ABSTRACT
PURPOSE: Tumor-infiltrating tissue-resident memory CD8 T cells (CD8 TRM; CD103+ CD8+) are considered tumor-specific and may correlate better with the tumor response to immune checkpoint blockade (ICB). This study evaluated the association of tumor-infiltrating CD8 TRM and their subsets with the efficacy of immunotherapy in patients with advanced hepatocellular carcinoma (HCC). EXPERIMENTAL DESIGN: Consecutive HCC patients who received ICB in prospective trials were analyzed. Formalin-fixed paraffin-embedded tumor sections were stained for DAPI, CD8, CD103, CD39, programmed cell death-1 (PD-1), and programmed cell death ligand 1 (PD-L1) using a multiplex immunohistochemical method. The densities of CD8 T cells, CD8 TRM, and CD39+ or PD-L1+ subsets of CD8 TRM were correlated with tumor response and overall survival (OS). RESULTS: A total of 73 patients were identified, and 48 patients with adequate pretreatment tumor specimens and complete follow-up were analyzed. A median of 32.7% (range: 0-92.6%) of tumor-infiltrating CD8 T cells were TRM. In subset analyses, 66.6% ± 34.2%, 69.8% ± 33.4%, and 0% of CD8 TRM cells coexpressed CD39, PD-L1, and PD-1, respectively. The objective response rates for CD8 T cell-high, CD8 TRM-high, CD39+ CD8 TRM-high, and PD-L1+ CD8 TRM-high groups were 41.7%, 37.5%, 37.5%, and 29.2%, respectively. Patients with CD8 T cell-high, but not those with CD8 TRM-high, CD39+ CD8 TRM-high, or PD-L1+ CD8 TRM-high, tumors, had significantly prolonged OS (p = 0.0429). CONCLUSIONS: Compared with total tumor-infiltrating CD8 T cells, tumor-infiltrating CD8 TRM or their subsets failed to provide additional advantages in predicting the efficacy of immunotherapy for HCC.
ABSTRACT
BACKGROUND: Gut microbiome has been associated with the efficacy of immune checkpoint inhibitors (ICI) in patients with various types of cancers but not yet in hepatocellular carcinoma (HCC). AIMS: To investigate the association between gut microbiome and efficacy of ICI in patients with HCC. METHODS: Patients with HCC who were scheduled to receive ICI were prospectively enrolled. Fecal samples were collected within 7 days before initiation of ICI (baseline) and 8 weeks later. Gut microbiome was assessed using 16S rRNA sequencing and shotgun whole-genome sequencing and correlated with objective response (complete or partial response), disease control (objective response or stable disease for ≥16 weeks), and overall survival. RESULTS: Thirty-six patients with HCC were enrolled, and 20 of them provided both baseline and 8-week feces. Alpha diversity, richness, and compositions of baseline gut microbiome indicated no difference between responders and nonresponders or between disease control and nondisease control groups. For the 20 paired feces, immunotherapy did not change any of the major microbiome features. No specific taxa were enriched in patients with objective response. Three taxa-Bifidobacterium, Coprococcus, and Acidaminococcus-were enriched in patients with disease control. However, the baseline abundance of these three taxa did not predict overall survival benefit. CONCLUSIONS: In this exploratory study, we failed to disclose any overt association of gut microbiome with the efficacy of ICI in patients with HCC. A larger prospective study is warranted for definite conclusion.
ABSTRACT
BACKGROUND & AIMS: Intratumor heterogeneity has frequently been reported in patients with hepatocellular carcinoma (HCC). Thus, the reliability of single-region tumor samples for evaluation of the tumor immune microenvironment is also debatable. We conducted a prospective study to analyze the similarity in tumor immune microenvironments among different regions of a single tumor. METHODS: Multi-region sampling was performed on newly resected tumors. The tumor immune microenvironment was evaluated by immunohistochemical staining of PD-L1, CD4, CD8, CD20, FoxP3, DC-LAMP (or LAMP3), CD68, MPO, and tertiary lymphoid structures (TLSs). PD-L1 expression was manually quantified according to the percentage of PD-L1-stained tumor or stromal cells. The densities (number/mm2) of immune cells and the number of TLSs per sample were determined by whole-section counting. RNA-sequencing was applied in selected samples. Similarities in tumor immune microenvironments within each tumor were evaluated by multivariate Mahalanobis distance analyses. RESULTS: Thirteen tumors were collected from 12 patients. The median diameter of tumors was 9â¯cm (range 3-16â¯cm). A median of 6 samples (range 3-12) were obtained from each tumor. Nine (69.2%) tumors exhibited uniform expression of PD-L1 in all regions of the tumor. Out of 13 tumors analyzed by immunohistochemical staining, 8 (61.5%) tumors displayed a narrow Mahalanobis distance for all regions within the tumor; while 8 (66.7%) of the 12 tumors analyzed by RNA-sequencing displayed a narrow Mahalanobis distance. Immunohistochemistry and RNA-sequencing had a high concordance rate (83.3%; 10 of 12 tumors) for the evaluation of similarities between tumor immune microenvironments within a tumor. CONCLUSIONS: A single-region tumor sample might be reliable for the evaluation of tumor immune microenvironments in approximately 60-70% of patients with HCC. LAY SUMMARY: Heterogeneity in the regional immune microenvironments of tumors has been reported in patients with hepatocellular carcinoma. This heterogeneity could be an obstacle when trying to reliably evaluate the immune microenvironment of an entire tumor using only a single-region tumor sample, which may be the only option in patients with more advanced disease. Our study utilized both immunohistochemical and transcriptomic analyses to demonstrate that a single-region sample is reliable for evaluation of tumor immune microenvironments in 60-70% of patients with hepatocellular carcinoma.
