SEARCH

SEARCH BY CITATION

Abstract

  1. Top of page
  2. Abstract
  3. Materials and Methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References

Assessment of liver fibrosis in patients with chronic hepatitis C (CHC) is critical for predicting disease progression and determining future antiviral therapy. LecT-Hepa, a new glyco-marker derived from fibrosis-related glyco-alteration of serum alpha 1-acid glycoprotein, was used to differentiate cirrhosis from chronic hepatitis in a single-center study. Herein, we aimed to validate this new glyco-marker for estimating liver fibrosis in a multicenter study. Overall, 183 CHC patients were recruited from 5 liver centers. The parameters Aspergillus oryzae lectin (AOL) / Dature stramonium lectin (DSA) and Maackia amurensis lectin (MAL)/DSA were measured using a bedside clinical chemistry analyzer in order to calculate LecT-Hepa levels. The data were compared with those of seven other noninvasive biochemical markers and tests (hyaluronic acid, tissue inhibitor of metalloproteases-1, platelet count, aspartate aminotransferase-to-platelet ratio index [APRI], Forns index, Fib-4 index, and Zeng's score) for assessing liver fibrosis using the receiver-operating characteristic curve. LecT-Hepa correlated well with the fibrosis stage as determined by liver biopsy. The area under the curve (AUC), sensitivity, and specificity of LecT-Hepa were 0.802, 59.6%, and 89.9%, respectively, for significant fibrosis; 0.882, 83.3%, and 80.0%, respectively, for severe fibrosis; and 0.929, 84.6%, and 88.5%, respectively, for cirrhosis. AUC scores of LecT-Hepa at each fibrosis stage were greater than those of the seven aforementioned noninvasive tests and markers. Conclusion: The efficacy of LecT-Hepa, a glyco-marker developed using glycoproteomics, for estimating liver fibrosis was demonstrated in a multicenter study. LecT-Hepa given by a combination of the two glyco-parameters is a reliable method for determining the fibrosis stage and is a potential substitute for liver biopsy. (HEPATOLOGY 2012)

Accurate staging of hepatic fibrosis in patients with chronic hepatitis C (CHC) is most important for predicting disease progression and determining the need for initiating antiviral therapy, such as interferon (IFN) therapy.1, 2 Liver biopsy has been considered the gold standard for fibrosis staging for many years.3 However, liver biopsy is invasive and painful,4, 5 with rare but potentially life-threatening complications.6 In addition, this method may suffer from sampling errors since only 1/50,000 of the organ is examined.7 Furthermore, inter- and intraobserver discrepancies reaching levels of 10% to 20% have been reported using this method, leading to misdiagnosis of cirrhosis.8 Therefore, finding a noninvasive method for diagnosing liver fibrosis is an emerging issue in the care of patients with CHC.

Several methods have been studied for the noninvasive diagnosis of hepatic fibrosis or cirrhosis, including clinical9 or blood markers,10, 11 and signal analysis (ultrasonography, magnetic resonance imaging, and elastography).12, 13 Although each method can play a substantial role in the diagnosis of cirrhosis, it is evident that the best way of monitoring hepatitis progression employs an accurate serological method for the quantitative evaluation of fibrosis. We developed a new glyco-marker using multiple lectins that performed well in estimating liver fibrosis in a single-center study.14, 15

