We appreciate Dr. Nishida for his interest on our recent study [1] and insightful editorial on personalized approaches to treating hepatocellular carcinoma (HCC) using immune checkpoint inhibitors (ICIs) [2]. Dr. Nishida highlights the significant advances made in treatment for HCC patients based on the tumor immune microenvironment (TiME) and the critical need for reliable molecular biomarkers to enhance overall therapeutic outcomes. We would like to extend this discussion by highlighting the importance of integrating single-cell level data, pathology, and medical imaging data to further refine biomarkers for immunotherapy responses.

Expanding on Dr. Nishida’s insights, we agree that detailed characterization of the tumor microenvironment (TME) is vital for predicting responses to ICIs. However, current transcriptome-based approaches, including the immune signature score (ISS10), while promising, may overlook critical nuances of tumor heterogeneity. Single cell RNA-sequencing (scRNA-seq) technologies offer an unprecedented level of detail by allowing us to analyze individual cells within the context of the TME [3]. This enables the identification of specific immune cell populations that may drive immunosuppression or resistance to ICIs. Furthermore, the dynamic profiling facilitated by scRNA-seq allows us to monitor these immune populations over time, assessing how they evolve in response to treatment and providing real-time insights into therapeutic efficacy. Therefore, integrating scRNA-seq data into predictive biomarkers could refine patient stratification and help overcome the challenge of immune resistance in HCC.

In addition to molecular data, the integration of advanced pathology and medical imaging techniques could significantly enhance our understanding of the spatial architecture of the TME. As HCC exhibit profound heterogeneity in immune cell infiltration, spatial transcriptomics and proteomics using imaging mass cytometry could help map immune cell distributions across different regions of the tumor [4]. This would allow for a more nuanced understanding of how immune effector cells interact with tumor cells and how immunosuppressive cells are positioned within the tumor landscape. Moreover, medical imaging technologies such as radiomics could provide non-invasive means of assessing tumor features, such as immune infiltration, angiogenesis, and tumor vasculature, that correlate with responses to ICIs [5]. Combining these imaging data with molecular biomarkers would give us a more comprehensive tool set for predicting immunotherapy outcomes in realworld clinical settings.

Another critical aspect of HCC biomarker research that requires further attention is the role of cancer etiology. HCC is a highly heterogeneous disease with distinct etiological factors, such as hepatitis B virus (HBV), hepatitis C virus (HCV), alcohol consumption, and non-alcoholic steatohepatitis (NASH), contributing to its development [6,7]. These etiologies have profound effects on the TME, immune cell composition, and the mutational landscape of the tumor. For example, HBV-driven HCC often exhibits higher levels of T-cell exhaustion, whereas HCV-associated HCC is characterized by chronic inflammation and increased infiltration of immune cells [8,9]. Incorporating etiology data into biomarker studies could help refine the predictive power of existing molecular signatures. Moreover, it could lead to the discovery of etiology-specific biomarkers that predict responses to ICIs more accurately.

Recent advancements in personalized cancer vaccines have shown significant promise in improving outcomes for patients with HCC, as demonstrated in a recent phase 1/2 trial [10]. In this study, a personalized neoantigen vaccine was combined with pembrolizumab (anti-PD-1 antibody) in patients with advanced HCC. The trial demonstrated that the vaccine, tailored to each patient’s specific tumor mutations, elicited strong immune responses and showed encouraging clinical activity. By targeting tumor-specific neoantigens, these vaccines have the potential to overcome the immunosuppressive tumor microenvironment seen in the “immune-excluded” or “immune-desert” subtypes of HCC [11]. The study also highlighted that combining the neoantigen vaccine with pembrolizumab enhanced T-cell infiltration into tumors, improving the efficacy of immune checkpoint inhibition. This approach represents a promising strategy for turning “cold” tumors into “hot” ones, thereby increasing their sensitivity to immunotherapy.

We remain grateful for Dr. Nishida’s insightful comments and expert perspective. We see significant potential in combining scRNA-seq technologies, pathology, and medical imaging data to advance biomarker discovery. Additionally, personalized cancer vaccines represent a promising new direction in HCC treatment. We believe that combining these approaches with deep learning and artificial intelligence could address the complex heterogeneity of the tumor microenvironment and lead to more targeted and effective immunotherapies for HCC.