Prognostic biomarkers in and selection of surgical patients with hepatocellular carcinoma.

Abstract

The Faculty of Health and Medical Sciences at the University of Copenhagen has accepted this dissertation for public defence for the doctoral degree in medicine, Copenhagen, January 10, 2023 Bente Merete Stallknecht, Head of Faculty Assessment committee Chair Professor Susanne Dam Poulsen, University of Copenhagen Opponents Professor Pål-Dag Line, University of Oslo Professor Christian Toso, University of Geneva Chair of the defence ceremony Professor Kirsten Grønbæk, University of Copenhagen The defense took place on February 14, 2023 at 14:00 CET at Rigshospitalet, Auditorium 2, Blegdamsvej 9, 2100 Copenhagen Ø This thesis summarizes a significant part of the research that I carried out from 2015 to 2020 alongside my surgical specialization. Since defending my PhD in 2013, my research area of interest has shifted towards liver surgery and transplantation, in particular prognostic aspects related to malignancy. The complexity of liver surgery, including anatomy, technique and medical aspects has always sparked my interest. The work in this thesis has paved the way for the current and future area of interest of my research group at Department of Surgery and Transplantation, Rigshospitalet. I would like to express my deepest appreciation to Allan Rasmussen for inspiration and guidance. Allan introduced me to this specific field of research and to many important research partners in Rigshospitalet, Denmark and Scandinavia. I would like to thank Jens Hillingsø and all of my wonderful colleagues at Department of Surgery and Transplantation, Rigshospitalet for their support and for offering me the opportunity to carry out my research. I would also like to thank the collaborators on the studies for their excellent work and important scientific inputs. These include: Andreas Rostved, Carina Lund Sørensen, Peter Nørgaard Larsen and Nicolai Schultz, Department of Surgery and Transplantation, Rigshospitalet; Jakob Burcharth and Jacob Rosenberg, Department of Surgery, Herlev Hospital; Jane Preuss Hasselby and Gro Linno Willemoe, Department of Pathology, Rigshospitalet; Jan-Michael Kugler, Institute for Molecular and Cellular Medicine, Panum Institute; Christina Westmose Yde, Lise Ahlborn and Finn Cillius Nielsen, Center for Genomic Medicine, Rigshospitalet; Claus Lindbjerg Andersen and Tenna Henriksen, Department for Molecular Medicine (MOMA), Aarhus University Hospital/Skejby; Kristoffer Rohrberg, Phase 1 Unit, Department of Oncology, Rigshospitalet; Lau Caspar Thygesen, National Institute of Public Health, University of Southern Denmark; and the European Liver and Intestine Transplant Association (ELITA) study group. Thanks to my research group for daily inspiration, good discussions and laughs (Delal Akdag, Jeanett Klubien, Lucas Knøfler og Sophie Nordkild). This work was financially supported by the Harboe foundation, the Risom foundation and the Dagmar Marchall foundation, for which I am grateful. Finally, I would like to thank my wife Line for her tremendous support over the many years where my research mostly has been conducted in my spare time. Thanks to my three children for filling me with love each day and thanks to my mother and my brother for continuous support. A special thanks to my father who passed away in the fall. He was always proud, supportive and encouraging. Hans-Christian PommergaardMD, DMSc, PhDDepartment of Surgery and TransplantationRigshospitalet, University of Copenhagen Charlottenlund, January 2023 Pommergaard HC, Burcharth J, Rosenberg J, Rasmussen A. Serologic and molecular biomarkers for recurrence of hepatocellular carcinoma after liver transplantation: a systematic review and meta-analysis. Transplant Rev (Orlando). 2016 Jul;30(3):171–7. Pommergaard HC, Hasselby JP, Willemoe GL, Ralbovska A, Rostved AA, Rasmussen A, Schultz NA, Hillingsø J, Larsen PN, Kugler JM. Peroxisome proliferator-activated receptor activity correlates with poor survival in patients resected for hepatocellular carcinoma. J Hepatobiliary Pancreat Sci. 2021 Apr;28(4):327–35. Pommergaard HC, Rasmussen A, Hillingsø J, Kugler JM. Aldehyde dehydrogenase expression may be a prognostic biomarker and associated with liver cirrhosis in patients resected for hepatocellular carcinoma. Surg Oncol. 