Diagnostic Accuracy of Plasma Osteopontin in Egyptian Hepatocellular ‎Carcinoma Patients

INTRODUCTION

Hepatocellular carcinoma (HCC) is the most common primary liver cancer. According to WHO, it is the 5th most common cancer in the world and it is considered as the second cause of cancer-caused deaths [1, 2].

One can observe the higher burden of HCC in developing countries, where it has unique geographic, sex, and age distributions that are usually determined by variable etiologic factors [3].

Chronic viral hepatitis-induced cirrhosis due to chronic hepatitis B virus (HBV) and hepatitis C virus (HCV) infections corresponds to a major risk factor for HCC, representing about 85% of HCC cases worldwide. Additional risk factors that should be considered include metabolic disorders such as nonalcoholic fatty liver disease (NAFLD) and chronic alcohol consumption [4].

Based on previous reports, HCC in Egypt is the fourth most prevalent cancer, and the second cause of cancer-induced death in both males and females. It is worth to mention that HCC incidence increased to the double over the last ten years [5,6].

Despite the fact that using direct-acting antivirals (DAAs) for the treatment of HCV has improved the possibility of significantly reducing the rate of developing HCC in such patients [7], the consequences of such treatment on the possibility of developing HCC in cirrhotic patients is still questionable [8,9].

Unfortunately, no changes in incidence and mortality rates for HCC are observed, signifying the poor overall survival of HCC patients. This makes the early diagnosis of HCC to be considered as the most effective approach to decrease the mortality rates [10,11]. In fact, lacking specific symptoms during the early stages of HCC developments plays a major role in the delayed diagnosis and hence, poor prognosis [12].

One of the recommended interventions for liver cirrhosis patients is the regular screening for alpha-fetoprotein (AFP) and ultrasonography every 6 months. This enables early detection of HCC, allowing for the application of effective treatment strategies [13].

Screening by this combination (ultrasound & AFP) has been useful in Asia where hepatitis infection mostly HBV are common. In contrast, no data supports the utilization of AFP in the United States & Europe due to its low sensitivity and specificity [14].

A retrospective epidemiological study in Egypt showed that 60% of HCC cases had AFP levels below 200ng/ml; the diagnostic level of this marker. Therefore, multiple efforts are directed towards detecting new biomarkers for diagnosis of HCC. In addition, up to date research in genomics & proteomics fields could provide new tools that can aid in the diagnosis and predicting prognosis of HCC [15, 16].

In the last decade, several additional biomarkers have been suggested for diagnosis and screening of HCC including Osteopontin (OPN). OPN is an integrin-binding glycophosphoprotein that is produced and secreted by activated macrophages and T-lymphocytes [17, 18]. In addition to its role in several physiological cellular functions, it has been reported to be overexpressed in different kinds of human tumors, including liver carcinomas [19, 20]. Both increased tumor vascular invasion and advanced tumor grade have been associated with over-expression of OPN. Some research groups suggested a predictive potential for OPN in case of HCC invasion and metastasis [21, 22].

Despite being introduced as a promising biomarker for HCC diagnosis, further studies should be conducted to assess the diagnostic value of plasma OPN in HCC patients in different populations.

So that, we aimed in this study at assessing the diagnostic value of plasma osteopontin (OPN) compared with alpha-fetoprotein (AFP) for the diagnosis of hepatocellular carcinoma (HCC) in Egyptian patients.

PATIENTS AND METHODS

Study design: It is a prospective case control study.

Study settings: This study was carried out in the Department of Tropical Medicine and Gastroenterology, Assiut University Hospital, Alrajhi liver hospital in Egypt, between May 2018 and July 2019 after approval from the insitutional ethical committee (IRB No.  17200782)

Study groups: The study was conducted on 120 subjects subdivided into 3 groups: Group I includ­ed 60 HCC patients; Group II includ­ed 30 liver cirrhosis patients; Group III in­cluded 30 healthy subjects (control group).

Inclusion criteria:

Group I: patients with HCC, diagnosed by one imaging technique showing radiological hallmark characteristic for HCC (contrast uptake in arterial phase and rapid wash out in venus or delayed phase) were included [23, 24].

