Studies were carried out to test the hypothesis that exposure to aflatoxin B1 (AFB1) is common among individuals with hepatocellular carcinoma (HCC) who are also chronically infected with hepatitis B virus (HBV). Experiments were also carried out to determine whether there is a close association between the presence of AFB1-DNA adducts and the expression of one or more HBV antigens in the tumor or non-tumor regions of the liver. Twenty-seven paired tumor and non-tumor liver tissues of HCC patients from Taiwan were analyzed. Monoclonal antibody 6A10, generated against the imidazole ring-opened persistent form of the major N-7 guanine adduct of AFB1, was used for adduct detection by both indirect immunofluorescence and competitive enzyme-linked immunosorbent assay. An avidin-biotin complex staining method was used for the detection of HBsAg and HBxAg in liver sections. A total of 8 (30%) HCC samples and 7 (26%) adjacent non-tumor liver tissue samples from Taiwan were positive for AFB1-DNA adducts. For HBsAg, 10 (37%) HCC samples and 22 (81%) adjacent non-tumorous liver samples were positive while 9 (33%) HCC samples and 11 (41%) adjacent non-tumor liver samples were HBxAg-positive. No association with AFB1-DNA adducts was observed for HBsAg and HBxAg. These results suggest that both AFB1 exposure and carrier status of HBsAg/HBxAg may be involved in the induction of HCC in Taiwan.

AimsHelicobacter pylori (HP) infection is the most common cause of chronic gastritis worldwide. Due to the small size of HP and limited resolution, diagnosing HP infections is more difficult when using digital slides.Methods and ResultsWe developed a two‐tier deep‐learning‐based model for diagnosing HP gastritis. A whole‐slide model was trained on 885 whole‐slide images (WSIs) with only slide‐level labels (positive or negative slides). An auxiliary model was trained on 824 areas with HP in nine positive WSIs and 446 negative WSIs for localizing HP. The whole‐slide model performed well, with an area under the receiver operating characteristic curve (AUC) of 0.9739 (95% confidence interval [CI], 0.9545–0.9932). The calculated sensitivity and specificity were 93.3% and 90.1%, respectively, whereas those of pathologists were 93.3% and 84.2%, respectively. Using the auxiliary model, the highlighted areas of the localization maps had an average precision of 0.5796.ConclusionsHP gastritis can be diagnosed on haematoxylin‐and‐eosin‐stained WSIs with human‐level accuracy using a deep‐learning‐based model trained on slide‐level labels and an auxiliary model for localizing HP and confirming the diagnosis. This two‐tiered model can shorten the diagnostic process and reduce the need for special staining.