2021 Nov 23;40:101677. Pommergaard HC, Yde CW, Ahlborn LB, Andersen CL, Henriksen TV, Hasselby JP, Rostved AA, Sørensen CL, Rohrberg KS, Nielsen FC, Rasmussen A. Personalized circulating tumour DNA in patients with hepatocellular carcinoma: a pilot study. Mol Biol Rep. 2021 Nov 22. Pommergaard HC, Rostved AA, Adam R, Thygesen LC, Salizzoni M, Gómez Bravo MA, Cherqui D, Filipponi F, Boudjema K, Mazzaferro V, Soubrane O, García-Valdecasas JC, Prous JF, Pinna AD, O'Grady J, Karam V, Duvoux C, Rasmussen A. Vascular invasion and survival after liver transplantation for hepatocellular carcinoma: a study from the European Liver Transplant Registry. HPB (Oxford). 2018 Aug;20(8):768–75. Pommergaard HC, Rostved AA, Adam R, Thygesen LC, Salizzoni M, Gómez Bravo MA, Cherqui D, De Simone P, Boudjema K, Mazzaferro V, Soubrane O, García-Valdecasas JC, Fabregat Prous J, Pinna AD, O'Grady J, Karam V, Duvoux C, Rasmussen A. Locoregional treatments before liver transplantation for hepatocellular carcinoma: a study from the European Liver Transplant Registry. Transpl Int. 2018 May;31(5):531–9. Pommergaard HC, Rostved AA, Adam R, Rasmussen A, Salizzoni M, Gómez Bravo MA, Cherqui D, De Simone P, Houssel-Debry P, Mazzaferro V, Soubrane O, García-Valdecasas JC, Prous JF, Pinna AD, O'Grady J, Karam V, Duvoux C, Thygesen LC. Mortality after transplantation for hepatocellular carcinoma: a study from the European Liver Transplant Registry. Liver Cancer. 2020 Aug;9(4):455–67. Hepatocellular carcinoma (HCC) is the sixth most common cancer and the fourth leading cause of cancer deaths globally with increasing incidence. In most cases, HCC develops as a result of an underlying chronic liver disease. The stage at the time of diagnosis is closely correlated with prognosis with a 5-year survival of more than 70% for patients with early-stage HCC. These patients are generally suitable for curative treatment, such as resection, ablation or liver transplantation. In contrast, 5-year survival may be as low as 5% for patients in advanced stages of the disease. Due to organ shortage and the risk of mortality for patients on the waiting list for liver transplantation, survival in patients undergoing transplantation due to HCC should be comparable to those with benign indications. However, there is an ongoing debate concerning the best criteria for the selection of patients with HCC for transplantation, as the current criteria may rule out patients with a good prognosis. Liver resection may be performed in patients with HCC in non-cirrhotic livers and cirrhotic livers with preserved liver function. However, careful selection of patients is crucial due to high recurrence risk (>50%) and severe complications in >10% of liver resections for HCC. The aims of this thesis were to identify prognostic biomarkers in and improve the selection of patients undergoing transplantation and resection for HCC, respectively. Yet another aim was to improve upon the existing early detection of recurrence. The seven studies included in this thesis provided novel insights into prognostic biomarkers and selection of surgical patients with HCC. Firstly, the literature was reviewed regarding biomarkers to detect recurrence after transplantation for HCC. Next, two novel tumour-related prognostic biomarkers were identified and the applicability of circulating tumour DNA (ctDNA) was evaluated. Moreover, using data from the European Liver Transplant Registry (ELTR), vascular invasion and locoregional treatment were identified as important prognostic markers. Lastly, using propensity score calibration, prognostic differences between HCC in cirrhotic and non-cirrhotic livers were clarified. The studies included in this thesis add important knowledge towards better selection of patients for transplantation and resection of HCC, respectively. In future guidelines, alpha-fetoprotein, radiological markers of vascular invasion and response to locoregional treatment may be included in selection criteria for transplantation of patients with HCC. In addition, the novel biomarkers identified may add important prognostic information and improve the selection of patients. For resected patients, a proper balance between survival benefit and risk