Liver biopsy is the best standard for the assessment of liver fibrosis [4
], although wound pain and possible severe complications restrict its clinical practice [5
]. Using non-invasive LSM by TE to assess the stage of liver fibrosis has been proposed in several cohort studies [8
]. A recent meta-analysis [18
] showed that the mean AUROC for the diagnosis of cirrhosis by TE was high, at 0.94 (95% CI 0.93–0.95). The sensitivity was 87% (95% CI 84–90%), specificity 91% (95% CI 89–92), positive likelihood ratio 11.7 (95% CI 7.9–17.1), and negative likelihood ratio 0.14 (95% CI 0.10–0.20). For the detection of cirrhosis, TE is currently the most accurate non-invasive tool and allowed the saving of the liver biopsy in 90% of cases [19
]. In this study, we further demonstrated that LSM was an independent factor in the prediction of HCC presence. And the risk of HCC presence increased with advancement of cirrhosis. With liver stiffness ≥24 kPa, the likelihood ratio of HCC presence was 5.9 for chronic hepatitis patients. The sensitivity and specificity in the prediction of HCC presence was 41.5 and 92.7%, respectively.
SSLR was the statistical method used to evaluate the risk of disease by a fixed optimal cut-off point [16
]. For continuous scores like liver stiffness, SSLRs provide sufficient information for the possibility of disease occurrence by stratifying the population with a given range. Because, with too many strata, the likelihood ratios become unstable and degenerate, we obtained three strata following the principle reported by Peirce and Cornell [16
]. Our previous study indicated 12, 12 and 10 kPa as the cut-offs for cirrhosis in all patients, chronic HCV and HBV patients [12
], the levels of liver stiffness were used as the cut-off between the lower and medium strata. We used 24 kPa as the cut-off between the medium and higher strata due to the point having a significant difference from the two strata. In the current study, it seems that the possibility of HCC presence was smaller in “non-cirrhotic” patients detected by TE. Even in so-called “cirrhotic” patients, the possibility of HCC presence was elevated slightly when the liver stiffness was less than 24 kPa. However, once the stiffness exceeded 24 kPa, this possibility rose steeply, which was comparable with the report on patients with HCV by Masuzaki et al. [13
]. He pointed out that the possibility of HCC presence in HCV-related cirrhosis patients increases from 1.3- to 5-fold when the liver stiffness is higher than 25 kPa. In our study, the risk of HCC presence increased greatly when liver stiffness exceeds 24 kPa, not only for patients with HCV but also for patients with HBV. Hence, the stiffer the liver, the higher risk of HCC the cirrhosis patient might experience.
The likelihood ratio for HCC presence in HCV patients between >24 and <12 kPa was 11.7, whereas in HBV patients between >24 and <10 kPa it was 16. In cirrhotic liver at the same stiffness level, HBV patients seemed to have higher risk of HCC presence than HCV patients. This might be due to the differences between HBV and HCV in the mechanisms of hepatocarcinogenesis [20
]. Except for virus protein in the induction of carcinogenesis in both viruses, HBV viral sequences could be directly integrated into the host genome, which increases the genomic instability and induces HCC development. However, integration of the viral sequences into the host genome does not occur in HCV infection. This difference might explain why patients with HBV infection are at higher risk of developing HCC than patients with HCV infection with the same liver stiffness level.
Those patients with chronic hepatitis enrolled in our study had a higher liver stiffness value due to hepatic inflammation. The degree of liver stiffness in these patients might reduce after their recovery from hepatic inflammation [22
], whose severity of liver stiffness in the current study might be over-estimated. Therefore, the likelihood ratio of HCC presence for liver stiffness exceeding 24 kPa might be higher than estimated in this study, if the ALT level was comparable between patients both with and without HCC.
Patients with liver cirrhosis are a high-risk group for the development of HCC. However, the risk of HCC development might be different in patients with different degrees of liver cirrhosis. Case–control study showed that Child–Pugh A cirrhotic patients with HCC had higher liver stiffness values than those without HCC [23
]. Our study also proved that those cirrhotic patients with HCC had stiffer liver than those without [LSM values (mean ± SD) 27.9 ± 22.1 kPa vs. 20.0 ± 11.9 kPa, p
< 0.001]. A recent prospective and longitudinal study also demonstrated the association between liver stiffness and HCC development in patients with chronic hepatitis C [24
]. LSM had the potential indication of systemic screening of populations at high risk of chronic liver disease [25
]. In our study, LSM was demonstrated useful in stratifying high-risk group of HCC presence in cirrhosis patients. Therefore, it might be a useful tool in screening patients at high risk of HCC.
In clinical practice, regular HCC surveillance with ultrasonography for cirrhosis patients is recommended to detect early stage HCC and improve the patients’ survival [26
]. To yield better cost-effective results, there might be different screening or surveillance programs for HCC stratified according to liver stiffness values for patients with chronic hepatitis or cirrhosis. However, the ideal cut-off value of liver stiffness in the implementation of HCC surveillance program is still unknown. In a cohort of patients with chronic liver disease, the cut-off was 53.7 kPa for diagnosis of HCC presence with a negative predictive value more than 90% [8
]. In our patients population, we determined that 24 kPa is a useful cut-off, with a specificity of more than 90%, to implement screening or surveillance program for HCC in patients with chronic hepatitis. The discrepancy of cut-offs between studies might be explained by differences in the study populations and in etiologies of enrolled patients. Further prospective studies are necessary in determining the ideal cut-off of liver stiffness in the initiation of HCC screening and surveillance programs for patients with chronic hepatitis.
Thrombocytopenia and high APRI were validated markers to identify the risk group for HCC development in cirrhosis patients [28
]. However, the diagnostic accuracies of these serum markers, including platelet count, total bilirubin and APRI, in their prediction of HCC presence were low, with an AUROC between 0.5 and 0.6 in this study. Serum AFP was a more accurate predictor for HCC presence in our study. However, none of these serum markers were liver specific and might be affected by other co-morbid disease. Compared with these serum markers, LSM was a liver-specific method and superior or equal to the serum markers in the prediction of HCC presence in the current study.
In conclusion, LSM with FibroScan® identified the risk group for HCC presence in chronic hepatitis patients, irrespective of the virus etiology. With the cut-off of 24 kPa, LSM had high specificity in the prediction of HCC for patients with chronic hepatitis.