Recent progress in glycoproteomics has had a great influence on work toward ideal, disease-specific biomarkers for a number of conditions. Glycoproteins that exhibit disease-associated glyco-alteration and are present in serum or other fluids have the potential to act as biomarkers for the diagnosis of a target disease,16 because the features of glycosylation depend on the extent of cell differentiation and the stage of the cell. Detecting hepatic disease-associated glyco-markers for clinical applications has been a continuous challenge since the early 1990s, because increased fucosylation on complex-type N-glycans has been frequently detected in glycoproteins from patients with hepatocellular carcinoma (HCC) and cirrhosis.17, 18 Of all the alpha-fetoprotein (AFP) glycoforms, more than 30% have been found to react to a fucose-binding lectin, Lens culinaris agglutinin. This fraction, designated AFP-L3, was approved by the U.S. Food and Drug Administration (FDA) in 2005 for the diagnosis and prognosis of HCC.19 We have found that two fibrosis-indicator lectins (Aspergillus oryzae lectin [AOL] and Maackia amurensis lectin [MAL]) together with an internal, standard lectin (Datura stramonium lectin [DSA]) on an alpha 1-acid glycoprotein (AGP) could, using lectin microarray, clearly distinguish between cirrhosis and chronic hepatitis patients.14 We have further simplified this quantitative method so that it could be performed using bedside, clinical chemistry analyzers.15

The aim of the current study was to evaluate this new glyco-marker (LecT-Hepa) using multiple lectins and bedside clinical chemistry analyzers for use in the assessment of liver fibrosis. In this multicenter study we compared the method's efficiency in estimating liver fibrosis with other noninvasive fibrosis markers and tests.

Materials and Methods

  1. Top of page
  2. Abstract
  3. Materials and Methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References

Study Population.

This study included 183 consecutive adult patients with CHC who had undergone percutaneous liver biopsy at one of the following institutions: Hokkaido University Hospital, Musashino Red Cross Hospital, National Center for Global Health and Medicine, Hyogo College of Medicine Hospital, or Nagoya City University Hospital in Japan. A diagnosis of CHC was defined as detectable serum anti-hepatitis C virus (HCV) antibody and HCV-RNA, found using polymerase chain reaction assays, of at least 2 points. Exclusion criteria were coinfection with hepatitis B virus or human immunodeficiency virus (HIV), and other disorders that commonly cause liver diseases. Informed consent was obtained from each patient who participated in the study. This study was conducted in accordance with the provisions of the Declaration of Helsinki and was approved by our Institutional Review Board.

Histological Staging.

Ultrasonography-guided liver biopsy was performed according to a standardized protocol. Specimens were fixed, paraffin-embedded, and stained with hematoxylin-eosin and Masson's trichrome. A minimum of six portal tracts in the specimen were required for diagnosis. All liver biopsy samples were independently evaluated by two senior pathologists who were blinded to the clinical data. Liver fibrosis stages were assessed using METAVIR fibrosis (F) staging.20 Significant fibrosis was defined as METAVIR F ≥2, severe fibrosis as METAVIR F ≥3, and cirrhosis as METAVIR F4. Two patients were excluded from the study because of inadequate histological samples.

Clinical and Biological Data.

The age and sex of the patients were recorded. Serum samples were collected immediately before or no more than 2 months after liver biopsy and were stored at −80°C until analysis. The concentrations of the following variables were obtained by analyzing the serum samples: aspartate aminotransferase (AST), alanine aminotransferase (ALT), gamma-glutamyltransferase (GGT), total bilirubin, albumin, cholinesterase, total cholesterol, platelet count (platelets), prothrombin time, haptoglobin, hyaluronic acid (HA), α2-macroglobulin (α2-MG), tissue inhibitors of metalloproteinases 1 (TIMP1). The aspartate aminotransferase-to-platelet ratio index (APRI), Fib-4 index, Forns index, and Zeng's score were calculated according to published formulae appropriate to each measure.2, 7, 21, 22

Rapid Lectin-Antibody Sandwich Immunoassay Using HISCL.

Fibrosis-specific glyco-alteration of AGP was qualified from simultaneous measurements of the lectin-antibody sandwich immunoassays using three lectins (DSA, MAL, and AOL). In principle, the glycan part of the AGP was captured by the lectin immobilized on the magnetic beads, and the captured AGP was then quantified by an antihuman AGP mouse monoclonal antibody probe that was cross-linked to an alkaline phosphatase (ALP-αAGP). The assay manipulation was fully automated using a chemiluminescence enzyme immunoassay machine (HISCL-2000i; Sysmex, Kobe, Japan). We used the following criterion formula, named the “LecT-Hepa Test,” to enhance the diagnostic accuracy by combining two glyco-parameters (AOL/DSA and MAL/DSA) as described before: F = Log10[AOL/DSA]*8.6-[MAL/DSA].15

Statistical Analyses.

