以血清学指标比值为变量的肝纤维化诊断模型构建与评价
|
摘要
目的构建以血清学指标比值为变量的肝纤维化诊断模型,并对该模型的诊断效能进行评估。方法选取398例慢性肝病患者,均行肝组织病理活检,检测与肝脏病变及纤维化相关的血清学指标15项[透明质酸(HA)、Ⅲ前胶原肽(PⅢNP)、层粘连蛋白(LN)、Ⅳ型胶原(CⅣ)、谷氨酰转肽酶(GGT)、丙氨酸氨基转移酶(ALT)、碱性磷酸酶(ALP)、天门冬氨酸氨基转移酶(AST)、白蛋白(ALB)、腺苷脱氨酶(ADA)、直接胆红素(DBIL)、总胆红素(TBI)、尿素(UREA)、载脂蛋白A1(Apo A1)、血小板计数(PLT)];随机选取其中338例,将这些指标分别与PLT、ALB做比值,经Spearman等级相关分析筛选与肝纤维化分级相关的指标,采用二元Logistic回归构建诊断模型,绘制ROC曲线找到有诊断价值的最佳截断点;另外60例用以评价模型诊断效能,参照病理活检分期结果,将以血清学指标比值为变量建立肝纤维化诊断模型的预测结果与其比较。结果 398例慢性肝病患者中,肝纤维化S0期81例,S1期87例,S2期126例,S3期50例,S4期54例。除CⅣ/PLT、Apo A1/PLT、CⅣ/ALB、UREA/ALB、Apo A1/ALB与肝纤维化分期相关性较小外,其他指标均与肝纤维化分期相关(r均>0.3,P均=0.000)。建立的诊断模型为:ln(p1/1-p1)=-4.603+21.838×HA/PLT-151.979×PⅢNP/PLT+117.009×ADA/PLT-5.507×HA/ALB+36.553×PⅢNP/ALB;ln(p2/1-p2)=-3.314-1.839×HA/ALB+10.345×ADA/ALB+0.645×GGT/ALB+9.450×HA/PLT;ln(p3/1-p3)=-3.524+27.842×ADA/PLT+3.282×PⅢNP/ALB+0.582×ALP/ALB-4.932×ADA/ALB;ln(p4/1-p4)=-4.170-5.979×AST/PLT+44.142×ADA/PLT+5.001×PⅢNP/ALB-6.926×ADA/ALB+1.174×AST/ALB。p1为S≥1预测概率,p2为S≥2的预测概率,p3为S≥3的预测概率,p4为S≥4的预测概率。利用p1、p2、p3、p4绘制ROC曲线,曲线下面积均>0.7。用以验证的60例样本中,与病理分期结果完全相同24例,参考病理学诊断可接受上下一个等级的诊断结论,增加可接受病例18例,合计42例(70%)。结论成功构建以血清学指标比值为变量的肝纤维化诊断模型,该模型对肝纤维化的诊断有较高临床价值。
Objective To establish a liver fibrosis diagnostic model with the ratio of serum markers as variables and to evaluate its diagnostic efficacy. Methods Totally 398 cases of patients with chronic liver disease undergoing liver biopsy were collected. We detected 15 serum markers related to liver lesions and fibrosis [hyaluronic acid( HA),procollagen Ⅲpeptide( PⅢ NP),laminin( LN),collage Ⅳ( C Ⅳ),gamma-glutamyl transferase( GGT),alanine aminotransferase( ALT),alkaline phosphatase( ALP),aspartate aminotransferase( AST),albumin( ALB),adenosine deaminase( ADA),direct bilirubin( DBIL),total bilirubin( TBIL),UREA,apolipoprotein A1( Apo A1),and blood platelet( PLT) ]. Meanwhile,338 cases were randomly selected,and then the radio of serum markers divided by PLT and ALB were calculated. Binary Logistic regression was used to construct the diagnosis model. After that,we drew ROC curve to find cut-off points which were most valuable in diagnosis. Sixty cases were collected to evaluate the diagnostic efficacy of the model,and the predicted staging results from the diagnosis model were compared with the results of liver biopsy. Results Among the 398 patients with chronic liver disease,81 cases were in S0 stage,87 in S1,126 in S2,50 in S3,and54 in S4. In addition to the radios of serum markers of CⅣ/PLT,APOA1/PLT,CⅣ/ALB,UREA/ALB,and APOA1/ALB,all the remaining radios were related to liver fibrosis stage( | r | > 0. 3,P = 0. 000). The diagnostic model was successfully established,followed by:(1)ln( p1/1-p1) =-4. 603 + 21. 838 × HA/PLT-151. 979 × PⅢNP/PLT + 117. 009 ×ADA/PLT-5. 507 × HA/ALB + 36. 553 × P ⅢNP/ALB;(2)ln( p2/1-p2) =-3. 314-1. 839 × HA/ALB + 10. 345 ×ADA/ALB + 0. 645 × GGT/ALB + 9. 450 × HA/PLT;(3)ln( p3/1-p3) =-3. 524 + 27. 842 × ADA/PLT + 3. 282 × PⅢNP/ALB + 0. 582 × ALP/ALB-4. 932 × ADA/ALB;(4)ln( p4/1-p4) =-4. 170-5. 979 × AST/PLT + 44. 142 × ADA/PLT + 5. 001 × PⅢNP/ALB-6. 926 × ADA/ALB + 1. 174 × AST/ALB. P1 was the predicted probability for S≥1,p2 for S≥2,p3 for S≥3,and p4 for S≥4. The area under the curve was > 0. 7 by using p1,p2,p3 and p4. In the 60 cases,24 cases were identical with the pathological staging results,referencing pathological diagnosis accepted one upper and one lower stages,18 cases were added,and the total acceptable cases were 42( 70%). Conclusion The diagnostic model of liver fibrosis is successfully established by taking the ratio of serum markers as indexes,which has high clinical value in the diagnosis of hepatic fibrosis.
