Advice

Clinical and technical evidence

Clinical and technical evidence

A literature search was carried out for this briefing in accordance with the interim process and methods statement for medtech innovation briefings. This briefing includes the most relevant or best available published evidence relating to the clinical effectiveness of the technology. Further information about how the evidence for this briefing was selected is available on request by contacting mibs@nice.org.uk.

Published evidence

Six studies are summarised in this briefing including a total of 3,144 people with liver disease. These included 5 prospective cohort studies and 1 retrospective study. The clinical evidence and its strengths and limitations are summarised in the overall assessment of the evidence.

There is also an additional abstract (de Lédinghen et al. 2021a) and full-text paper (Aravind et al. 2020) that were not summarised for brevity. Aravind describes the development of LIVERFASt and evaluates its performance in 13,068 people with suspected fatty liver disease in Asia. It showed LIVERFASt scores were highly correlated with FibroTest (BioPredictive). But its accuracy was not evaluated against established diagnosis or liver biopsy as references.

Overall assessment of the evidence

The evidence in this briefing included 1 full-text peer-reviewed paper and 5 abstracts or posters. Abstracts were limited in detail on methods and patient demographics. Many contained detailed findings but often lacked discussion. The evidence would be strengthened by more full-text peer-reviewed papers to better explain findings and their implications for clinical practice.

Most studies were prospective cohort studies comparing LIVERFASt to other non- or minimally invasive tests. Several studies used comparators that were relevant to standard care in the NHS, including transient elastography and Fibrosis‑4 (FIB‑4). Liver biopsy was also used as a reference in 3 studies. Studies had within-group designs, with tests done at the same time. This meant participants acted as their own controls, and reduced confounding factors such as time and between-group differences. Although no studies were done in the UK, the populations were reflective of NHS patient groups and included non-alcoholic fatty liver disease (NAFLD) and hepatitis B.

The evidence suggests that LIVERFASt performs as well as transient elastography and better than FIB‑4 in detecting liver disease. It may also improve detection of severe disease when used with other tests. But there is no evidence on the resource or cost benefits of using LIVERFASt. The evidence base would therefore benefit from prospective studies on the cost effectiveness or resource consequences of using LIVERFASt in the NHS compared with existing tests in primary and secondary care.

Tangvoraphonkchai et al. (2022)

Study size, design and location

Prospective cohort study comparing non- or minimally invasive tests in 192 people with chronic liver disease from a tertiary liver centre in Thailand. Overall, 26% of people had NAFLD, 24% chronic hepatitis B and 50% chronic hepatitis C. Also, 50% had advanced fibrosis and 10% had cirrhosis.

Intervention and comparators

LIVERFASt-fibrosis compared with transient elastography (FibroScan), enhanced liver fibrosis test and FIB‑4.

LIVERFASt-steatosis compared with vibration control transient elastography (controlled attenuation parameter).

Liver biopsy was used as a reference.

Key outcomes

Mean LIVERFASt scores (ranging 0 to 1) were 0.29 for fibrosis, 0.29 for inflammatory activity and 0.35 for steatosis. Area under the receiver operating characteristics curve (AUC-ROC) for advanced fibrosis (n=157) was FibroScan 0.79 (95% confidence interval [CI] 0.71 to 0.85), LIVERFASt-fibrosis 0.77 (95% CI 0.61 to 0.77), FIB‑4 0.69 (95% CI 0.61 to 0.77) and enhanced liver fibrosis test 0.63 (95% CI 0.53 to 0.71). AUC-ROC was similar for bridging fibrosis.

LIVERFASt-steatosis and controlled attenuation parameter AUC‑ROC were 0.75 (95% CI 0.59 to 0.85) and 0.81 (95% CI 0.67 to 0.89) for mild steatosis, 0.75 (95% CI 0.61 to 0.85) and 0.81 (95% CI 0.68 to 0.89) for moderate steatosis and 0.70 (95% CI 0.47 to 0.83) and 0.74 (95% CI 0.55 to 0.86) for marked steatosis, respectively. There was no significant difference between LIVERFASt performance and FibroScan or controlled attenuation parameter.

