切换至 "中华医学电子期刊资源库"
高级检索
中华消化病与影像杂志(电子版)

中华消化病与影像杂志(电子版) ›› 2024, Vol. 14 ›› Issue (02) : 114 -120. doi: 10.3877/cma.j.issn.2095-2015.2024.02.003

论著

基于列线图模型对慢性乙型肝炎合并肝脏脂肪变性患者并发晚期肝纤维化的临床预测
江浩1, 余宏圣1, 杨碧兰1, 阿布都克尤木·斯马依1, 吴斌1, 杨逸冬1,†()   
  1. 1. 510630　广州，中山大学附属第三医院消化内科；510630　广州，广东省肝病重点实验室
  • 收稿日期:2023-09-08 出版日期:2024-04-01
  • 通信作者: 杨逸冬
  • 基金资助:
    广东省自然科学基金杰出青年项目(2022B1515020024)

Nomogram model-based clinical prediction of advanced hepatic fibrosis in patients with chronic hepatitis B complicated by hepatic steatosis

Hao Jiang1, Hongsheng Yu1, Bilan Yang1, Smayi Abdukyamu1, Bin Wu1, Yidong Yang1,†()   

  1. 1. Department of Gastroenterology, the Third Affiliated Hospital of Sun Yat-Sen University, Guangzhou 510630, China; Guangdong Provincial Key Laboratory of Liver Disease Research, Guangzhou 510630, China
  • Received:2023-09-08 Published:2024-04-01
  • Corresponding author: Yidong Yang
全文 ( ) 下载PDF ( ) 导出引用
引用本文：

江浩, 余宏圣, 杨碧兰, 阿布都克尤木·斯马依, 吴斌, 杨逸冬. 基于列线图模型对慢性乙型肝炎合并肝脏脂肪变性患者并发晚期肝纤维化的临床预测[J]. 中华消化病与影像杂志(电子版), 2024, 14(02): 114-120.

Hao Jiang, Hongsheng Yu, Bilan Yang, Smayi Abdukyamu, Bin Wu, Yidong Yang. Nomogram model-based clinical prediction of advanced hepatic fibrosis in patients with chronic hepatitis B complicated by hepatic steatosis[J]. Chinese Journal of Digestion and Medical Imageology(Electronic Edition), 2024, 14(02): 114-120.

目的

探讨慢性乙型肝炎（CHB）合并肝脏脂肪变性患者并发晚期肝纤维化（S3～S4）的独立危险因素，构建及验证列线图风险预测模型。

方法

选取2008年8月至2020年12月在中山大学附属第三医院就诊并行经皮肝脏穿刺活检的439例未经抗病毒治疗的CHB合并肝脏脂肪变性患者作为研究对象，按2∶1随机分为训练组293例和验证组146例。采用logistic回归筛选出晚期纤维化的相关危险因素并构建列线图风险预测模型。

结果

logistic回归分析显示血小板（OR=0.987，P=0.003）、凝血酶原时间（OR=1.952，P=0.011）、球蛋白（OR=1.260，P=0.001）是晚期纤维化的独立危险因素。列线图风险预测模型在训练组中预测晚期纤维化的曲线下面积为0.866，显著优于天冬氨酸转氨酶和血小板比率指数（0.782，P=0.017）、γ-谷氨酰转肽酶/血小板比值（0.753，P=0.004）、肝纤维化4因子指数（0.780，P=0.024）、非酒精性脂肪肝纤维化评分（0.737，P<0.001）、BARD评分（0.595，P<0.005）模型；在验证组中也得到了类似的结果（P<0.05）。校准曲线和决策曲线显示列线图模型具有较高的一致性和临床净收益。

结论

本研究基于CHB合并肝脏脂肪变性的晚期肝纤维化患者的独立危险因素构建了列线图风险预测模型，经验证，该模型具有较高的预测效能、一致性和临床净收益。

关键词: 慢性乙型肝炎, 非酒精性脂肪性肝病, 晚期肝纤维化, 列线图
Objective

To identify the independent risk factors associated with advanced fibrosis (S3-S4) in patients suffering from chronic hepatitis B (CHB) complicated with hepatic steatosis, and to develop and validate a nomogram risk prediction model.

Methods

A total of 439 treatment-naïve CHB patients who had hepatic steatosis and underwent liver biopsy in the Third Affiliated Hospital of Sun Yat-sen University between August 2008 and December 2020 were recruited as research objects. These patients were then randomly allocated at a 2∶1 ratio into the training set (n=293) and the validation set (n=146). Logistic regression was used to identify the risk factors of advanced fibrosis. A nomogram prediction model was subsequently created.