Objective To establish a liver fibrosis diagnostic model with the ratio of serum markers as variables and to evaluate its diagnostic efficacy. Methods Totally 398 cases of patients with chronic liver disease undergoing liver biopsy were collected. We detected 15 serum markers related to liver lesions and fibrosis [hyaluronic acid( HA),procollagen Ⅲpeptide( PⅢ NP),laminin( LN),collage Ⅳ( C Ⅳ),gamma-glutamyl transferase( GGT),alanine aminotransferase( ALT),alkaline phosphatase( ALP),aspartate aminotransferase( AST),albumin( ALB),adenosine deaminase( ADA),direct bilirubin( DBIL),total bilirubin( TBIL),UREA,apolipoprotein A1( Apo A1),and blood platelet( PLT) ]. Meanwhile,338 cases were randomly selected,and then the radio of serum markers divided by PLT and ALB were calculated. Binary Logistic regression was used to construct the diagnosis model. After that,we drew ROC curve to find cut-off points which were most valuable in diagnosis. Sixty cases were collected to evaluate the diagnostic efficacy of the model,and the predicted staging results from the diagnosis model were compared with the results of liver biopsy. Results Among the 398 patients with chronic liver disease,81 cases were in S0 stage,87 in S1,126 in S2,50 in S3,and54 in S4. In addition to the radios of serum markers of CⅣ/PLT,APOA1/PLT,CⅣ/ALB,UREA/ALB,and APOA1/ALB,all the remaining radios were related to liver fibrosis stage( | r | > 0. 3,P = 0. 000). The diagnostic model was successfully established,followed by:(1)ln( p1/1-p1) =-4. 603 + 21. 838 × HA/PLT-151. 979 × PⅢNP/PLT + 117. 009 ×ADA/PLT-5. 507 × HA/ALB + 36. 553 × P ⅢNP/ALB;(2)ln( p2/1-p2) =-3. 314-1. 839 × HA/ALB + 10. 345 ×ADA/ALB + 0. 645 × GGT/ALB + 9. 450 × HA/PLT;(3)ln( p3/1-p3) =-3. 524 + 27. 842 × ADA/PLT + 3. 282 × PⅢNP/ALB + 0. 582 × ALP/ALB-4. 932 × ADA/ALB;(4)ln( p4/1-p4) =-4. 170-5. 979 × AST/PLT + 44. 142 × ADA/PLT + 5. 001 × PⅢNP/ALB-6. 926 × ADA/ALB + 1. 174 × AST/ALB. P1 was the predicted probability for S≥1,p2 for S≥2,p3 for S≥3,and p4 for S≥4. The area under the curve was > 0. 7 by using p1,p2,p3 and p4. In the 60 cases,24 cases were identical with the pathological staging results,referencing pathological diagnosis accepted one upper and one lower stages,18 cases were added,and the total acceptable cases were 42( 70%). Conclusion The diagnostic model of liver fibrosis is successfully established by taking the ratio of serum markers as indexes,which has high clinical value in the diagnosis of hepatic fibrosis.
引文
[1]Gaggini MC,Navarro RS,Stefanini AR,et al.Correlation between METAVIR scores and Raman spectroscopy in liver lesions induced by hepatitis C virus:a preliminary study[J].Lasers Med Sci,2015,30(4):1347-1355.
[2]Yoav L,Muriel W,Ruth C,et al.Non-invasive diagnosis of liv er fibrosis and cirrhosis[J].World J Gestroenterol,2015,21(41):11567-11583.
[3]肝脏硬度评估小组.瞬时弹性成像技术诊断肝纤维化专家意见[J].中华肝病杂志,2013,21(6):420-424.
[4]中华医学会传染病与寄生虫病学会,肝病学分会.病毒性肝炎防治方案[J].中华肝脏病杂志,2000,8(6):324-329.
[5]赵晓薇.ADA与ALT AST GGT联合检测在肝脏疾病诊断中的意义[J].内蒙古医学杂志,2014,46(11):1357-1358.
[6]Adams LA,Bulsara M,Rossi E,et al.Hepascore:an accurate validated predictor of liver fibrosis in chronic hepatitis C infection[J].Clin Chem,2005,51(10):1867-1873.
[7]Zeng MD,Lu LG,Mao YM,et al.Prediction of significant fibrosis in HBe Ag-positive patients with chronic hepatitis B by a noninvasive model[J].Hepatology,2005,42(6):1437-1445.