Quantitative variables were expressed as the mean ± standard deviation (SD) unless otherwise specified. Categorical variables were compared using a chi-squared test or Fisher's exact test, as appropriate, and continuous variables were compared using the Mann-Whitney U test. P < 0.05 was considered statistically significant. A multivariate forward stepwise logistic regression analysis was performed to determine the independent predictors of the absence or presence of significant fibrosis, severe fibrosis, and cirrhosis, respectively. Pearson's correlation coefficient was used as necessary. To assess the classification efficiencies of various markers for detecting significant fibrosis, severe fibrosis, and cirrhosis,23 and to determine area under the curve (AUC) values, receiver-operating characteristic (ROC) curve analysis was also carried out. Diagnostic accuracy was expressed as the diagnostic specificity (specificity), diagnostic sensitivity (sensitivity), positive predictive values (PPV), negative predictive values (NPV), positive likelihood ratio (LR [+]), negative likelihood ratio (LR [−]), and AUC (95% confidence interval [95% CI]). We performed statistical analyses using STATA v. 11.0 (StataCorp, College Station, TX).

Results

  1. Top of page
  2. Abstract
  3. Materials and Methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References

Baseline Characteristics of the 183 Patients with Chronic Hepatitis C at the Time of Liver Biopsy.

Patient characteristics at the time of liver biopsy are shown in Table 1. The mean age of the 183 patients was 57.6 ± 11.4 years, and 75 (41%) of them were men. F0-F1 was diagnosed in 89 cases (48.6%), F2 in 46 (25.1%), F3 in 22 (12.0%), and F4 (cirrhosis) in 26 (14.2%).

Table 1. Baseline Characteristics of the 183 Patients with Chronic Hepatitis C at the Time of Liver Biopsy
FeaturesTotal (n = 183)
Age (years)57.6 ± 11.4
Male sex75 (41.0)
AST (IU/L)57.4 ± 43.9
ALT (IU/L)62.8 ± 56.8
GGT (IU/L)51.1 ± 62.6
Bilirubin (mg/dL)0.7 ± 0.4
Albumin (g/L)4.1 ± 0.4
Cholinesterase (IU/L)283.5 ± 97.0
Cholesterol (mg/dL)174.1 ± 35.5
Platelets (109/L)163 ± 57
Prothrombin time (%)87.2 ± 33.4
α2-MG (g/L)356.8 ± 133.1
HA (μg/L)205.3 ± 428.0
TIMP1 (pg/ml)210.6 ± 87.7
AOL/DSA6.3 ± 12.3
MAL/DSA9.0 ± 3.1
Fibrosis stage (%): 
F0-189 (48.6)
F246 (25.1)
F322 (12.0)
F426 (14.2)

Comparison of Variables Associated with the Presence of Significant Fibrosis by Univariate and Multivariate Analysis.

Variables associated with the presence of significant fibrosis were assessed by univariate and multivariate analysis (Table 2). The variables of age (P = 0.001), AST (P < 0.0001), ALT (P < 0.0001), GGT (P < 0.0001), bilirubin (P = 0.014), α2-MG (P = 0.002), HA (P < 0.0001), TIMP1 (P < 0.0001), and AOL/DSA (P < 0.0001) were significantly higher in the significant fibrosis group than in the not significant fibrosis group. The variables albumin (P < 0.001), cholinesterase (P < 0.0001), cholesterol (P = 0.005), platelets (P < 0.0001), prothrombin time (P = 0.0001), and MAL/DSA (P < 0.0001) were significantly lower in the significant fibrosis group than in the not significant fibrosis group. Multivariate analysis showed that platelets (odds ratio [OR]: 0.87, 95% CI: 0.77-0.99), HA (OR: 1.01, 95% CI: 1.01-1.02), and AOL/DSA (OR: 1.51, 95% CI: 1.07-2.15) were independently associated with the presence of significant fibrosis.