Results

Logistic regression analysis revealed that platelet (OR=0.987, P=0.003), prothrombin time (OR=1.952, P=0.011), and globulin (OR=1.260, P=0.001) were the independent risk factors for advanced fibrosis. The area under the receiver operating characteristic (ROC) curve for the proposed nomogram model in the training group, which predicted advanced fibrosis, was noted to be 0.866. This was considerably higher than the aspartate aminotransferase-to-platelet ratio index (0.782, P=0.017), gamma-glutamyl transpeptidase-to-platelet ratio (0.753, P=0.004), fibrosis-4 score (0.780, P=0.024), non-alcoholic fatty liver disease fibrosis score (0.737, P<0.001), and BARD score (0.595, P<0.001) models; similar findings were observed in the validation set (P<0.05). Calibration curve and decision curve demonstrated that the nomogram model had high consistency and clinical net benefit.

Conclusion

A nomogram risk prediction model based on independent risk factors of advanced liver fibrosis patients with CHB combined with hepatic steatosis is constructed in this study. After verification, the model has high predictive efficiency, consistency, and clinical net gain.

Key words: Chronic hepatitis B, Nonalcoholic fatty liver disease, Advanced cirrhosis, Nomogram
表1 训练组与验证组患者基线资料比较
基线资料 训练组(n=293) 验证组(n=146) 统计量值 P值
年龄[岁,M(Q1,Q3)] 36.00(31.00,42.00) 36.00(30.75,43.00) Z=0.272 0.786
男性[例(%)] 255(87.0) 129(88.4) c2=0.156 0.693
体重指数(kg/m2,±s) 24.88±3.50 24.35±3.20 t=-1.522 0.129
血小板(×10-9/L,±s) 210.40±54.71 220.9±52.59 t=-1.370 0.171
天门冬氨酸氨基转移酶[U/L,M(Q1,Q3)] 38.00(28.00,58.00) 42.00(31.00,65.25) Z=-1.646 0.100
丙氨酸氨基转移酶[U/L,M(Q1,Q3)] 29.00(24.00,37.00) 31.00(24.00,41.50) Z=-1.371 0.171
γ-谷氨酰转移酶[U/L,M(Q1,Q3)] 31.00(22.00,45.00) 34.00(24.00,50.25) Z=-1.360 0.174
白蛋白[g/L,M(Q1,Q3)] 45.80(43.80,47.80) 45.55(43.28,47.30) Z=1.502 0.133
球蛋白(g/L,±s) 26.85±4.20 27.79±4.45 t=1.927 0.052
甲胎蛋白[μg/mL,M(Q1,Q3)] 3.02(2.59,4.59) 3.38(2.10,5.25) Z=-1.317 0.188
总胆固醇(mmol/L,±s) 4.93±0.91 4.98±0.89 t=0.525 0.600
甘油三酯[mmol/L,M(Q1,Q3)] 1.28(0.94,1.92) 1.26(0.94,1.93) Z=0.259 0.796
高密度脂蛋白胆固醇[mmol/L,M(Q1,Q3)] 1.13(0.97,1.28) 1.12(0.98,1.30) Z=-0.395 0.693
低密度脂蛋白胆固醇(mmol/L,±s) 3.24±0.81 3.28±0.81 t=0.510 0.610
凝血酶原时间[s,M(Q1,Q3)] 13.20(12.70,13.70) 13.20(12.80,13.83) Z=-1.106 0.629
2型糖尿病[例(%)] 14(4.8) 9(6.2) c2=0.377 0.539
糖耐量受损[例(%)] 12(4.1) 5(3.4) c2=0.118 0.731
高血压[例(%)] 41(14.0) 23(15.8) c2=0.242 0.622
乙型肝炎e抗原阳性[例(%)] 126(43.0) 61(41.8) c2=0.060 0.807
乙型肝炎DNA水平[Log10 IU/mL,M(Q1,Q3)] 4.79(3.43,6.68) 5.78(3.53,7.40) Z=-1.648 0.099
