Lead Investigator: Guang Sheng Ling, University of Hong Kong
Title of Proposal Research: Using clinical characteristics of hepatocellular carcinoma (HCC) patients as predictors of their immunotherapy response
Vivli Data Request: 8853
Funding Source: Hong Kong General Research Fund: 17116622. HK’s GRF is a competitive process that supports both basic and applied research, with proposals evaluated on academic merit; contribution to academic/professional development; and potential for social, cultural or economic application
Potential Conflicts of Interest: None
Summary of the Proposed Research:
Cancer is a growing threat to human health, especially as our lifespan increases with advances in medicine. Cancer treatment has previously evolved from surgical removal to chemotherapy, a drug treatment which kills fast-growing cells, both cancerous and non-cancerous, in the body, and the most recent advancement is the application of immunotherapy approaches. Immunotherapy drugs aims to help the immune system identify cancer cells so it can specifically attack and kill them, leaving other cells unharmed. However, the outcomes for patients with solid tumors, such as hepatocellular carcinoma (HCC), when treated with immunotherapy are still poor, where only 20% of patients will benefit from the therapy. HCC is the most common form of liver cancer and is the second leading cause of cancer-related death in Asia. It is a growing threat due to rising obesity rates, which is associated with liver dysfunction that promotes tumor development.
Chronic viral infection with hepatitis B or C virus (HBV or HCV), carcinogens (food contaminants, tobacco smoking, and environmental toxins), excessive alcohol consumption (alcoholic fatty liver disease), metabolic syndrome (diabetes and obesity), and high-calorie intake (non-alcoholic fatty liver disease) are major risk factors for HCC. Each etiology (the manner in which the disease develops) has its own pathogenesis (set of factors that cause the disease) and immune microenvironment (the cells and fluids that surround a tumor some of which may promote tumor growth and some suppress it), which may contribute to immunotherapy resistance.
Our study aims to use the patient’s clinical characteristics, such as those described above, to predict their response to immunotherapy, and therefore inform decisions related to their treatment regime. An additional factor that will be considered in our analysis is the expression of the protein programmed death-ligand 1 (PD-L1) by the patients’ cells. Current popular immunotherapy treatments target the interactions of the protein PD-L1 with its receptor programmed cell death protein 1 (PD1), expressed by CD8 T cells (the main immune cells that control and kill cancer cells). Numerous research has demonstrated that the PD1-PD-L1 interactions are a key driver of CD8 T cell dysfunction (as termed ‘exhaustion’), hence treatments that disrupt this interaction have unprecedented efficacy. Therefore, we will also consider the patients’ PD-L1 expression as an independent variable in our analysis.
Our previous work identified that the exhaustion of CD8 T cells is a key factor for immunotherapy resistance, and we provided insights into drivers of their exhaustion. Our current work on publicly available datasets (from the Cancer Genome Atlas (TCGA), International Cancer Genomics Consortium (ICGC), and Liver Cancer Institute (LCI)) suggests that HCC’s underlying etiology, the tumour’s developmental stage, and evidence of microvascular invasion (presence of tumor cells in the blood vessels) can predict CD8 T cell exhaustion, which has been shown to predict immunotherapy outcomes in other cancers. Hence, we hypothesize that these three clinical characteristics can be directly used to predict immunotherapy response among HCC patients. Our requested studies contain the clinical characteristics and immunotherapy responses required for validating this hypothesis.
Overall, our research aims to advance precision medicine through using clinical characteristics as predictors of immunotherapy response for HCC patients.