[8]Forns X,Ampurdanes S,Llovet JM,et al.Identification of chronic hepatitis C patients without hepatic fibrosis by a simple predictive model[J].Hepatology,2002,36:986-992.
[9]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.
[10]叶柱均,叶淑霞,张飞勇,等.ADA与ALT、AST、GGT联合检测在肝脏疾病诊断中的意义[J].国际医药卫生导报,2016,22(9):1300-1302.
[11]Friedman SL.Evolving challenges in hepatic fibrosis[J].Nat Rev Gastroenterol Hepatol,2010,7(8):425-436.
[12]Hou XY,Ellis MK,Mc Manus DP,et al.Diagnostic value of noninvasive bio-markers for stage-specific fibrosis in patients with advanced schist-osomiasis japonca[J].Int J Parasitol,2011,41(34):325-332.
[13]苑同业,刘国凤,张立营.血清Ⅲ型前胶原测定在肝纤维化诊断中的应用[J].中国疗养医学,2014,23(9):238-240.
[14]张伟龙,肖翔,潘理想,等.ADA与ALT、AST、GGT联合检测在肝脏疾病诊断的价值研究[J].中国实用医药,2017,12(1):53-55.
[15]Keef EB,Dieterich DT,Han SH,et al.A treatment algorithm for the management of chronic hepatitis B virus in ghe United States:an update[J].Clin Gastroenterol Hepatol,2006,4(8):936-962.
[16]许庆忠,雷霆雯,马小平.实验性肝纤维化血清白蛋白含量变化[J].贵阳医学院学报,2002,27(2):128-129.
[17]夏琦,符节海,蔡卫民,等.AST/PLT和GGT/PLT比值对肝纤维化的诊断价值[J].临床肝胆病杂志,2007,23(5):364-366.
[18]董菁菁,郭宏华.GGT与PLT比值指数对慢性乙型肝炎肝纤维化的诊断价值[J].临床肝胆病杂志,2017,33(3):548-551.
[2]Yoav L,Muriel W,Ruth C,et al.Non-invasive diagnosis of liv er fibrosis and cirrhosis[J].World J Gestroenterol,2015,21(41):11567-11583.
[3]肝脏硬度评估小组.瞬时弹性成像技术诊断肝纤维化专家意见[J].中华肝病杂志,2013,21(6):420-424.
[4]中华医学会传染病与寄生虫病学会,肝病学分会.病毒性肝炎防治方案[J].中华肝脏病杂志,2000,8(6):324-329.
[5]赵晓薇.ADA与ALT AST GGT联合检测在肝脏疾病诊断中的意义[J].内蒙古医学杂志,2014,46(11):1357-1358.
[6]Adams LA,Bulsara M,Rossi E,et al.Hepascore:an accurate validated predictor of liver fibrosis in chronic hepatitis C infection[J].Clin Chem,2005,51(10):1867-1873.
[7]Zeng MD,Lu LG,Mao YM,et al.Prediction of significant fibrosis in HBe Ag-positive patients with chronic hepatitis B by a noninvasive model[J].Hepatology,2005,42(6):1437-1445.
[8]Forns X,Ampurdanes S,Llovet JM,et al.Identification of chronic hepatitis C patients without hepatic fibrosis by a simple predictive model[J].Hepatology,2002,36:986-992.
[9]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.
[10]叶柱均,叶淑霞,张飞勇,等.ADA与ALT、AST、GGT联合检测在肝脏疾病诊断中的意义[J].国际医药卫生导报,2016,22(9):1300-1302.
[11]Friedman SL.Evolving challenges in hepatic fibrosis[J].Nat Rev Gastroenterol Hepatol,2010,7(8):425-436.
[12]Hou XY,Ellis MK,Mc Manus DP,et al.Diagnostic value of noninvasive bio-markers for stage-specific fibrosis in patients with advanced schist-osomiasis japonca[J].Int J Parasitol,2011,41(34):325-332.
[13]苑同业,刘国凤,张立营.血清Ⅲ型前胶原测定在肝纤维化诊断中的应用[J].中国疗养医学,2014,23(9):238-240.
[14]张伟龙,肖翔,潘理想,等.ADA与ALT、AST、GGT联合检测在肝脏疾病诊断的价值研究[J].中国实用医药,2017,12(1):53-55.
[15]Keef EB,Dieterich DT,Han SH,et al.A treatment algorithm for the management of chronic hepatitis B virus in ghe United States:an update[J].Clin Gastroenterol Hepatol,2006,4(8):936-962.
[16]许庆忠,雷霆雯,马小平.实验性肝纤维化血清白蛋白含量变化[J].贵阳医学院学报,2002,27(2):128-129.
[17]夏琦,符节海,蔡卫民,等.AST/PLT和GGT/PLT比值对肝纤维化的诊断价值[J].临床肝胆病杂志,2007,23(5):364-366.
[18]董菁菁,郭宏华.GGT与PLT比值指数对慢性乙型肝炎肝纤维化的诊断价值[J].临床肝胆病杂志,2017,33(3):548-551.