Group II: liver cirrhosis patients were included. According to previous studies, diagnosis of cirrhosis was based on clinical, biochemical and ultrasonographic data (jaundice, palmar erythema ascites, cutaneous spider, muscle wasting, ecchymosis, prolonged prothrombin time (PT), low serum albumin, coarse bright liver echopattern and nodular surface) [25,26].

Child-Pugh classification was used for grading the severity of liver cirrhosis depending on clinical and biochemical parameters [27].

Group III: apparently healthy subjects, age and sex matched, free from any acute or chronic hepatic disorder and not on any medications were included in the study as control subjects. 

Exclusion criteria

Exclusion criteria of groups included:

1. Having any other tumor than HCC.
2. Receiving any previous treatment for HCC.
3. Suffering from poorly controlled metabolic disorder (diabetes mellitus) or systemic hypertension.

Patients and controls were all Egyptians and were sex and age matched. The study protocol was approved by the institutional review board and the ethical committee, and written informed consents were obtained from all the participants.

Methods

After inclusion in the study and carefully taking subjects’ history, all participants and controls were subjected to a thorough clinical examination and an abdominal ultrasound examination (using 3.5 MHz transducer on Digital Ultraasounic Diagnostic Imaging System, Model SSA-590A (SUP.SYMBOL; SN99A1253291), TOSHIBA Medical Systems Corporation, Japan). In addition, MSCT abdomen/MRI was performed for cases of HCC only.

Laboratory investigations

For laboratory investigations, Blood Samples were withdrawn from selected participants. The blood was collected into EDTA containing test tubes and sodium citrate containing test tubes. Plasma was separated by centrifugation for 15 minutes at 3,000 rpm. Samples were stored in aliquots in eppendorff tubes and frozen at -80 ºC until measurements of OPN, AFP.

Samples were used for routine laboratory tests including:

●   Complete blood picture (CBC): was performed on a Cell-DYN 3700 haematology analyzer; (Abbott -Germany).

●   Kidney function test, blood glucose and liver function tests (LFT): including serum bilirubin, alanine transferase, aspartate transferase, serum albumin, and alkaline phosphatase: performed on Modular P autoanalyzer; Roche diagnostics. Mannheim Ltd., Germany.

●   Hepatitis markers (HBs Ag and HCV antibody): were done on Architect i 1000 sr immunoligy analyzer (abbott- Germany).

●   Prothrombin time and concentration: done on Sysmex coagulation analyzer (Siemes-Germany).

● Tumor markers including:

1.  Assessment of Plasma AFP was done using immulite 2000^® immunoassay system (siemens-Germany).

2.  Assessment of Plasma OPN level: Plasma OPN levels were measured using Human OsteopontinELISA Kit^®according to manufacturer’s instructions (SinoGene Clon Biotech Co., Ltd).

Statistical analysis

Data were analyzed using SPSS version 21*. Means, standard deviations, medians and percentages were calculated. Difference in distribution of frequencies among different groups were compared using Chi square test, whereas, proportions of repeated measures were compared using McNemar test. T-test and ANOVA test were utilized to compare the mean difference between groups. Post-hoc test was calculated using Bonferroni corrections for pairwise comparisons between each two study groups. Independent Sample Kruskal Wallis test was used to test the median differences of the data that don’t follow normal distribution.

ROC curve was depicted to explore the diagnostic performance of OPN and AFP for HCC and Cirrhosis prediction, analyzed as area under the curve (AUC), standard error (SE) and 95% CI. Sensitivity, specificity, PPV, NPV and accuracy were calculated. A p-value equals or less than 0.05 was considered significant.

* IBM_SPSS. (Statistical Package for Social Science). Ver.21. Standard version. Copyright © SPSS Inc., 2011-2012. NY, USA. 2012.

RESULTS

In The current case-control study 120 participants were included. Patients were divided as follows: (Group I) 60 patients with HCC, (Group II) 30 patients with liver cirrhosis and Group III 30 healthy subjects as controls.

1. Patients Characteristics

I. baseline socio-demographic data and risk factors in different groups

Socio-demographic data and risk factors including age, gender, residence being diabetic or smoker are summarized in table 1

II. Clinical and Laboratory Findings among cases (HCC vs. Cirrhosis)

The distribution of clinical data findings among cases was illustrated in table 2.

