Non-invasive predictors of prognosis in chronic liver disease

<p dir="ltr">If the liver is exposed to continuous damage over time, accumulation of liver fibrosis occurs in all chronic liver diseases. The amount of fibrosis is traditionally evaluated via liver biopsy and is most often staged as fibrosis stage 0-4, where stage 4 is termed liver cirrhosis, and is the end-stage of chronic liver diseases, regardless of the etiology. This is considered an irreversible condition and is associated with an increased morbidity and mortality, including an increased risk of hepatocellular carcinoma (HCC), and biannual screening with ultrasound scans is recommended to identify smaller tumors eligible for curative treatment.</p><p dir="ltr">Fibrosis stage by liver biopsy is the marker that best correlates with progression to cirrhosis, decompensated cirrhosis, and hepatocellular carcinoma. However, a liver biopsy is invasive and not an optimal method since it is resource-intensive, comes with a risk of complications, carries a risk of sampling error and interobserver variability, and is not suitable to screen large populations with, and the prevalence of the most common chronic liver disease, metabolic dysfunction- associated steatotic liver disease (MASLD), is 38% globally. Non-invasive tests have the potential to be used instead of a liver biopsy. In this thesis, we aimed to investigate different non-invasive tests and their performance in chronic liver disease, with a special emphasis on MASLD. We investigated non-invasive tests both as screening tools to identify cases with a higher risk of progression, and their performance to predict important outcomes such as progression to cirrhosis, decompensated cirrhosis, HCC, liver transplantation, and liver-related death.</p><p dir="ltr">In study I, a case-control study, cases were patients with cirrhosis who had died from HCC between 2004 and 2020 in Stockholm County. They were matched to controls, patients with cirrhosis who had not died from HCC, based on age, sex, etiology of cirrhosis, year of cirrhosis diagnosis, hospital of cirrhosis-diagnosis, MELD score, and Child-Pugh Score. Using an adjusted logistic regression model, we investigated the exposure to HCC-surveillance between cases and controls. 72 cases and 72 controls were matched, and we found that HCC surveillance was associated with decreased HCC-related mortality, with an odds ratio (OR) of 0.37 (95%CI=0.16-0.88). The analysis showed that controls who did not die from HCC were more exposed to screening ultrasounds compared to cases, suggesting a protective effect.</p><p dir="ltr">In study II, we collected liver stiffness measurements (LSM) by vibration- controlled transient elastography (VCTE) from 16 hospitals in Sweden. After linkage to nationwide healthcare registers, we created a cohort of 14,000 patients where disease etiology, comorbidities and outcomes were based on data from the registers. We used a Cox regression model to investigate if LSM by VCTE could be used to predict progression to an outcome associated with portal hypertension or hepatocellular carcinoma. A total of 402 (2.7%) patients developed an outcome associated with portal hypertension, and the risk of an outcome was 48-fold higher in patients with an LSM >25 kPa compared to patients with an LSM <10 kPa, HR 48.3 (95%CI=37.6-62.0). We also showed that an increased LSM by VCTE was associated with and increased rate of progression to HCC. A total of 220 patients (1.5%) developed hepatocellular carcinoma and patients with an LSM >25 kPa had a 34-fold increased rate of hepatocellular carcinoma compared to patients with an LSM <10 kPa, HR 34.0 (95%CI=23.9-48.4).