脂肪变性分级[例(%)]     c2=2.812 0.421
0 89(30.4) 49(33.6)    
1 177(60.4) 90(61.6)    
2 20(6.8) 5(3.4)    
3 7(2.4) 2(1.4)    
炎症分级[例(%)]     c2=7.910 0.095
0~1 135(46.0) 64(43.9)    
2 113(38.6) 47(32.2)    
3 34(11.6) 21(14.4)    
4 11(3.8) 14(9.6)    
纤维化分期[例(%)]     c2=9.016 0.061
0 38(13.0) 20(13.7)    
1 130(44.7) 44(30.0)    
2 71(24.2) 39(26.7)    
3 34(11.6) 27(18.5)    
4 20(6.8) 16(11.0)    
表2 训练组中与晚期纤维化相关的临床参数的单因素Logistic回归分析
变量 OR 95%置信区间 P值
年龄 1.010 1.045~1.083 0.013
男性      
否 reference    
是 1.001 0.415~2.410 0.999
体重指数 0.902 0.984~1.072 0.709
血小板 0.974 0.980~0.987 <0.001
丙氨酸氨基转移酶 0.995 1.001~1.007 0.065
天门冬氨酸氨基转移酶 1.017 1.004~1.030 0.009
γ-谷氨酰转移酶 1.010 1.002~1.017 0.012
白蛋白 0.906 0.836~0.983 0.018
球蛋白 1.306 1.195~1.428 <0.001
甲胎蛋白 1.060 1.017~1.104 0.006
总胆固醇 1.126 0.801~1.582 0.495
甘油三酯 0.811 0.571~1.151 0.241
高密度脂蛋白胆固醇 3.200 0.929~11.024 0.065
低密度脂蛋白胆固醇 1.091 0.743~1.603 0.656
凝血酶原时间(s) 2.441 1.669~3.570 <0.001
2型糖尿病      
否 reference    
是 1.832 0.532~6.078 0.322
糖耐量受损      
否 reference    
是 1.503 0.393~5.749 0.551
高血压      
否 reference    
是 2.069 0.977~4.380 0.057
乙型肝炎e抗原阳性      
否 reference    
是 0.812 0.444~1.485 0.499
乙型肝炎DNA水平 0.864 0.743~1.005 0.057
表3 训练组中与晚期纤维化相关的临床参数的多因素Logistic回归分析
变量 回归系数 标准误 Wald值 OR(95%置信区间) P值
年龄 0.023 0.024 0.903 1.023(0.976~1.072) 0.342
血小板 -0.013 0.004 9.117 0.987(0.978~0.995) 0.003
天门冬氨酸氨基转移酶 0.002 0.010 0.025 1.002(0.982~1.022) 0.874
γ-谷氨酰转移酶 0.008 0.007 1.463 1.008(0.995~1.022) 0.226
白蛋白 -0.002 0.057 0.002 0.998(0.893~1.115) 0.968
球蛋白 0.231 0.054 18.615 1.260(1.135~1.400) 0.001
凝血酶原时间 0.669 0.261 6.548 1.952(1.169~3.258) 0.011
甲胎蛋白 0.034 0.024 2.031 1.035(0.987~1.085) 0.154
图1 预测慢性乙型肝炎合并肝脏脂肪变性患者并发晚期肝纤维化风险的列线图预测模型注：Points为单项得分；PLT为血小板；PT为凝血酶原时间；GLB为球蛋白；Total Points为总得分，以总得分向下划一条线所得的值即为诊断为晚期肝纤维化的概率（S3～S4）。
图2 各预测模型在训练组和验证组的受试者操作特征曲线注：APRI为天冬氨酸转氨酶和血小板比率指数；GPR为γ-谷氨酰转肽酶/血小板比值；FIB-4为肝纤维化4因子指数；NFS为非酒精性脂肪肝纤维化评分；BARD为bard评分；Nomogram为列线图模型；Sensitivity为灵敏度；Specificity为特异度；（2A）训练组，（2B）验证组。
表4 不同预测模型预测晚期纤维化的受试者操作特征曲线分析
组别 敏感性(%) 特异性(%) AUC(95% CI) P值 P value of ROC delong test
训练组          
APRI 88.89 59.00 0.782(0.731~0.828) <0.001 0.017
GPR 74.07 70.04 0.753(0.699~0.801) <0.001 0.004
FIB-4 68.52 81.17 0.780(0.728~0.826) <0.001 0.024
NFS 59.26 81.59 0.737(0.683~0.786) <0.001 <0.001
BARD 53.70 64.02 0.595(0.536~0.651) 0.020 <0.001
Nomogram 79.63 84.81 0.866(0.821~0.903) <0.001 -
验证组          
APRI 79.07 73.79 0.766(0.689~0.832) <0.001 <0.001
GPR 67.44 75.73 0.755(0.677~0.822) <0.001 0.042
FIB-4 51.16 91.26 0.754(0.676~0.822) <0.001 0.014