of recurrence and complications may be ensured. For transplanted patients, optimized selection means that the utility of donor organs is maximized by more accurately selecting patients with favourable tumour biologies. Hepatocellulært karcinom (HCC) er den sjette mest almindelige kræftsygdom og den fjerde hyppigste årsag til kræftdødsfald globalt med stigende forekomst. I de fleste tilfælde udvikler HCC sig som følge af en underliggende kronisk leversygdom. Stadie på diagnosetidspunktet er tæt korreleret med prognose med femårs overlevelse på mere end 70% hos patienter med HCC i tidligt stadium. Disse patienter kan generelt modtage kurativt intenderet behandling, såsom resektion, ablation eller levertransplantation. I modsætning hertil kan femårs overlevelse være helt ned til 5% hos patienter med fremskredne stadier. På grund af organmangel og død på venteliste til levertransplantation bør overlevelse hos patienter der transplanteres for HCC være sammenlignelig med overlevelse for benigne indikationer. Der er aktuelt debat om de bedste kriterier til at selektere patienter med HCC til transplantation, da nuværende kriterier udelukker visse patienter med en god prognose. Leverresektion kan udføres hos patienter med HCC i en ikke-cirrotisk lever eller i en cirrotisk lever med bevaret leverfunktion. Imidlertid er omhyggelig udvælgelse af patienter afgørende på grund af en høj risiko for recidiv (>50%) og alvorlige komplikationer i >10% af leverresektioner for HCC. Formålet med denne disputats var at identificere prognostiske biomarkører og forbedre selektion af patienter til henholdsvis transplantation og resektion for HCC. Desuden var målet at forbedre tidlig identifikation af recidiv. De syv studier inkluderet i denne disputats giver ny indsigt i prognostiske biomarkører og selektion af kirurgiske patienter med HCC. Eksisterende biomarkører for recidiv efter transplantation for HCC blev systematisk undersøgt. Dernæst blev to nye tumorrelaterede prognostiske biomarkører identificeret, og anvendeligheden af cirkulerende tumor-DNA (ctDNA) blev evalueret. Ved hjælp af data fra European Liver Transplant Registry (ELTR) blev mikrovaskulær invasion og lokalbehandling af HCC identificeret som vigtige prognostiske markører. Endelig blev prognostiske forskelle mellem HCC i cirrotisk og ikke-cirrotisk lever afklaret ved hjælp af propensity score kalibrering. Studierne inkluderet i denne disputats tilfører vigtig viden til bedre selektion af patienter til henholdsvis transplantation og resektion af HCC. I fremtidige retningslinjer kan alfa-føto protein, radiologiske markører for microvaskulær invasion og respons på lokalbehandling muligvis inkluderes i selektionskriterier for transplantation af HCC. Derudover kan de nye identificerede biomarkører tilføje vigtig prognostisk information og forbedre udvælgelsen af patienter. For resecerede patienter kan der skabes en balance mellem overlevelsesgevinst og risiko for recidiv og komplikationer. For transplanterede patienter betyder optimeret selektion, at udnyttelse af donororganer maksimeres ved mere præcis udvælgelse patienter med en gunstig tumorbiologi. Hepatocellular carcinoma (HCC) is the sixth most common cancer and the fourth leading cause of cancer deaths globally with increasing incidence. HCC represents 75–85% of primary liver cancers and has a strong male predominance, both regarding incidence and mortality [1, 2]. In 70–90% of cases, HCC arises in patients with cirrhotic livers [3, 4]. Approximately one-third of patients with cirrhosis will develop HCC in their lifetime (1–8% annual incidence) [5, 6]. On a global level, hepatitis B virus (HBV) is the primary cause of HCC (33%), followed by alcohol (30%), hepatitis C virus (HCV) (21%) and other aetiologies (16%) [2]. Most deaths due to HCC occur in Asia, and the primary cause is endemic HBV infections. However, the incidence of HCC is rising in Western countries as a result of the increasing prevalence of non-alcoholic fatty liver disease (NAFLD), non-alcoholic steatohepatitis (NASH) and alcohol-related liver disease [7]. HCC seems to be less frequent in patients with alcoholic and NASH cirrhosis than in patients with viral hepatitis [6, 8]. However, the annual incidence of HCC in patients with NASH is 1–2%, and it is estimated