</p><p dir="ltr">In study III, a cross-sectional study with MASLD-patients, we compared the ability of the non-invasive tests FIB-4 (a score based on ALT, AST, platelet count, and age) and ADAPT (a score based on age, diabetes, the collagen formation biomarker PRO-C3, and platelet count) to detect patients with an LSM >8 kPa or >12 kPa. We showed that ADAPT had modestly better discriminative ability than FIB-4 in detecting an LSM >8 kPa in patients with MASLD, and a higher net benefit among across the probability threshold range of 15% - 40%, which means that ADAPT has a higher net benefit as a first-line screening test of fibrosis compared to FIB-4, if we are willing to do 100 LSM by VCTE to find 15-40 patients with an LSM >8 kPa.</p><p dir="ltr">In study IV, we used longitudinal data from patients with MASLD to investigate how the non-invasive tests LSM by VCTE, FIB-4 and CAP changed over three years. We compared changes in mean values over time, both in the full population and between subgroups where the exposure to disease modifiers such as weight loss and alcohol consumption differed. We also described fluctuations between established categories of LSM and FIB-4, aimed to describe the risk of advanced fibrosis, over time. We could demonstrate that the mean change over time in LSM by VCTE (-0.13 kPa/year, 95%CI =- 0.39 to +0.13) and FIB-4 (0.04/year, 95%CI=0.01-0.06) was negligible, but there were considerable intraindividual fluctuations over time where 55 patients (29%) changed LSM category (<8 kPa, 8- 12 kPa, >12 kPa) between baseline and one year, and 51 patients (27%) changed LSM category between one and three years of follow-up.</p><p dir="ltr">In conclusion, this thesis found that HCC surveillance in patients with cirrhosis is associated with a decreased HCC-related mortality, which suggests that patients with cirrhosis eligible for curative treatment should continue to be included in HCC-surveillance programs. We also showed that LSM by VCTE is a non-invasive biomarker suitable for risk stratification in patients with chronic liver disease. ADAPT performs modestly better than FIB-4 when used as a first-line test to screen for patients with an LSM >8 kPa, however the added value of ADAPT as a first-line test is unknown and warrants further evaluation. Mean LSM by VCTE and FIB-4 does not change much over three years on a group level, but disease monitoring with non-invasive tests remains challenging with intraindividual fluctuations over time. Further research is needed to find better ways of disease monitoring, where true progressors can be separated from non-progressors with fluctuations due to measurement errors of non-invasive tests.</p><h3>List of scientific papers</h3><p dir="ltr">I. <b>Hegmar H,</b> Bengtsson B, Ullman Nilsson S, Rowe I, Stål P, Hagström H. HCC-surveillance is associated with improved HCC-related mortality in patients with cirrhosis - a case-control study. Scand J Gastroenterol. 2025 Jun 17:1-11. doi: 10.1080/00365521.2025.2517208. Online ahead of print.<br><a href="https://doi.org/10.1080/00365521.2025.2517208">https://doi.org/10.1080/00365521.2025.2517208<br></a><br></p><p dir="ltr">II. <b>Hegmar H,</b> Wester A, Aleman S, Backman J, Degerman E, Ekvall H, Lund K, Lundgren Å, Nasr P, Shahnavaz A, Vessby J, Westin J, Önnerhag K, Hagström H. Liver stiffness predicts progression to liver-related events in patients with chronic liver disease - A cohort study of 14 414 patients. Liver Int. 2024 Jul;44(7):1689-1699. <a href="https://doi.org/10.1111/liv.15919" rel="noreferrer" target="_blank">https://doi.org/10.1111/liv.15919</a></p><p dir="ltr">III. <b>Hegmar H,</b> Wiggers T, Nasr P, Vessby J, Kechagias S, Nyhlin N, Marschall HU, Danielsson Borssen Å, Strandberg R, Karsdal M, Julie Leeming D, Ekstedt M, Hagström H. Performance of novel collagen turnover biomarkers to detect increased liver stiffness in MASLD. J Intern Med. 2024 Aug;296(2):177-186. <a href="https://doi.org/10.1111/joim.13813" rel="noreferrer" target="_blank">https://doi.org/10.1111/joim.13813</a></p><p dir="ltr">IV. <b>Hegmar H,</b> Strandberg R, Shang Y, Vessby J, Danielsson Borssén Å, Nyhlin N, Kechagias S, Stål P, SWEHEP study group, Ekstedt M, Hagström H. Dynamics of non-invasive tests in metabolic dysfunction-associated steatotic liver disease. [Manuscript]</p>