Table 2. Variables Associated with the Presence of Significant Fibrosis (F2-4) and Severe Fibrosis (F3-4) by Univariate and Multivariate Analysis
FeaturesNo Significant Fibrosis (n = 89)Significant Fibrosis (n = 94)P Value (Univariate)Odds Ratio (95% CI) (Multivariate)No Severe Fibrosis (n = 135)Severe Fibrosis (n = 48)P ValueOdds Ratio (95% CI) (Multivariate)
Age (years)54.7 ± 11.860.5 ± 10.40.001 55.8 ± 11.962.9 ± 7.80.0011.15
(1.02-1.31)
Male sex (%)30 (33.7)45 (47.9)0.051 52 (38.5)23 (47.9)0.255 
AST (IU/L)45.7 ± 41.668.3 ± 43.5<0.0001 49.7 ± 40.179.1 ± 47.4<0.0001 
ALT (IU/L)51.0 ± 56.674.0 ± 54.9<0.0001 55.9 ± 54.982.5 ± 57.9<0.0001 
GGT (IU/L)40.6 ± 61.762.1 ± 63.1<0.0001 45.5 ± 67.165.8 ± 46.7<0.0001 
Bilirubin (mg/dL)0.6 ± 0.30.7 ± 0.40.014 0.6 ± 0.30.8 ± 0.40.005 
Albumin (g/L)4.2 ± 0.34.0 ± 0.5<0.001 4.2 ± 0.33.8 ± 0.5<0.0001 
Cholinesterase (IU/L)329.2 ± 76.0247.2 ± 96.9<0.0001 312.4 ± 84.4217 ± 91.9<0.0001 
Cholesterol (mg/dL)181.0 ± 31.5167.5 ± 36.20.005 178.1 ± 34.1162.4 ± 33.50.016 
Platelets (109/L)186 ± 53142 ± 52<0.00010.87180 ± 52119 ± 46<0.00010.74
(0.77-0.99)(0.58-0.94)
Prothrombin time (%)94.7 ± 33.480.1 ± 32.10.0001 89.5 ± 36.280.8 ± 23.2<0.001 
α2-MG (g/L)326 ± 117.7389.2 ± 141.10.002 331.1 ± 122.5423.9 ± 137.5<0.0001 
HA (μg/L)85.6 ± 154.3318.7 ± 556.1<0.00011.01115.4 ± 201.1458.2 ± 711.0<0.0001 
(1.01-1.02)
TIMP1 (pg/ml)183.5 ± 53.3238.6 ± 106.1<0.0001 189.7 ± 64.5263.9 ± 113.8<0.0001 
AOL/DSA1.4 ± 1.210.9 ± 15.9<0.00011.512.0 ± 2.618.3 ± 19.3<0.0001 
(1.07-2.15)
MAL/DSA10.6 ± 1.77.5 ± 3.4<0.0001 10.2 ± 2.05.6 ± 3.4<0.00010.52
(0.37-0.76)

Comparison of Variables Associated with the Presence of Severe Fibrosis by Univariate and Multivariate Analysis.

Variables associated with the presence of severe fibrosis were assessed by univariate and multivariate analysis (Table 2). The variables of age (P = 0.001), AST (P < 0.0001), ALT (P < 0.0001), GGT (P < 0.0001), bilirubin (P = 0.005), α2-MG (P < 0.0001), HA (P < 0.0001), TIMP1 (P < 0.0001), and AOL/DSA (P < 0.0001) were significantly higher in the severe fibrosis group than in the no severe fibrosis group. The variables albumin (P < 0.0001), cholinesterase (P < 0.0001), cholesterol (P = 0.016), platelets (P < 0.0001), prothrombin time (P < 0.001), and MAL/DSA (P < 0.0001) were significantly lower in the severe fibrosis group than in the no severe fibrosis group. Multivariate analysis showed that age (OR: 1.15, 95% CI: 1.02-1.31), platelets (OR: 0.74, 95% CI: 0.58-0.94), and MAL/DSA (OR: 0.52, 95% CI: 0.37-0.76) were independently associated with the presence of severe fibrosis.

