HBV: studies assessing LSM with VCTE

The worldwide prevalence of chronic hepatitis B virus (HBV) infection is estimated at around 400 million individuals [7]. The annual incidence of progression from chronic hepatitis to cirrhosis ranges from 2.1% to 6% and is influenced by older age, lack of viral suppression, and the co-infection with hepatitis C virus (HCV)/hepatitis D virus (HDV) [8]. Overall, the progression to liver cirrhosis and HCC is estimated to reach 40% of subjects with chronic HBV infection [9].

Assuming the need for a reliable non-invasive evaluation of liver fibrosis, Seo et al. [10] have demonstrated with a multivariate analysis that LSM value was an independent predictor of HCC occurrence [hazard ratio (HR): 1.041, p < 0.001], while histology was not (p > 0.05).

Taking into account other NITs for fibrosis staging, the study by Chon et al. [11] demonstrated a higher diagnostic accuracy of baseline LSM [area under the receiver operating curve (AUROC) 0.789] compared to APRI (AUROC 0.729) and FIB-4 (AUROC 0.744) in the prediction of HCC occurrence during follow-up.

Less satisfactory results were found in the study by Wang et al. [12] in which LSM had AUROC 0.636 with a proposed cut-off value of 21.5 kPa, (Sens 55.6%, Spec 64.2%). In addition, liver stiffness variations during follow-up were not demonstrated to be associated with HCC occurrence (Table 1).

HCV: studies assessing LSM with VCTE

It is estimated that HCC develops late, almost after 20 years from HCV infection, during the cirrhosis stage of the disease [23, 24]. Due to its worldwide diffusion, HCV is the leading cause of HCC occurrence in the setting of liver cirrhosis [25]. The natural history of HCV infection has been revolutionized by the antiviral therapy with direct antiviral agent (DAA); however, in patients with advanced liver fibrosis, the risk of HCC development remains high [26] even if in sustained virological response (SVR).

The study by Wang et al. [27] showed at multivariate analysis, that patients with baseline liver stiffness values > 24 kPa and between 12–24 kPa held higher risk of HCC development (HR 24.6, 95% CI 2.7–220.4, p = 0.004 and HR 11.7, 95% CI 1.3–105.2, p = 0.028) in comparison to those with LSM < 12 kPa. Furthermore, during follow-up LSM values persistently > 12 kPa carried a higher risk to develop HCC compared with those with LSM < 12 kPa.

In the study by Nakagomi et al. [28] baseline liver stiffness predicted the cumulative risk of HCC occurrence until 10 years. In detail, at enrollment, values of LSM ≤ 5 kPa, 5.1–10 kPa, 10.1–15 kPa, 15.1–20 kPa, 20.1–25 kPa, and > 25 kPa were associated with a progressively higher risk of 1.8%, 4.8%, 7.3%, 11.9%, 16.9%, and 22.7% of developing HCC at 1, 2, 3, 5, 7, and 10 years, respectively (Table 2).

HCV: studies assessing LSM with MRE

Promising results come from the application of MRE in the prediction of HCC occurrence.

In the study by Kumada et al. [39], a total of 537 viremic patients with HCV were evaluated at baseline with MRE. During follow-up, the patients received DAA treatment obtaining SVR. During the median follow-up of 43.8 months, those patients with baseline MRE value > 4.5 kPa (adjusted HR 7.301) showed higher rates of HCC occurrence and MRE resulted to be independently associated with HCC development at multivariate analysis.

Conversely, Tamaki et al. [40] evaluated 346 patients with SVR in prediction of HCC occurrence. In their study, MRE was compared to other NITs such as FIB-4. LSM ≥ 3.75 kPa was shown to be an independent predictive factor for HCC occurrence (HR 3.51, 95% CI 1.24–9.99) at multivariate analysis. In the patients with LSM ≥ 3.75 kPa the development of HCC was 6.6%, 11.9%, and 14.5% at 1/2/3‐year, respectively. The AUROC analysis for LSM resulted in 0.661, 0.724, and 0.743 at 1, 2 and 3 years, respectively.

In a further study, Kumada et al. [41] evaluated patients with HCV-related chronic liver disease at baseline (viremic) with MRE and SWE (LOGIQ S8 GE Healthcare) in the prediction of HCC occurrence. Their study found a good accuracy of both methods in the prediction of HCC at 5 years of follow-up. More in detail, LSM-MRE > 4.5 kPa showed a time-dependent AUROC 0.808 and LSM-SWE > 11.7 kPa showed a time-dependent AUROC 0.826 for the development of HCC (Table 3).