Statistical Analysis Plan:
The GO30140 (NCT02715531) and IMbrave150 (NCT03434379) cohorts will be analyzed together in this meta-analysis of HCC patients undergoing treatment with atezolizumab and bevacizumab. To evaluate if the clinical characteristics of HCC patients can be used as predictors of their immunotherapy response, the patients will be classified into different groups based on the three main independent variables: HCC etiology, tumor stage, and presence of microvascular invasion. Their therapy outcomes, the dependent variable, consists of 4 groups (complete response (CR); partial response (PR); stable disease (SD); or progressive disease (PD)), and will be treated as a categorial variable. The correlation between patient groups and their therapeutic response will be explored by using Cramer’s V Correlation. The null hypothesis would be that the patient therapy outcomes are independent of our three main independent variables. We will perform bootstrapping to generate a confidence interval of the correlation. A 95% confidence interval will be used to judge the significance of the correlation. Multivariate logistic regression of the main independent variables with therapy outcomes will also be performed to calculate the odds ratio and 95% confidence intervals. Other clinical characteristics of the patient groups, such as their PD-L1 levels, will also be descriptively reported and expressed as percentages, or median values and interquartile ranges (IQR) as appropriate. Differences among these clinical characteristics will be analyzed using the Kruskal-Wallis test, while the Chi-squared test will be adopted for categorial variables. p-values calculated will be two-sided, with significance predefined to be at < 0.05. Imputation for missing values will only be done if there are insufficient samples after removing those with missing values. The statistical analysis will be performed using R on the Vivli platform using proper R packages (e.g., stats, gglot2, etc.) Requested Studies:
An Open-Label, Multicenter Phase Ib Study of The Safety and Efficacy of Atezolizumab (Anti-PD-L1 Antibody) Administered in Combination With Bevacizumab and/or Other Treatments in Patients With Solid Tumors
Data Contributor: Roche
Study ID: NCT02715531
Sponsor ID: GO30140
A Phase III, Open-Label, Randomized Study of Atezolizumab in Combination With Bevacizumab Compared With Sorafenib in Patients With Untreated Locally Advanced or Metastatic Hepatocellular Carcinoma
Data Contributor: Roche
Study ID: NCT03434379
Sponsor ID: YO40245
Summary of results:
Although we did not complete our statistical analysis plan (SAP), the key learnings from this proposal indicate that hepatocellular carcinoma (HCC) etiology, tumor stage, and microvascular invasion status do not serve as effective biomarkers for predicting immunotherapy response within the scope of our analysis. Initially, we classified patients into hepatitis B, hepatitis C, and non-viral groups, subsequently comparing their responsiveness to immunotherapy. The results demonstrated that there was no significant difference in immunotherapy response among these groups, as indicated by outcomes such as complete response (CR), partial response (PR), stable disease (SD), or progressive disease (PD). Furthermore, we categorized patients based on their tumor stages, spanning from early to advanced stages, and meticulously assessed their clinical outcomes following immunotherapy. While patients with early-stage tumors exhibited a seemingly improved response rate, this observation warrants cautious interpretation since the majority of patients typically receive immunotherapy during advanced stages. This confounding factor diminishes the informative value of the apparent differential response between early and late-stage tumors. Additionally, our comparative analysis between patients with and without microvascular invasion revealed no significant impact on the effectiveness of immunotherapy. To expand our investigative scope, we stratified patients according to their alpha-fetoprotein (AFP) levels, differentiating between high and low AFP cohorts. Similarly, we found no significant variations in immunotherapy response rates between these stratified groups. Our analysis also encompassed sex-based differences, comparing immunotherapy responses in male and female patients, as well as age-related differences by organizing patients into various age brackets. The results uniformly indicated that neither sex nor age bore a significant influence on immunotherapy outcomes. It is noteworthy to mention that our analysis of PD-L1 levels on tumor cells, as measured by histological sections, was impeded by incomplete data. This limitation precluded us from conducting a thorough analysis in this specific area. In conclusion, our comprehensive evaluation underscores that HCC etiology, tumor stage, microvascular invasion status, AFP levels, sex, and age are not reliable biomarkers for predicting immunotherapy response within the parameters of our study. These findings highlight the critical necessity for continued research efforts aimed at identifying and validating more robust and predictive biomarkers. Such advancements are essential for enhancing the efficacy and personalization of immunotherapy for patients with HCC, thereby improving therapeutic outcomes and clinical decision-making.