NFS 55.81 86.41 0.732(0.653~0.802) <0.001 0.011
BARD 48.84 64.08 0.561(0.476~0.642) 0.236 <0.001
Nomogram 74.42 86.41 0.836(0.765~0.892) <0.001 -
图3 预测模型的校准曲线图注：Actually probability为实际风险概率；Predicted Pr为预测风险概率；Apparent为参考线；Bias-corrected为矫正后曲线；Ideal为理想曲线；（3A）训练组，（3B）验证组。
图4 决策曲线分析图注：Nomogram为列线图模型；APRI为天冬氨酸转氨酶和血小板比率指数；GPR为γ-谷氨酰转肽酶/血小板比值；FIB-4为肝纤维化4因子指数；NFS为非酒精性脂肪肝纤维化评分；BARD为bard评分；ALL为均发生晚期纤维化；None为均不发生晚期纤维化；Standardized Net Benefit为标准净收益；High Risk Threshold为风险阈值；Cost：Benefit Ratio为成本效益比；（4A）训练组，（4B）验证组。
表5 不同预测模型预测晚期纤维化的Hosmer-Lemeshow检验分析
  训练组 验证组
统计量值(2) P值 统计量值(2) P值
APRI 32.302 <0.001 19.121 0.014
GPR 18.699 0.017 8.991 0.343
FIB-4 5.480 0.705 5.336 0.721
NFS 5.929 0.655 16.503 0.036
BARD 0.349 0.840 6.203 0.045
Nomogram 5.473 0.706 4.044 0.853
[1]
Shi YW, Yang RX, Fan JG. Chronic hepatitis B infection with concomitant hepatic steatosis: Current evidence and opinion[J]. World J Gastroenterol, 2021, 27(26): 3971-3983.
[2]
Kleiner DE, Brunt EM, Van Natta M, et al. Nonalcoholic Steatohepatitis Clinical Research Network. Design and validation of a histological scoring system for nonalcoholic fatty liver disease[J]. Hepatology, 2005, 41(6): 1313-1321.
[3]
Wang MM, Wang GS, Shen F, et al. Hepatic steatosis is highly prevalent in hepatitis B patients and negatively associated with virological factors[J]. Dig Dis Sci, 2014, 59(10): 2571-2579.
[4]
Peleg N, Issachar A, Sneh Arbib O, et al. A. Liver steatosis is a strong predictor of mortality and cancer in chronic hepatitis B regardless of viral load[J]. JHEP Rep, 2019, 1(1): 9-16.
[5]
Scheuer PJ. Classification of chronic viral hepatitis: a need for reassessment[J]. J Hepatol, 1991, 13(3): 372-374.
[6]
Hanif H, Khan MM, Ali MJ, et al. A New Endemic of Concomitant Nonalcoholic Fatty Liver Disease and Chronic Hepatitis B[J]. Microorganisms, 2020, 8(10): 1526.
[7]
Li F, Ou Q, Lai Z, et al. The Co-occurrence of Chronic Hepatitis B and Fibrosis Is Associated With a Decrease in Hepatic Global DNA Methylation Levels in Patients With Non-alcoholic Fatty Liver Disease[J]. Front Genet, 2021, 12: 671552.
[8]
Wai CT, Greenson JK, Fontana RJ, et al. A simple noninvasive index can predict both significant fibrosis and cirrhosis in patients with chronic hepatitis C[J]. Hepatology, 2003, 38(2): 518-526.
[9]
Thachil J. Relevance of clotting tests in liver disease[J]. Postgrad Med J, 2008, 84(990): 177-181.
[10]
Schmilovitz-Weiss H, Tovar A, et al. Predictive value of serum globulin levels for the extent of hepatic fibrosis in patients with chronic hepatitis B infection[J]. J Viral Hepat, 2006, 13(10): 671-677.
