Helicobacter pylori infection induces a number of pro-inflammatory signaling pathways contributing to gastric inflammation and carcinogenesis and has been identified as a major risk factor for the development of gastric cancer (GC). Janus kinase-signal transducer and activator of transcription (JAK-STAT) signaling mediates immune regulatory processes, including tumor-driven immune escape. Programmed death ligand 1 (PD-L1) expressed on gastric epithelium can suppress the immune system by shutting down T cell effector function. In a human cohort of subjects with gastric lesions and GC analyzed by proteomics, STAT1 increased along the cascade of progression of precancerous gastric lesions to GC and was further associated with a poor prognosis of GC (Hazard Ratio (95% confidence interval): 2.34 (1.04–5.30)). We observed that STAT1 was activated in human H. pylori-positive gastritis, while in GC, STAT1, and its target gene, PD-L1, were significantly elevated. To confirm the dependency of H. pylori, we infected gastric epithelial cells in vitro and observed strong activation of STAT1 and upregulation of PD-L1, which depended on cytokines produced by immune cells. To investigate the correlation of immune infiltration with STAT1 activation and PD-L1 expression, we employed a mouse model of H. pylori-induced gastric lesions in an Rnf43-deficient background. Here, phosphorylated STAT1 and PD-L1 were correlated with immune infiltration and proliferation. STAT1 and PD-L1 were upregulated in gastric tumor tissues compared with normal tissues and were associated with immune infiltration and poor prognosis based on the TCGA-STAD database. H. pylori-induced activation of STAT1 and PD-L1 expression may prevent immune surveillance in the gastric mucosa, allowing premalignant lesions to progress to gastric cancer.

Tumor budding is a histopathological biomarker associated with metastases and adverse survival outcomes in colorectal carcinoma (CRC) patients. It is characterized by the presence of single tumor cells or small clusters of cells within the tumor or at the tumor-invasion front. In order to obtain a tumor budding score for a patient, the region with the highest tumor bud density must first be visually identified by a pathologist, after which buds will be counted in the chosen hotspot field. The automation of this process will expectedly increase efficiency and reproducibility. Here, we present a deep learning convolutional neural network model that automates the above procedure. For model training, we used a semi-supervised learning method, to maximize the detection performance despite the limited amount of labeled training data. The model was tested on an independent dataset in which human- and machine-selected hotspots were mapped in relation to each other and manual and machine detected tumor bud numbers in the manually selected fields were compared. We report the results of the proposed method in comparison with visual assessment by pathologists. We show that the automated tumor bud count achieves a prognostic value comparable with visual estimation, while based on an objective and reproducible quantification. We also explore novel metrics to quantify buds such as density and dispersion and report their prognostic value. We have made the model available for research use on the grand-challenge platform.

INTRODUCTION: Oral microbiota may contribute to the development of upper gastrointestinal (UGI) disorders. In this study, we evaluated the association between microbiota of saliva, subgingival, and buccal mucosa and UGI disorders, particularly precancerous lesions. We also aimed to identify which oral site have the greatest potential as biomarkers for the development of UGI cancers. METHODS: In this population-based study, 388 adults underwent upper endoscopy with biopsies for histopathological analysis. UGI symptoms were assessed using a validated questionnaire, and 16S rRNA gene sequencing characterized the microbiota in 380 saliva, 200 subgingival, and 267 buccal mucosa samples collected during endoscopy. RESULTS: Dysbiosis of the salivary microbiota was observed in subjects with gastroesophageal reflux symptoms (GERSs) alone, as well as in those with combined conditions such as GERS and esophagitis, or esophagitis and Barrett's esophagus. Significant microbial alterations were also found in individuals with several stomach disorders including Helicobacter pylori infection, chemical reactive gastritis, atrophic gastritis, and intestinal metaplasia. However, microbiota dissimilarity in subgingival and buccal mucosa samples was primarily associated with Barrett's esophagus or gastric atrophy. Among identified genera in saliva, the association between Prevotella and Fusabacterium and atrophic gastritis and intestinal metaplasia was notable. In subgingival samples, the link of Fretibacterium with Barrett's esophagus and Fusabacterium with gastric atrophy and intestinal metaplasia has also been found to be important. DISCUSSION: Dysbiosis of saliva microbiota is linked to a broad spectrum of UGI disorders. However, microbiota dysbiosis in subgingival and buccal mucosa sites is specifically associated with the premalignant conditions such as Barrett's esophagus and gastric atrophy. Among oral sites, the subgingival microbiota shows more potential as a infectious biomarker for UGI cancers.

