Purpose: We aimed to characterize the role of selenium-binding protein 1 (SBP1) in hepatocellular carcinoma (HCC) invasiveness and underlying clinical significance.Experimental Design: SBP1 expression was measured in stepwise metastatic HCC cell lines by Western blotting. The role of SBP1 in HCC was investigated using siRNA. Immunofluorescence analyses were used to detect the interaction between SBP1 and glutathione peroxidase 1 (GPX1). Nineteen fresh tumor tissues and 323 paraffin-embedded samples were used to validate in vitro findings and to detect the prognostic significance of SBP1, respectively.Results: Inhibition of SBP1 effectively increased cell motility, promoted cell proliferation, and inhibited apoptosis only under oxidative stress; it also greatly enhanced GPX1 activity without altering GPX1 expression and downregulated hypoxia-inducible factor-1a (HIF-1a) expression. SBP1 and GPX1 formed nuclear bodies and colocalized under oxidative stress. In freshly isolated clinical HCC tissues, decreased SBP1 was linked with increased GPX1 activity and correlated with vascular invasion. Tumor tissue microarrays indicated that SBP1 was an independent risk factor for overall survival and disease recurrence; patients with lower SBP1 expression experienced shorter overall survival periods and higher rates of disease recurrence (P < 0.001). Further analyses indicated that the predictive power of SBP1 was more significant for patients beyond the Milan criteria than patients within the Milan criteria.Conclusions: Decreased expression of SBP1 could promote tumor invasiveness by increasing GPX1 activity and diminishing HIF-1a expression in HCC; SBP1 could be a novel biomarker for predicting prognosis and guiding personalized therapeutic strategies, especially in patients with advanced HCC.

Although immune checkpoint blockade have demonstrated promising results, their effects on gastric cancer (GC) are under investigation. Understanding the clinical significance of PD1 and its ligands' expression, together with T cell infiltration might provide clues for biomarkers screening in GC immunotherapy. Immunohistochemistry were performed on a tissue microarray including 1,014 GC specimens using PD1, PDL1 and PDL2 antibodies. T cell markers CD3 and CD8 were also stained and quantified by automated image analysis. Correlation with clinical features and outcome were analyzed after controlling for potential confounders including EBV infection, HER2, C-met and PCNA expression. 37.8% of the cases showed membranous PD-L1 expression in tumor cells and 74.9% in infiltrating immune cells. PDL1 expression rate was rather higher in patients without metastasis, in EBV positive group and those with C-met and PCNA expression. GC patients with high level PDL1 expression exhibited better survival. GC Patients with higher T cell infiltration also showed elevated PDL1, PDL2 and PD1 expression and predict favorable outcome, indicating an adaptive immune resistance mechanism may exist. The group of patients infiltrated with lower density CD3+ T cells also without PDL1 expression in tumor cells predict the worst outcome in the subgroup of different PTNM stage, which may suggest an inactive immune status. These results highlights the need to assess both PDL1 expression in all tumor context and the characterization of the GC immune microenvironment.

ObjectiveTumour pathology contains rich information, including tissue structure and cell morphology, that reflects disease progression and patient survival. However, phenotypic information is subtle and complex, making the discovery of prognostic indicators from pathological images challenging.DesignAn interpretable, weakly supervised deep learning framework incorporating prior knowledge was proposed to analyse hepatocellular carcinoma (HCC) and explore new prognostic phenotypes on pathological whole-slide images (WSIs) from the Zhongshan cohort of 1125 HCC patients (2451 WSIs) and TCGA cohort of 320 HCC patients (320 WSIs). A ‘tumour risk score (TRS)’ was established to evaluate patient outcomes, and then risk activation mapping (RAM) was applied to visualise the pathological phenotypes of TRS. The multi-omics data of The Cancer Genome Atlas(TCGA) HCC were used to assess the potential pathogenesis underlying TRS.ResultsSurvival analysis revealed that TRS was an independent prognosticator in both the Zhongshan cohort (p<0.0001) and TCGA cohort (p=0.0003). The predictive ability of TRS was superior to and independent of clinical staging systems, and TRS could evenly stratify patients into up to five groups with significantly different prognoses. Notably, sinusoidal capillarisation, prominent nucleoli and karyotheca, the nucleus/cytoplasm ratio and infiltrating inflammatory cells were identified as the main underlying features of TRS. The multi-omics data of TCGA HCC hint at the relevance of TRS to tumour immune infiltration and genetic alterations such as the FAT3 and RYR2 mutations.ConclusionOur deep learning framework is an effective and labour-saving method for decoding pathological images, providing a valuable means for HCC risk stratification and precise patient treatment.