Comparison of Variables Associated with the Presence of Cirrhosis by Univariate and Multivariate Analysis.

Variables associated with the presence of cirrhosis were assessed by univariate and multivariate analysis (Table 3). Age (P = 0.0016), AST (P = 0.016), GGT (P = 0.0031), bilirubin (P < 0.0001), α2-MG (P = 0.019), HA (P < 0.0001), TIMP1 (P < 0.0001), and AOL/DSA (P < 0.0001) were significantly higher in the cirrhosis group than in the no cirrhosis group. Albumin (P < 0.0001), cholinesterase (P < 0.0001), cholesterol (P < 0.0001), platelets (P < 0.0001), prothrombin time (P = 0.0004), and MAL/DSA (P < 0.0001) were significantly lower in the cirrhosis group than in the no cirrhosis group. Multivariate analysis showed that platelets (OR: 0.76, 95% CI: 0.58-0.99) and MAL/DSA (OR: 0.67, 95% CI: 0.49-0.90) were independently associated with the presence of cirrhosis.

Table 3. Variables Associated with the Presence of Cirrhosis (F4) by Univariate and Multivariate Analysis
FeaturesNo Cirrhosis (n=157)Cirrhosis (n = 26)P ValueOdds Ratio (95% CI) (Multivariate)
Age (years)56.6 ± 11.763.8 ± 7.30.0016 
Male sex (%)60 (38.2)15 (57.7)0.061 
AST (IU/L)54.6 ± 41.774.9 ± 53.70.016 
ALT (IU/L)62.1 ± 58.167.2 ± 48.20.446 
GGT (IU/L)48.5 ± 63.964.9 ± 53.80.0031 
Bilirubin (mg/dL)0.6 ± 0.31.0 ± 0.5<0.0001 
Albumin (g/L)4.2 ± 0.43.6 ± 0.5<0.0001 
Cholinesterase (IU/L)305.3 ± 83.9181.7 ± 90.1<0.0001 
Cholesterol (mg/dL)178.4 ± 33.3146.9 ± 29.8<0.0001 
Platelets (109/L)172 ± 54106 ± 36<0.00010.76
(0.58-0.99)
Prothrombin time (%)88.7 ± 35.579.2 ± 16.10.0004 
α2-MG (g/L)346.2 ± 131.6416.9 ± 127.80.019 
HA (μg/L)137.1 ± 215.7617.4 ± 915.1<0.0001 
TIMP1 (pg/ml)196.4 ± 70.4287.3 ± 126.6<0.0001 
AOL/DSA3.4 ± 7.124.0 ± 20.4<0.0001 
MAL/DSA9.8 ± 2.44.2 ± 2.8<0.00010.67
(0.49-0.90)

Evaluation of the Two Glyco-Parameters AOL/DSA and MAL/DSA for Estimating the Progression of Liver Fibrosis.

To assess the correlation of the two obtained glyco-parameters with the progression of fibrosis, we analyzed the data of triple lectins from HISCL measurements on the 183 CHC patients. The boxplots of AOL/DSA and MAL/DSA in relation to the fibrosis staging are shown in Fig. 1A,B, respectively. The AOL/DSA values gradually increased with the progression of fibrosis and Pearson's correlation efficient was R = 0.61. On the other hand, the MAL/DSA values gradually decreased with the progression of fibrosis and Pearson's correlation efficient was R = −0.69. Both parameters fitted the quantification of the progression of fibrosis from F2 to F4.

thumbnail image

Figure 1. Boxplot of (A) AOL/DSA, (B) MAL/DSA, and (C) LecT-Hepa in relation to the fibrosis score. The box represents the interquartile range. The whiskers indicate the highest and lowest values, and the dots represent outliers. The line across the box indicates the median value. Correlation of AOL/DSA, MAL/DSA, and LecT-Hepa was measured by HISCL with the progression of liver fibrosis. R: Pearson's correlation coefficient.