that 6 million people in the United States alone have NASH [9] and that obesity and/or diabetes accounts for approximately 37% of the HCC cases [10]. Current guidelines suggest that patients with cirrhosis (Child-Pugh A/B and Child-Pugh C (awaiting transplantation)) and non-cirrhotic HBV patients with intermediate or high risk of HCC should undergo surveillance for HCC with abdominal ultrasound every 6 months [11]. This modality has a limited sensitivity (45%) for the detection of HCC [12], which may be reduced further by obesity and cirrhosis [13]. However, evidence suggests that surveillance improves the identification of HCC at an early stage and hence the chances of curative treatment and survival [14]. Patients with early-stage HCC are generally suitable for curative treatments, such as resection, ablation or liver transplantation. In contrast, patients with advanced-stage disease are typically only offered palliative oncologic treatment or the best supportive care. Patients with intermediate-stage disease may be offered chemo-embolization [15]. Consequently, the stage at the time of diagnosis is closely correlated with the prognosis with a 5-year survival of more than 70% for patients with early-stage HCC compared with only 5% for patients in advanced stages of the disease [11, 16]. In most centres, a multidisciplinary approach is adopted with treatment allocation based on the stage of the disease, using, for example, the Barcelona Clinic Liver Cancer (BCLC) classification [17]. A contemporary approach includes treatment stage migration, wherein a patient with a specific disease stage not suited for the first-line treatment may be allocated a treatment recommended for a different disease stage [18-20]. The concepts are integrated with the 2022 BCLC guideline update [15] (Fig. 1). With evidence suggesting that prognosis is related to first-line treatment irrespective of tumour stage, an alternative approach has been suggested [21]. In the concept of therapeutic hierarchy (Fig. 2), treatments are decided independent of stage and hierarchical use of available therapies according to their specific survival benefit [22]. Previously, sorafenib was the only available systemic treatment. It is associated with significant adverse events and only minor survival benefits [23]. Based on the evidence from phase III trials, several additional systemic therapies have been approved including atezolizumab plus bevacizumab, lenvatinib, regorafenib, cabozantinib and ramucirumab [24]. Moreover, trials are now evaluating combination therapies. According to the 2022 guideline update from BCLC, Atezolizumab-Bevacizumab/Durvalumab-Tremelimumab is now recommended as first-line systemic treatment [15] (Fig. 1). Hepatocellular carcinoma may be diagnosed solely based on imaging using the Liver Imaging Reporting and Data System (LI-RADS) [25], thereby eliminating the need for preoperative histological diagnosis by tumour biopsy. LI-RADS is valid in patients with high-risk of developing HCC, namely patients with cirrhosis, chronic HBV and current or prior HCC. However, LI-RADS may not be applied in patients without these risk factors, in patients under 18 years of age and in those with cirrhosis due to a vascular disorder or congenital hepatic fibrosis [25]. Moreover, a recent study showed that LI-RADS may be valid in a larger proportion of patients with an increased risk of HCC than previously considered [26]. However, biopsy in patients without increased risk, including patients with non-cirrhotic livers, is still needed to confirm diagnosis. Due to organ shortage and the risk of mortality for patients on waiting list for liver transplantation, survival in patients undergoing transplantation due to HCC should be comparable to those with benign indications [27]. However, there is an ongoing debate concerning the best criteria for selection of patients with HCC for transplantation. Some argue that the current criteria may rule out patients with a good prognosis [28]. Thus, better selection criteria are needed. Patients with HCC in a non-cirrhotic liver without macrovascular invasion or extra-hepatic spread may undergo curative liver resection. Liver resection may also be performed in patients with HCC in a cirrhotic liver with preserved liver