[11]
Buechler C, Wanninger J, Neumeier M. Adiponectin, a key adipokine in obesity related liver diseases[J]. World J Gastroenterol, 2011, 17(23): 2801-2811.
[12]
Feldstein AE, Canbay A, Angulo P, et al. Hepatocyte apoptosis and fas expression are prominent features of human nonalcoholic steatohepatitis[J]. Gastroenterology, 2003, 125(2): 437-443.
[13]
Zhang RN, Pan Q, Zhang Z, et al. Saturated Fatty Acid inhibits viral replication in chronic hepatitis B virus infection with nonalcoholic Fatty liver disease by toll-like receptor 4-mediated innate immune response[J]. Hepat Mon, 2015, 15(5): e27909.
[14]
European Association for Study of Liver; Asociacion Latinoamericana para el Estudio del Higado. EASL-ALEH Clinical Practice Guidelines: Non-invasive tests for evaluation of liver disease severity and prognosis[J]. J Hepatol, 2015, 63(1): 237-264.
[1] 赵亚楠, 方明, 徐绍岩, 魏树梅, 张慧, 黄奕宁, 刘亚静, 黄品同. 多声学技术参数联合血清学指标对非酒精性脂肪性肝病肝脂肪变性阶段的诊断价值[J]. 中华医学超声杂志（电子版）, 2023, 20(11): 1164-1173.
[2] 姚放鸣, 焦莹莹, 何敏聪, 曾子俊, 何晓铭, 刘良燕, 何伟, 魏秋实, 刘文刚. 膝骨关节炎患者的肌少症发病率及发病特点分析[J]. 中华关节外科杂志(电子版), 2024, 18(01): 30-38.
[3] 柯晴潆, 沈延飞, 许强宏, 蔡国龙. 预测脓毒性心肌病患者院内死亡风险列线图模型的构建及验证[J]. 中华危重症医学杂志(电子版), 2024, 17(01): 10-18.
[4] 王晓梅, 刘冰, 马丽琼, 卢祖静, 苗建军. 基于LASSO-Cox回归分析的非轻症急性胰腺炎死亡风险列线图预测模型的建立和临床应用效果分析[J]. 中华普通外科学文献(电子版), 2024, 18(01): 44-50.
[5] 唐旭, 韩冰, 刘威, 陈茹星. 结直肠癌根治术后隐匿性肝转移危险因素分析及预测模型构建[J]. 中华普外科手术学杂志(电子版), 2024, 18(01): 16-20.
[6] 甄子铂, 刘金虎. 基于列线图模型探究静脉全身麻醉腹腔镜胆囊切除术患者术后肠道功能紊乱的影响因素[J]. 中华普外科手术学杂志(电子版), 2024, 18(01): 61-65.
[7] 刘化胜, 郑龙波, 秦琛, 王伟芹, 郑学风, 马金龙, 马正, 王洪霞, 刘磊, 胡继霖. 基于影像学指标构建永久性乙状结肠造口旁疝发病风险预测模型[J]. 中华疝和腹壁外科杂志(电子版), 2024, 18(01): 75-82.
[8] 蔡小芳, 高慧, 葛军, 邢慧芸, 庄小燕, 李小丁. 多重耐药性肺结核治疗依从性预测分析[J]. 中华肺部疾病杂志(电子版), 2024, 17(01): 51-56.
[9] 陈晓毅, 尹雪霞, 刘静, 邬国松. 阻塞性睡眠呼吸暂停低通气综合征并发肺动脉高压的危险因素及预测分析[J]. 中华肺部疾病杂志(电子版), 2024, 17(01): 41-45.
[10] 孙振, 谭天华, 郑洋洋, 李喆, 宋京海. 基于术前纤维蛋白原与白蛋白比值构建肝癌微血管侵犯的预测模型[J]. 中华肝脏外科手术学电子杂志, 2024, 13(01): 27-32.
[11] 潘晓帆, 徐勤义, 陆瑨, 王丹, 刘路路, 董万利. 颅内动脉瘤破裂介入术后并发脑疝的风险因素分析[J]. 中华脑科疾病与康复杂志(电子版), 2024, 14(01): 37-44.
[12] 袁梦晨, 刘译阳, 赵帅, 陈林, 高宇, 肖晓燕, 尤亚茹, 梁何俊, 高剑波. 增强CT的列线图在鉴别EB病毒相关的胃淋巴上皮瘤样癌与胃腺癌中的应用[J]. 中华消化病与影像杂志(电子版), 2024, 14(02): 107-113.
[13] 薛文欣, 常婉英, 肖瑶, 张盼盼, 党晓智, 宋宏萍. 乳腺恶性非肿块型病变列线图预测模型的构建及外部验证[J]. 中华临床医师杂志(电子版), 2023, 17(10): 1045-1050.
[14] 王林源, 熊鑫, 杨坤, 邓勇志. 基于冠状动脉CT血管成像的影像组学列线图鉴别诊断易损斑块的价值[J]. 中华诊断学电子杂志, 2024, 12(01): 1-8.
[15] 李智毓, 李昌顶, 聂鑫, 倪琨涵, 韩泳涛, 冷雪峰. 食管鳞状细胞癌手术切缘阳性大样本回顾性研究及其临床预测模型的建立与验证[J]. 中华胸部外科电子杂志, 2024, 11(01): 31-39.
阅读次数
全文


摘要

版权所有 中华医学会 中华医学电子音像出版社有限责任公司  京ICP备14006079号-1  网络出版服务许可证：(署)网出证(京)字第075号