	0.008