function, no portal hypertension and sufficient volume of the future liver remnant (FLR) [11]. Careful selection of patients is essential due to a high recurrence risk (>50%) [29] and severe complications in >10% of liver resections for HCC. Age, model for end-stage liver disease (MELD) score, liver stiffness and serum albumin were independent predictors of complications [30], while portal hypertension, extend of resection, MELD score, bilirubin and indocyanine green (ICG) clearance were associated with the risk of postoperative liver failure [31-33]. However, knowledge of cancer-related prognostic biomarkers is lacking. To optimize the selection of patients who will benefit from a high-risk surgical procedure, prognostic biomarkers are needed. Furthermore, as liver disease may not be easily diagnosed in some patients, additional biomarkers to identify high-risk patients, for example, with fibrosis/cirrhosis may be useful [34]. However, so far, no prognostic biomarkers have been widely accepted in clinical practice. Early identification of recurrence after surgical treatment may be associated with better prognosis [29]. Imaging studies currently used for follow-up may be insufficient for smaller lesions (<2 cm) [35, 36]. Thus, tools for early detection of recurrence are needed. Hypothesis 1 was investigated by conducting a systematic review and a meta-analysis using the PubMed and Scopus databases on studies investigating biomarkers for recurrence of HCC after liver transplantation. Hypotheses 2 and 3 were investigated using tissue stored in the Danish Cancer Biobank from patients undergoing liver resection for HCC at Rigshospitalet, Denmark between 2014 and 2018. Hypothesis 4 was investigated by evaluating the detection of ctDNA in blood preoperatively in a prospective pilot study in patients undergoing liver resection for HCC at Rigshospitalet, Denmark in 2018 and 2019. Hypotheses 5, 6 and 7 were investigated using data from the European Liver Transplant Registry (ELTR) for the period from 1990 to 2016, where pre-, peri- and postoperative data from 174 liver transplantation centres in 28 European countries have been collected, including data on tumour characteristics and locoregional treatment in patients with HCC. Regarding liver transplantation, donor livers are scarce with substantial mortality of patients in waiting lists [37]. Thus, the prognosis for recipients transplanted for HCC should be comparable to that of benign indications to justify transplantation [27]. Current selection criteria, based on the size and number of HCC, are insufficient, as they may rule out patients with good prognosis despite being outside the established criteria [28]. Inclusion of biomarkers in the selection criteria may improve the selection of patients for liver transplantation for HCC. However, except for alpha-fetoprotein (AFP), no biomarkers are included in the existing selection criteria for transplantation due to HCC. Surgical resection for HCC may be a high-risk procedure, especially in patients with cirrhosis, with a substantial risk of recurrence [11] and complications, including liver failure [30]. Prognostic biomarkers may improve the selection of patients who will benefit from a surgery and also provide more precise preoperative information to them. Currently, tumour burden, liver function and performance score are used to select patients for surgical resection [15]. Thus, biomarkers may be used in conjunction with these criteria to determine the optimal treatment strategy. Currently, no biomarkers have been widely incorporated into clinical practice. A prognostic biomarker may be defined as a factor indicating the likelihood of a future recurrence, progression or death in a population [38]. For the purpose of this thesis, a prognostic biomarker was defined as a blood or tumour marker associated with prognosis not currently included in standard diagnostic workups. A vast amount of research has been conducted in an effort to identify clinically useful biomarkers in patients with HCC. However, relatively less effort has gone into the validation of positive results. In an attempt to gain an overview of the level of evidence, directions for the design of studies and the possibilities for clinical implementation, a systematic