Strengths and limitations

The study was reported in an abstract and therefore lacked detailed methods, patient demographics and findings. It included people who had liver biopsy that was used to determine performance of each test. LIVERFASt was compared with other tests that are reflective of standard care in the NHS. The findings are therefore relevant but limited by brevity.

Decraecker et al. (2022)

Intervention

Repeated LIVERFASt-fibrosis testing along with body mass index (BMI) and liver enzymes alanine transaminase (ALT). The abstract did not report what treatment people had over the course of the study. Significant improvement in clinical endpoints was defined as a decrease in BMI of 10% or more and ALT of 50% or more from baseline.

Key outcomes

At baseline, LIVERFASt-fibrosis found 45% of people had no fibrosis (stage F0), 29% had minimal fibrosis (F1), 6% moderate (F2), 12% significant (F3) and 8% severe (F4). Overall, 13 people (3%) had liver fibrosis regression based on a decrease in LIVERFASt-fibrosis score of 0.15 or more from baseline. ALT regression of 50% or more from baseline was found in 109 people (27%) while 75 people (19%) had significant improvement in BMI. LIVERFASt-fibrosis half-stage liver fibrosis improvement was more likely with ALT regression of 50% or more from baseline (Cox–Mantel hazard ratio [HR] 3.47; 95% CI 1.08 to 11.19) than without (HR 0.29; 95% CI 0.09 to 0.93). Authors concluded LIVERFASt-fibrosis correlates with clinical endpoints and could be used for long-term monitoring.

Strengths and limitations

The study was reported in an abstract and poster only and was limited in detail. It did not report treatment after baseline or if LIVERFASt scores were used in clinical decision making. Data was collected prospectively for people with at least 1 repeated LIVERFASt-fibrosis test. Overall, 87 people had 7 repeated tests but the median number of tests was not reported. The study had a relatively large sample size but only 13 people (3%) showed half-stage liver fibrosis improvement based on a decrease in LIVERFASt score of 0.15 or more. This should be considered when interpreting findings comparing improvement on LIVERFASt scores with clinical endpoints.

Decraecker et al. (2021)

Intervention and comparators

LIVERFASt compared with FibroScan and FIB‑4. Everyone had a minimum follow-up of 1 year and at least 2 measurements from each test.

Key outcomes

Independent predictors of overall survival included LIVERFASt score (HR 2.5; 95% CI 1.91 to 3.27; p<0.001), FibroScan measurement (HR 1.62; 95% CI 1.41 to 1.86; p<0.001) and FIB‑4 value (HR 1.26; 95% CI 1.16 to 1.36; p<0.001). The risk of overall complications was also significantly associated with LIVERFASt-fibrosis score (odds ratio [OR] 2.04; 95% CI 1.78 to 2.34; p<0.001), FibroScan measurement (OR 1.92; 95% CI 1.67 to 2.21; p<0.001) and FIB‑4 value (OR 1.89; 95% CI 1.62 to 2.21; p<0.001).

Clinical models showed LIVERFASt, FibroScan and FIB‑4 all had good performance for predicting overall and liver-related mortalities and morbidities. Performance increased when tests were combined, with maximum accuracy achieved by combining any 2 tests and clinical parameters.

Strengths and limitations

Data was collected prospectively and consecutively from a single centre. A total of 4,378 people with a diagnosis of alcoholic fatty liver disease or NAFLD were screened. Of these, 265 did not have MAFLD and 1,494 were excluded because of a lack of data from tests. Despite this, the study had a large sample size and benefited from everyone having all tests. Comparators align with standard care tests for NAFLD although the study population was specific to MAFLD.

de Lédinghen et al. (2021b)

Study size, design and location

Prospective cohort study comparing liver fibrosis tools in 588 people with NAFLD in France. Of these, 51% had type 2 diabetes. This study included the same sample as de Lédinghen et al. (2020).

Intervention and comparators

LIVERFASt compared with FibroScan and FIB‑4, with liver biopsy as a reference.