**Chi-square test was used to compare proportions between groups

2- OPN level differences among the Total HCC cases regarding Demographic and Tumor Characteristics

Older patients (> 60 years) had statistically significant higher levels of OPN (261.8 ng/ml) compared with younger cases (191.9 ng/ml) (p= 0.034). Likely, male patients had statistically significant higher levels of OPN (254.6 ng/ml) compared with females (146.7 ng/ml) (p= 0.010).

Also, patients with large tumors (> 5 cm) had statistically significant higher OPN levels (254.2 ng/ml) compared with those with size < 5cm (128.1 ng/ml) (p= 0.018). Likewise, patients with BCLC stage C & D had statistically significant higher levels of OPN (297.1 ng/ml) compared with those BCLC stage A & B (219.1 ng/ml) (p= 0.021).

On the other hand, the mean OPN level showed non-statistically significant differences regarding Child Pugh, multiplicity or lymph node (LN) (p>0.05) table 3.

3- Tumor Markers Differences between groups

Comparing OPN plasma levels showed that HCC group patients had significantly higher median levels compared with both cirrhotic and control groups. In the same time, cirrhotic cases had significantly higher median levels than control groups (200 vs. 77.5 vs. 25.5 ng/ml, p<0.05) (figure 1).

Also the same patient group of HCC showed significantly higher median levels compared with both cirrhotic and control groups regarding AFP. Similarly, cirrhotic patients had significantly higher AFP median levels than control groups (199.5 vs. 4.5 vs. 2.4 ng/ml, p<0.05) (figure 2).

4- Diagnostic performance of biomarkers for diagnosis of HCC vs. Control

As regard AFP, with the best cut-off point (3.3 ng/ml) for diagnosis of HCC cases vs. controls; AFP had 95% sensitivity (i.e. the marker correctly identified 95% of the HCC cases as cases). Also, it had 93% specificity (i.e. the marker correctly identified 93% of the controls as free from the disease). Additionally, it had 94% PPV (i.e. the marker correctly identified 94% of the HCC cases among all positives). Equally, it had 95% NPV (i.e. the marker correctly identified 95% of the controls among all negatives). The overall accuracy of the marker was 94%.

The diagnostic value of serum AFP levels to diagnose HCC vs. control (AUC = 0.978, p<0.001; 95% CI 0.953-1.000) was lower than that of serum OPN serum level (AUC = 1.000, p<0.001) figure 3.

Considering the best cut-off point for OPN (90 ng/ml) for diagnosis of HCC cases vs. controls; the overall accuracy of the marker was 99%. OPN sensitivity was 98%, specificity 100%; and the predictive value of the positive test was 100%, whereas that of the negative test was 98%.

5- Diagnostic performance of biomarkers for diagnosis of HCC vs. Cirrhotic cases

The diagnostic performance values of serum AFP levels to diagnose HCC vs. cirrhotic cases (AUC = 0.902, p<0.001; 95% CI 0.841-0.964) was lower than that of serum OPN serum level (AUC = 0.923, p<0.001; 95% CI 0.871-0.976) fig 4.

As regard AFP, with the best cut-off point (5.5 ng/ml) for diagnosis of HCC cases vs. cirrhotic cases; AFP had 90% sensitivity (i.e., the marker correctly identified 90% of the HCC cases as cases). Also, it had 63% specificity (i.e., the marker correctly identified 63% of cirrhotic cases as having cirrhosis). Additionally, it had 71% PPV (i.e., the marker correctly identified 71% of the HCC cases among all positives). Equally, it had 68% NPV (i.e., the marker correctly identified 68% of cirrhotic cases among all negatives). The overall accuracy of the marker was 77%.

Regarding the best cut-off point for OPN (90 ng/ml) for diagnosis of HCC cases vs. cirrhotic cases. The sensitivity of OPN was found to be 97%, specificity 70%; the predictive value of the positive test was 77%, while that of the negative test was 96%. The overall accuracy of the marker was 84%.

Table 1: Socio-demographic and Risk Factor Differences between Groups.