review and a meta-analysis were conducted in Study I. PubMed and Scopus databases were searched up to the year 2016 to identify observational studies investigating blood-borne or tumour markers for recurrence of HCC after liver transplantation. Meta-analyses were performed in biomarkers investigated in multiple studies with sufficient homogeneity. Meta-analysis was omitted when heterogeneity was above 75% as assessed by the I2 test. The generic inverse variance method was utilized to incorporate adjusted hazard ratio (HR) estimates from individual studies. This was applied to adjust for confounding in individual studies. Alpha-fetoprotein was investigated in 49 studies of which 17 studies were included in the meta-analysis. These studies were defined by using a pre-transplant AFP >400 NT/mL as cut-off [39-55]. The pooled estimate (HR) regarding risk of recurrence was 2.69 [2.08–3.47], with an acceptable heterogeneity of 46% (Fig. 3). However, the funnel plot indicated a risk of publication bias with a lack of smaller negative studies. Des-gamma-carboxy prothrombin (DCP/PIVKA) was investigated in seven studies. Four studies [45, 48, 57, 58] with comparable outcome definition and cut-off levels (PIVKA level >100–400 mAU/mL) were included in a meta-analysis with an HR of 5.99 [3.27–10.98] regarding the risk of recurrence. Heterogeneity of the four studies was 0% (Fig. 4). Neutrophil to lymphocyte ratio (NLR) was found to be associated with recurrence in four studies [59-62]. However, a meta-analysis was not conducted due to the high heterogeneity (>75%) between the studies. Moreover, circulating DNA – particularly single nucleotide polymorphisms rs894151 and rs12438080 – was found to be associated with recurrence in one of the studies [63]. Following protein markers in HCC tumours has been shown to be associated with recurrence: anti-Ki67 [64], calpain small subunit 1 (Capn4) [65], CD34 antigen [66], CD147 [67], NDRG1 [68], cadherins [64], x-linked inhibitor of apoptosis protein-associated factor 1 (XAF1) [69], hepatocyte nuclear factor 1 (HNF1) [70] and hippo cell signal pathway [71]. Enzymes were investigated in four studies. Two of them investigated matrix metalloproteinases (MMPs) measured in tumour [67, 72]. In a meta-analysis, there was no association with recurrence with an HR of 2.12 [0.21–21.43], with a high heterogeneity of 74% (Fig. 5). However, MMP measured in adjacent stromal tissue may be associated with recurrence [67]. In two additional studies, histone deacetylases (HDACs) [73] and O-GlcNAcylation [52] were associated with recurrence. Micro-RNA in tumour was associated with recurrence in four studies [41, 44, 54, 74]. Moreover, long noncoding RNAs, in particular, HOTAIR in tumour [50] and MALAT-1 [75], were associated with recurrence. DNA/genes associated with recurrence included CIMP [76], MACC1 [42, 51], and cell-cycle genes pRb and p16 [77]. Lastly, allelic imbalance in microsatellites was associated with recurrence in a meta-analysis with two studies and an HR of 13.49 [3.17–57.30] and 0% heterogeneity [78, 79] (Fig. 6). The majority of biomarkers were only investigated in a single study, and results have not been validated. The included studies considered different types of aetiology for HCC, and patients may therefore not be comparable. Moreover, outcome definition differed between the studies. For meta-analyses, there was an insufficient number of studies for meaningful evaluation of publication bias by means of funnel plots, except for AFP where a lack of smaller negative studies suggested a risk of publication bias. Since the publication of Study I in 2016, additional studies have been published on biomarkers for recurrence after liver transplantation for HCC. Thus, an updated literature search was performed. Regarding blood-borne biomarkers, platelet-to-lymphocyte ratio (PLR), controlling nutritional status (CONUT) and the prognostic nutritional index (PNI) were associated with recurrence [80]. Serum homocysteine was associated with recurrence in recipients with AFP-negative disease [81]. Moreover, markers of angiogenesis and inflammation (plasma VEGFR1, VEGF and VEGF-C) and circulating exosomal miR-92b were associated with recurrence [82, 83]. Regarding tumour tissue biomarkers, increased LGALS3 