Key outcomes

Overall, 45% of people had bridging fibrosis. Test cut-offs detected bridging fibrosis in 44% of people with FibroScan, 40% with LIVERFASt and 15% with FIB‑4. Discordance rate compared with liver biopsy was 32% for LIVERFASt, 24% for FibroScan and 17% for FIB‑4. This did not include non-applicable FibroScan values (n=67) or indeterminate FIB‑4 scores (n=280). Concordance rate with liver biopsy was 27% (n=160) for all tests and 61% (n=356) for at least 2 tests. Evaluation of sequential and combinatory models showed best performance for LIVERFASt plus FibroScan (sensitivity 75%, specificity 84%, positive predictive value 79%, negative predictive value 81%).

Strengths and limitations

The study was reported in a poster and had limited details on methods, patient demographics and findings. Liver biopsy was used as a reference and comparators included non- and minimally invasive tests that are used in NHS standard care. The findings are therefore relevant to current clinical practice. The study also explored the use of tests in combination and suggested that using LIVERFASt with standard care tests could improve detection of severe NAFLD. More evidence on this is needed to show the value to the NHS.

de Lédinghen et al. (2020)

Study size, design and location

Retrospective study comparing performance of tests for severe fibrosis in 583 people with NAFLD in France. This study included the same sample as de Lédinghen et al. (2021b).

Intervention and comparators

LIVERFASt compared with FibroScan, Fibrosure FibroTest, Hepascore, FIB‑4, aspartate aminotransferase (AST) to platelet ratio index (APRI) and Forns index for cirrhosis, with liver biopsy as a reference.

Key outcomes

LIVERFASt‑fibrosis AUC‑ROC was 0.72 (95% CI 0.68 to 0.76) for severe fibrosis. Intention-to-diagnose analysis reported LIVERFASt AUC‑ROC for cirrhosis as 0.80 (95% CI 0.75 to 0.84), with no significant difference from Fibrosure (0.81; 95% CI 0.75 to 0.85), Hepascore (0.77; 95% CI 0.71 to 0.82), Forns (0.78; 95% CI 0.73 to 0.83), FibroScan (0.75; 95% CI 0.68 to 0.80) and FIB‑4 (0.76; 95% 0.69 to 0.81). It performed better than APRI (0.65; 95% CI 0.59 to 0.71; p<0.001). There was no difference in LIVERFASt performance for severe fibrosis or cirrhosis in people with or without type 2 diabetes.

Strengths and limitations

The study was reported in an abstract and poster with limited details on methods, patient demographics and discussion of findings. Liver biopsy was used as a reference and comparators included non- and minimally invasive tests that are used in NHS standard care. The findings are therefore relevant to current clinical practice. Some study authors were affiliated with the company.

Lim et al. (2020)

Intervention and comparators

LIVERFASt compared with acoustic radiation force impulse, FIB‑4 and APRI.

Key outcomes

Spearman correlations of LIVERFASt-fibrosis and LIVERFASt-activity (respectively) with other tests were acoustic radiation force impulse 0.18 and 0.21, FIB‑4 0.47 and 0.11 and APRI 0.23 and 0.69 (all p<0.01). No or minimal liver disease was discriminated from moderate or severe disease by LIVERFASt-fibrosis, LIVERFASt-activity and APRI but not FIB‑4 and acoustic radiation force impulse. The negative predictive value (NPV) for people who were hepatitis B e antigen (HBeAg) negative was 77% for LIVERFASt-fibrosis, 82% for LIVERFASt-activity and 27% for APRI. NPV for people who were HBeAg positive was 89% for LIVERFASt-fibrosis, 62% for LIVERFASt-activity and 12% for APRI.

Strengths and limitations

The study was reported in an abstract and had limited methods, patient demographics and discussion. Findings reported negative predictive value and 3 cases of false positive, but overall positive predictive value was not stated. The abstract presented quite detailed results but would be strengthened by more discussion particularly of their relevance to clinical practice and outcomes. Two study authors were affiliated with the company.

Sustainability

The company claims LIVERFASt may result in the use of less resources associated with hospitalisation for liver biopsy. There is no published evidence to support these claims.

Recent and ongoing studies

Clinical performance of LIVERFASt compared with liver biopsy in people with NAFLD: a prospective single-arm trial. ClinicalTrials.gov identifier: NCT04579874. Status: completed (July 2022). Devices: LIVERFASt. Last updated: August 2022. Country: US.