HCV: studies assessing LSM with SWE

The study by Gyotoku et al. [42] assessed shear-wave velocity (LOGIQ E9 multi-purpose GE Healthcare) at baseline, at the end of DAA therapy, at 12 and 24 weeks after treatment for purposes of evaluating its diagnostic accuracy in HCC prediction. Their study showed a greater accuracy of shear wave velocity at 24 weeks with AUROC 0.86 as compared to baseline (AUROC 0.80), end of treatment (AUROC 0.74), and at 12 weeks after SVR (AUROC 0.72). The authors found that the patients with HCC had baseline SWE values higher than those who did not develop primary liver cancer: 1.86 ± 0.20 m/s (1.52–2.26) vs. 1.58 ± 0.26 m/s (1.03–2.56) (p = 0.0036).

The study by Hamada et al. [43] evaluated the diagnostic accuracy of LSM-SWE (SuperSonic Imagine S.A., France) in the prediction of HCC development in HCV-SVR patients. This study showed an AUROC 0.93 with an optimal LSM cut-off value ≥ 11 kPa (Sens 75%, Spec 95%) (Table 4).

HCV: studies assessing SSM with VCTE

Dajti et al. [44] evaluated the prediction of LSM and SSM for the diagnosis of HCC in patients with SVR. They found in univariate analysis that SSM SVR-24 was significantly associated with HCC development (HR 1.029, 95% CI 1.006–1.053, p = 0.013). At adjusted multivariate analysis, SSM SVR-24 showed HR 1.025 (95% CI 1.001–1.050, p = 0.049).

The SSM cut-off value defined in this study was 42 kPa (Sens 75%, Spec 61%, NPV 93.6%, PPV 24.2%) (Table 5).

Mixed etiology: studies assessing LSM with VCTE

The studies reported below include populations with different etiologies of liver disease, this fact has to be considered in the evaluation of liver stiffness diagnostic performance in prediction of HCC occurrence.

The study by Kuo et al. [45] evaluated 435 patients with HCV/HBV infection at different stages of liver disease. The AUROC of LSM in predicting HCC presence was 0.736. At multivariate analysis, liver stiffness proved to be an independent factor for HCC presence (odds ratio 1.07, 95% CI 1.05–1.09). LSM diagnostic performance was further stratified according to stiffness values < 12 kPa, 12–24 kPa, and > 24 kPa. In the prediction of HCC occurrence, LSM > 24 kPa showed 41.5% sensitivity and a higher level of specificity at 92.7%, as compared to the other cut-off values.

Many studies compared LSM and HVPG in the prediction of clinically significant liver-related events as a composite outcome.

In the study by Robic et al. [46], LSM proved to be non-inferior to HVPG in the prediction of liver-related events (including HCC). Therefore, LSM could be considered a worthy alternative to invasive HVPG in the prediction of liver complications. In support of this, Pérez-Latorre et al. [47] have demonstrated the non-inferiority of VCTE (AUROC 0.85, 95% CI 0.73–0.97) compared to HVPG (AUROC 0.76, 95% CI 0.63–0.89) in the prediction of liver-related events (including HCC), in cirrhotic patients with HCV-HIV co-infection (Table 6).

Mixed etiology: studies assessing LSM with MRE

Lee et al. [54], evaluated 217 patients with various causes of liver disease at risk for HCC occurrence. At multivariate analysis, the authors found that LSM values obtained by MRE were predictive of HCC occurrence (p < 0.001, HR = 1.59 per unit, 95% CI 1.25–2.03). In addition, they identified LSM 4.4 kPa as a cut-off value able to predict the occurrence of HCC.

In the study by Ichikawa et al. [55] the predictive value of LSM-MRE was stratified based on three levels: LSM < 3 kPa, 3–4.7 kPa, or > 4.7 kPa. Patients have undergone MRE examinations at baseline and during follow-up. The authors subclassified patients into three groups according to sequential changes in liver stiffness: Group A (high LSM on the first MRE), Group C (low on both examinations), Group B (other combinations). This study showed that patients of Group A had a risk ratio of 1.018–6.030 (p = 0.0028–0.0268) for the development of HCC.

On the other hand, the previous case-control study by Anaparthy et al. [56] investigated the difference in LSM-MRE between cACLD patients with or without HCC. This study found no significant difference (p = 0.7) in LSM-MRE between individuals with HCC (6.3 ± 2.1 kPa) and patients without (6.1 ± 2.3 kPa) (Table 7).