An altered Wnt-signaling activation has been reported during Barrett’s esophagus progression, but with rarely detected mutations in APC and β-catenin (CTNNB1) genes. In this study, a robust in-depth expression pattern analysis of frizzled receptors, co-receptors, the Wnt-ligands Wnt3a and Wnt5a, the Wnt-signaling downstream targets Axin2, and CyclinD1, as well as the activation of the intracellular signaling kinases Akt and GSK3β was performed in an in vitro cell culture model of Barrett’s esophagus. Representing the Barrett’s sequence, we used normal esophageal squamous epithelium (EPC-1, EPC-2), metaplasia (CP-A) and dysplasia (CP-B) to esophageal adenocarcinoma (EAC) cell lines (OE33, OE19) and primary specimens of squamous epithelium, metaplasia and EAC. A loss of Wnt3a expression was observed beginning from the metaplastic cell line CP-A towards dysplasia (CP-B) and EAC (OE33 and OE19), confirmed by a lower staining index of WNT3A in Barrett’s metaplasia and EAC, than in squamous epithelium specimens. Frizzled 1–10 expression analysis revealed a distinct expression pattern, showing the highest expression for Fzd2, Fzd3, Fzd4, Fzd5, Fzd7, and the co-receptor LRP5/6 in EAC cells, while Fzd3 and Fzd7 were rarely expressed in primary specimens from squamous epithelium. Despite the absence of an in-depth characterization of Wnt-signaling-associated receptors in Barrett’s esophagus, by showing variations of the Fzd- and co-receptor profiles, we provide evidence to have a significant role during Barrett’s progression and the underlying pathological mechanisms.

Ulcerative colitis (UC) is characterized by chronic inflammation of the colorectum. Histological remission has emerged as a potential future treatment goal; however, the histopathological assessment of intestinal inflammation in UC remains challenging with a multitude of available scoring systems and the need for a pathologist with expertise in inflammatory bowel disease (IBD). In previous studies, quantitative phase imaging (QPI) including digital holographic microscopy (DHM) was successfully applied as an objective method for stain-free quantification of the degree of inflammation in tissue sections. Here, we evaluated the application of DHM for the quantitative assessment of histopathological inflammation in patients with UC. In our study, endoscopically obtained colonic and rectal mucosal biopsy samples from 21 patients with UC were analyzed by capturing DHM-based QPI images that were subsequently evaluated using the subepithelial refractive index (RI). The retrieved RI data were correlated with established histological scoring systems including the Nancy index (NI) as well as with endoscopic and clinical findings. As a primary endpoint, we found a significant correlation between the DHM-based retrieved RI and the NI (R² = 0.251, p < 0.001). Furthermore, RI values correlated with the Mayo endoscopic subscore (MES; R² = 0.176, p < 0.001). An area under the receiver operating characteristics (ROC) curve of 0.820 confirms the subepithelial RI as a reliable parameter to distinguish biopsies with histologically active UC from biopsies without evidence of active disease as determined by conventional histopathological examination. An RI higher than 1.3488 was found to be the most sensitive and specific cut-off value to identify histologically active UC (sensitivity of 84% and specificity of 72%). In conclusion, our data demonstrate DHM to be a reliable tool for the quantitative assessment of mucosal inflammation in patients with UC.

Objective: Considering the essential role of KRAS mutation in NSCLC and the limited experience of PET radiomic features in KRAS mutation, a prediction model was built in our current analysis. Our model aims to evaluate the status of KRAS mutants in lung adenocarcinoma by combining PET radiomics and machine learning. Method: Patients were retrospectively selected from our database and screened from the NSCLC radiogenomic dataset from TCIA. The dataset was randomly divided into three subgroups. Two open-source software programs, 3D Slicer and Python, were used to segment lung tumours and extract radiomic features from ¹⁸F-FDG-PET images. Feature selection was performed by the Mann–Whitney U test, Spearman’s rank correlation coefficient, and RFE. Logistic regression was used to build the prediction models. AUCs from ROCs were used to compare the predictive abilities of the models. Calibration plots were obtained to examine the agreements of observed and predictive values in the validation and testing groups. DCA curves were performed to check the clinical impact of the best model. Finally, a nomogram was obtained to present the selected model. Results: One hundred and nineteen patients with lung adenocarcinoma were included in our study. The whole group was divided into three datasets: a training set (n = 96), a validation set (n = 11), and a testing set (n = 12). In total, 1781 radiomic features were extracted from PET images. One hundred sixty-three predictive models were established according to each original feature group and their combinations. After model comparison and selection, one model, including wHLH_fo_IR, wHLH_glrlm_SRHGLE, wHLH_glszm_SAHGLE, and smoking habits, was validated with the highest predictive value. The model obtained AUCs of 0.731 (95% CI: 0.619~0.843), 0.750 (95% CI: 0.248~1.000), and 0.750 (95% CI: 0.448~1.000) in the training set, the validation set and the testing set, respectively. Results from calibration plots in validation and testing groups indicated that there was no departure between observed and predictive values in the two datasets (p = 0.377 and 0.861, respectively). Conclusions: Our model combining ¹⁸F-FDG-PET radiomics and machine learning indicated a good predictive ability of KRAS status in lung adenocarcinoma. It may be a helpful non-invasive method to screen the KRAS mutation status of heterogenous lung adenocarcinoma before selected biopsy sampling.