Download figure to PowerPoint

LecT-Hepa, Combined with Two Glyco-Parameters, Was Evaluated in the Diagnosis of Significant Fibrosis, Severe Fibrosis, and Cirrhosis.

LecT-Hepa was calculated using two glyco-parameters (AOL/DSA and MAL/DSA). The boxplots of LecT-Hepa in relation to the fibrosis staging are shown in Fig. 2. The LecT-Hepa values gradually increased with the progression of fibrosis. Pearson's correlation coefficient between LecT-Hepa and liver fibrosis was very high (R = 0.72), and was superior to those for AOL/DSA (R = 0.61) and MAL/DSA (R = −0.69). We next examined AUC to characterize the diagnostic accuracy of LecT-Hepa at each stage of fibrosis, i.e., significant fibrosis (F2/F3/F4), severe fibrosis (F3/F4), and cirrhosis (F4). For the prediction of significant fibrosis, AUC (95% CI), sensitivity, specificity, PPV, NPV, LR (+), and LR (−) of the test were 0.802 (0.738-0.865), 59.6%, 89.9%, 85.7%, 66.7%, 5.89, and 0.45, respectively (Fig. 3A). For the prediction of severe fibrosis, AUC (95% CI), sensitivity, specificity, PPV, NPV, LR (+), and LR (−) were 0.882, 83.3%, 80.0%, 59.7%, 93.1%, 4.17, and 0.21, respectively (Fig. 3B). For the prediction of cirrhosis, AUC (95% CI), sensitivity, specificity, PPV, NPV, LR (+), and LR (−) were 0.929 (0.896-0.976), 84.6%, 88.5%, 58.8%, 97.2%, 7.38, and 0.17, respectively (Fig. 3C).

thumbnail image

Figure 2. ROC curves of LecT-Hepa to distinguish between significant fibrosis and no significant fibrosis in patients with chronic hepatitis C (A); severe fibrosis and no severe fibrosis (B); cirrhosis and no cirrhosis (C). AUC: area under the receiver operating characteristic curve; PPV: positive predictive values; NPV: negative predictive values; LR (+): positive likelihood ratio; LR (−): negative likelihood ratio.

Download figure to PowerPoint

thumbnail image

Figure 3. Comparison of ROC curves in the performance of LecT-Hepa, HA, TIMP1, Plt, APRI, Fib-4 Index, Forns index, Zeng's score for the diagnosis of significant fibrosis (A), severe fibrosis (B), and cirrhosis (C). ROC: receiver operating characteristic curve; TIMP1: tissue inhibitors of metalloproteinases 1; Plt: platelet count; HA: hyaluronic acid.

Download figure to PowerPoint

Comparison of AUC, Sensitivity, Specificity, PPV, and NPV for Predicting the Diagnosis of Significant Fibrosis, Severe Fibrosis, and Cirrhosis.