Mixed etiology: studies assessing LSM with SWE

Kasai et al. [59] evaluated the predictive value of LSM-SWE (Aixplorer US system SuperSonic Imagine S.A.) for HCC in 273 patients with various etiologies of liver disease.

This case-control study found a significant difference in LSM value (22.65 ± 10.19 kPa) of patients with HCC and those without (12.67 ± 9.45 kPa) (p < 0.0001) (Table 8).

NAFLD: studies assessing LSM with VCTE

Nowadays non-alcoholic fatty liver disease (NAFLD) represents the most frequent cause of liver disease worldwide with an estimated prevalence in the USA and Europe between 10% and 30% [60]. Many clinical conditions are associated with NAFLD: obesity, hyperlipidemia, hypertension, type-2 diabetes, and metabolic syndrome. NAFLD is defined as the presence of steatosis in more than 5% of hepatocytes. The histological spectrum varies and includes non-alcoholic fatty liver, steatohepatitis (NASH; steatosis with inflammation and hepatocyte ballooning) and advanced fibrosis which ultimately evolves in cirrhosis [61]. Patients whose liver disease progresses through stages are at major risk of liver-related complications, such as HCC development.

In the study by Shili-Masmoudi et al. [62], including 2,251 patients, LSM was showed to be at multivariate analysis an independent predictor of overall survival with an adjusted HR 2.85 (1.65–4.92), p = 0.0002. Moreover, patients with elevated baseline LSM had a major risk to develop cardiovascular and liver events. The study also showed that patients with HCC occurrence during follow-up had a higher level of baseline LSM, but no further data was provided.

The study by Lee et al. [63], including 2,666 patients, showed the high accuracy of a composite model including age, LSM, platelets (PLT) together with aspartate transaminase (AST) ≥ 34 IU/L in the prediction of HCC occurrence, with AUROC at 2, 3, and 5 years of 0.948, 0.947, and 0.939, respectively (Table 9).

NAFLD: studies assessing LSM with MRE

Ajmera et al. [64] evaluated 6 studies assessing LSM-MRE and the prediction of liver-related events and HCC occurrence. In this study, LSM-MRE ≥ 5kPa was found to be associated with a greater than 1.5% risk of HCC development per year thus supporting the use of this cut-off for HCC surveillance. Moreover, this study stratified patients for the risk of HCC development at 1 and 3 year/s based on the value of LSM-MRE values: < 5 kPa, 5–8 kPa, > 8 kPa, and found a risk of 0.1%–0.35%, 3.71%–5.25% and 3.61%–5.66%, respectively. The HRs for HCC development were 23.4 (95% CI 6.85–79.8, p < 0.001) for LSM > 5kPa and 33.8 (95% CI 8.94–127.7, p < 0.001) for LSM 5–8 kPa, compared to LSM < 5 kPa (Table 10).

Studies assessing LSM and SSM in the prediction of HCC recurrence after resection

HCC recurrence after curative treatment represents a strong predictor of poor prognosis. Thus, stratification of patients at risk for this condition, potentially improves outcomes [65].

Many studies have evaluated liver and SSM in the prediction of HCC recurrence after treatment.

The study by Marasco et al. [66] evaluated the diagnostic accuracy of LSM and SSM with VCTE in the prediction of HCC recurrence after resection in patients with chronic liver disease. They defined early recurrence and late recurrence considering HCC onset < 24 months from surgery and > 24 months, respectively. This study found that SSM > 70 kPa was associated with a higher recurrence rate with a statistical significance at multivariate analysis. Thus, the authors proposed a cut-off value of SSM > 70 kPa.

The wide use of MR during follow-up in patients with previous HCC to detect tumor reappearance, have led many clinicians to investigate the role of LSM-MR in the prediction of HCC recurrence. Cho et al. [67] evaluated 192 patients with HCC who underwent hepatic resection, radio-frequency ablation (RFA), or trans-arterial chemoembolization (TACE). Interestingly, this study found that higher values of LSM represented an independent risk factor for early tumor recurrence both in patients treated with RFA/resection and TACE. In the first group, a cut-off of LSM-MR > 4.5 kPa showed an HR = 2.95; 95% CI 1.26–6.94; p = 0.013; in the second group LSM-MR > 6 kPa they found an HR = 2.94; 95% CI 1.27–6.83; p = 0.012 (Table 11).