and LGALS3BP gene and protein expression were associated with recurrence [84]. In addition, TP53 mutations, high fractional allelic loss, hypomethylation of 8 tumour suppressors and absence of CTNNB1 mutations were of prognostic significance [85]. Lastly, tumour evolutionary distance by genome-wide single-nucleotide polymorphism genotyping was associated with recurrence and validated in a replication cohort [86]. As in liver transplant recipients, AFP is the most commonly evaluated biomarker after liver resection for HCC. However, only 10–20% of early-stage HCC patients present with abnormal AFP, which may be correlated with an aggressive HCC type (e.g., S2 class, epithelial cell adhesion molecule (EpCAM) positive) [87-89]. Moreover, AFP has low accuracy, as elevated AFP in patients with cirrhosis may reflect viral hepatitis and background liver disease in addition to HCC development [90]. Despite these limitations, AFP level was associated with recurrence with results externally validated in international cohorts [91, 92]. Furthermore, AFP level at recurrence predicted survival [93]. The following additional blood-borne biomarkers were associated with recurrence after resection for HCC: DCP [94], albumin-bilirubin score [91], circulating EpCAM positive tumour cells [95], preoperative CRP [96], preoperative neutrophil-to-lymphocyte ratio [97] and ratio of postoperative to preoperative circulating levels of cancer stem cell (CSC) markers K19 and CD44 [98]. Peroxisome proliferator-activated receptors (PPARs) are ligand-dependent nuclear receptors that regulate lipid metabolism and transduce metabolic and nutritional signals into transcriptional responses [99-101]. The association between PPAR and the development of HCC is not fully understood, and studies are somewhat conflicting. Moreover, the relationship may depend on context and species [102]. In some studies, high PPAR activity seems beneficial. Thus, it may reduce tumour growth via anti-inflammatory mechanisms [103], improve liver function [104] and resolve NASH [105]. As opposed to this, studies suggested unwanted consequences of PPAR activation, such as the induction of HCC due to regulation of the cell cycle [106] or formation of steatosis [107]. However, the prognostic importance in patients with HCC has not been investigated in a clinical setting. In Study II, we presented the results of RNA sequencing of HCC and non-tumour liver tissue in 52 patients undergoing liver resection. Based on the expression of PPAR target genes in heat maps (Fig. 7), we identified a group of patients (cancer cluster 3) with decreased overall survival (3-year overall survival of 48.6% vs 81.7% amongst controls, Fig. 8). The correlation remained significant in multivariate analysis adjusted for age, cirrhosis, microvascular invasion, number of tumours and resection margins (Table 1). Furthermore, we found that PPAR expression in non-tumour liver tissue was correlated with cirrhosis. In an earlier study using publicly available data from The Cancer Genome Atlas (TCGA), it was found that patients with HCC and low PPAR activity in tumour had significantly reduced overall survival [109]. However, in that study, PPAR activity from tumour tissue in patients with HCC was compared with liver tissue from healthy patients. Moreover, in a large study, PPAR expression evaluated with immunohistochemistry was correlated with shorter survival [110]. While immunohistochemistry may be a good surrogate for PPAR activity, transcriptome analysis allows a more direct evaluation of activity, as PPAR requires ligands and co-receptors to regulate gene expression [111]. In a phase 2 trial, patients with NASH were randomized to receive Lanifibranor, a PPAR agonist or a placebo [112]. Patients in the treatment arm receiving 1200 mg daily had reduced disease activity measured by the steatosis, activity and fibrosis [SAF] scoring system without the worsening of fibrosis. This suggested a possible clinical implication of PPAR not confined to HCC. Together with the correlation between PPAR and cirrhosis found in Study II, these findings may reflect that PPAR activity is important for liver cell homoeostasis. To our knowledge, Study II was the first clinical study to establish the prognostic value of PPAR gene activit