ROC curves of LecT-Hepa, HA, TIMP1, platelets, APRI, Forns index, Fib-4 index, and Zeng's score for predicting significant fibrosis, severe fibrosis, and cirrhosis were plotted, as shown in Fig. 3A-C. The AUC of LecT-Hepa for predicting significant fibrosis (0.802) was superior to HA (0.756), TIMP1 (0.697), platelets (0.729), APRI (0.777), Fib-4 index (0.747), Forns index (0.783), and Zeng's score (0.791). For predicting severe fibrosis, AUC of LecT-Hepa (0.882) was superior to HA (0.839), TIMP1 (0.753), platelet count (0.821), APRI (0.840), Fib-4 index (0.811), Forns index (0.861), and Zeng's score (0.863). For predicting cirrhosis, AUC of LecT-Hepa (0.929) was superior to HA (0.866), TIMP1 (0.783), platelets (0.851), APRI (0.787), Fib-4 index (0.856), Forns index (0.887), and Zeng's score (0.853). Sensitivity, specificity, PPV, and NPV by eight noninvasive tests and markers are shown in Table 4. In general, indicators of LecT-Hepa were superior to other noninvasive tests and markers. Specificity and PPV used to distinguish significant fibrosis in LecT-Hepa were superior to those in other tests and markers, although sensitivity and NPV by LecT-Hepa (59.6% and 66.7%, respectively) to distinguish significant fibrosis were inferior to those in other tests and markers. When distinguishing severe fibrosis, the categories of sensitivity (83.3%), specificity (80.0%), PPV (59.7%), and NPV (93.1%) for LecT-Hepa were superior to those in other tests and markers, except for specificity (82.2%) and PPV (61.0%) in HA. When distinguishing cirrhosis, the categories of sensitivity (84.6%), specificity (88.5%), PPV (58.8%), and NPV (97.2%) in LecT-Hepa were superior to those in other tests and markers, except for sensitivity by HA (88.5%), Forns index (84.6%), and Zeng's score (92.3%) and NPV by Zeng's score (98.3%).

Table 4. Diagnostic Performance of Biochemical Markers and Scores by Stage of Fibrosis
 No Significant Fibrosis (F0-1) vs. Significant Fibrosis (F2-4)No Severe Fibrosis (F0-2) vs. Severe Fibrosis (F3-4)No Cirrhosis (F0-3) vs. Cirrhosis (F4)
 AUC (95% CI)Se (%)Sp (%)PPV (%)NPV (%)AUC (95% CI)Se (%)Sp (%)PPV (%)NPV (%)AUC (95% CI)Se (%)Sp (%)PPV (%)NPV (%)
  1. AUC, area under the ROC curve; CI, confidence interval; Se, sensitivity; Sp, specificity; PPV, positive predictive values; NPV, negative predictive values.

LecT-Hepa0.802 (0.738-0.865)59.689.985.766.70.882 (0.830-0.949)83.38059.793.10.929 (0.896-0.976)84.688.558.897.2
HA0.756 (0.684-0.827)68.178.777.869.60.839 (0.771-0.908)77.182.26190.30.866 (0.790-0.942)88.575.837.396.8
TIMP10.697 (0.619-0.774)65.971.970.460.70.753 (0.665-0.841)7576.35388.90.783 (0.710-0.887)80.874.527.894.6
Platelets0.72978.761.968.573.50.82181.370.449.491.30.85184.670.732.395.8
(0.656-0.803)(0.751-0.891)(0.785-0.918)
APRI0.777 (0.709-0.844)71.371.972.268.80.840 (0.780-0.900)81.372.650.691.50.787 (0.703-0.871)76.968.227.993.9
Fib-40.747 (0.671-0.818)65.976.474.7680.811 (0.733-0.889)77.173.35089.20.856 (0.788-0.924)73.180.937.594.1
Forns0.783 (0.716-0.852)73.477.577.573.40.861 (0.802-0.920)81.371.15091.40.887 (0.831-0.943)84.675.236.196.7
Zeng0.791 (0.723-0.858)82.970.77579.70.863 (0.799-0.925)81.379.859.592.80.853 (0.783-0.933)92.373.936.998.3

Discussion

  1. Top of page
  2. Abstract
  3. Materials and Methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References

Our results showed that the LecT-Hepa test, calculated by combining two glyco-parameters (AOL/DSA and MAL/DSA), had higher sensitivity and specificity for diagnosing severe fibrosis and cirrhosis compared to other noninvasive tests and markers for these conditions. The new glyco-marker we have developed is based on the glyco-alteration on the AGP, which is mainly synthesized in the liver. AGP has been considered one of the best candidates for glyco-markers in liver fibrosis or HCC. This is because it is a well-characterized glycoprotein with five highly branched, complex-type N-glycans, whose alteration (e.g., desialylation, increased branching, and increased fucosylation) occurs during the progression of liver fibrosis and carcinogenesis.24 It has already been reported that an increased degree of fucosylation was detected in cirrhosis patients using a fucose-binding lectin (AAL)-antibody sandwich ELISA and an automated analyzer.24 The detection of asialo-AGP using lactosamine-recognition lectin RCA120 has also been reported as an alternative method for finding cirrhosis.25 Meanwhile, we detected many other aspects of glyco-alteration of AGP using a multiplex sandwich immunoassay with a 43-lectin microarray,26 resulting in the selection of three lectins—MAL, AOL, and DSA—to serve, collectively, as a fibrosis indicator and a signal normalizer.14 Since two glyco-parameters (AOL/DSA and MAL/DSA) on AGP are normalized by an internal standard lectin (DSA), LecT-Hepa is not influenced by the amount of AGP. We confirmed that the use of this lectin set was statistically superior to the previously selected lectins (AAL and RCA120).

This triplex-sandwich immunoassay employing DSA/MAL/AOL lectins and an anti-AGP antibody from the lectin microarray has already been converted to a fully automated immunoassay analyzer (HISCL-2000i) for clinical use.15 Pretreatment requires 3 hours, and quantifying the two glyco-parameters for the LecT-Hepa to use this automated analyzer takes 17 minutes. Currently, we can obtain data from LecT-Hepa to predict liver fibrosis on the same day of blood sample collection. This simple and reliable glyco-marker may be suitable for clinical use, and may substitute for liver biopsy in some cases.

We are confident that our study samples are representative of most patients. The AUC scores for distinguishing significant fibrosis, severe fibrosis, and cirrhosis by APRI, HA, Fib-4 index, Forns index, and Zeng's score were not significantly different from those in previous studies.11, 27, 28 Every serum sample in this study was obtained from a patient immediately before or no more than 2 months after liver biopsy. As many serum samples as possible were collected from each liver center to eliminate a selection bias in any center. Since we could not perform liver biopsy on the patients who had a tendency to develop hemorrhages, fewer samples of severe fibrosis and cirrhosis were collected than those of milder fibrosis. In fact, the population of fibrosis staging in this study was similar to that of a previous, large prospective study evaluating noninvasive fibrosis markers.29 In addition, we did not include patients with obvious decompensated cirrhosis. This is because inclusion of patients with severe liver disease would have artificially improved the predictive values of the logistic function. On the other hand, we included many patients with mild histological features (48.6% with F0-1). Sampling variation poses potential difficulties, especially in the early stages of disease, when fibrosis might be unevenly distributed.

There are several advantages in using reliable noninvasive markers for assessing liver fibrosis. First, they can be used to accurately determine the appropriate time for initiating IFN treatment in CHC patients. These markers can also help monitor and assess the therapeutic efficacy of IFN treatment in improving liver function in cases of liver fibrosis and cirrhosis. Finally, these markers will be essential in the development of new, antifibrotic treatments. Recently, many directed or targeted therapies against liver fibrosis, such as anti-transforming growth factor beta and anti-tumor necrosis factor alpha compounds have been developed.30, 31 To evaluate these new drugs, reliable and simple noninvasive fibrosis markers are needed. LecT-Hepa appears to be one of the most prominent candidates to serve as a marker for developing antifibrotic drugs.

In conclusion, both glyco-parameters (AOL/DSA and MAL/DSA) using lectins in a bedside, clinical chemical analyzer succeeded in the quantification of the progression of liver fibrosis. Using LecT-Hepa, the combination score of both AOL/DSA and MAL/DSA is a reliable method for determining fibrosis staging and can be a good substitute for liver biopsy.

Acknowledgements

  1. Top of page
  2. Abstract
  3. Materials and Methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References

We thank K. Saito, S. Unno, T. Fukuda, and M. Sogabe (AIST) for technical assistance. We also thank C. Tsuruno, S. Nagai, and Y. Takahama (Sysmex Co.) for critical discussion.

References

  1. Top of page
  2. Abstract
  3. Materials and Methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References