Prediction of Epidermal Growth Factor Receptor Mutation Subtypes in Non-Small Cell Lung Cancer From Hematoxylin and Eosin-Stained Slides Using Deep Learning.

Accurate assessment of epidermal growth factor receptor (EGFR) mutation status and subtype is critical for the treatment of non-small cell lung cancer patients. Conventional molecular testing methods for detecting EGFR mutations have limitations. In this study, an artificial intelligence-powered deep learning framework was developed for the weakly supervised prediction of EGFR mutations in non-small cell lung cancer from hematoxylin and eosin-stained histopathology whole-slide images. The study cohort was partitioned into training and validation subsets. Foreground regions containing tumor tissue were extracted from whole-slide images. A convolutional neural network employing a contrastive learning paradigm was implemented to extract patch-level morphologic features. These features were aggregated using a vision transformer-based model to predict EGFR mutation status and classify patient cases. The established prediction model was validated on unseen data sets. In internal validation with a cohort from the University of Science and Technology of China (n = 172), the model achieved patient-level areas under the receiver-operating characteristic curve (AUCs) of 0.927 and 0.907, sensitivities of 81.6% and 83.3%, and specificities of 93.0% and 92.3%, for surgical resection and biopsy specimens, respectively, in EGFR mutation subtype prediction. External validation with cohorts from the Second Affiliated Hospital of Anhui Medical University and the First Affiliated Hospital of Wannan Medical College (n = 193) yielded patient-level AUCs of 0.849 and 0.867, sensitivities of 79.2% and 80.7%, and specificities of 91.7% and 90.7% for surgical and biopsy specimens, respectively. Further validation with the Cancer Genome Atlas data set (n = 81) showed an AUC of 0.861, a sensitivity of 84.6%, and a specificity of 90.5%. Deep learning solutions demonstrate potential advantages for automated, noninvasive, fast, cost-effective, and accurate inference of EGFR alterations from histomorphology. Integration of such artificial intelligence frameworks into routine digital pathology workflows could augment existing molecular testing pipelines.

Hypoxia-induced snail expression through transcriptional regulation by HIF-1α in pancreatic cancer cells.

BACKGROUND: Intratumoral hypoxia and epithelial-mesenchymal transition are involved in tumor invasion and metastasis. AIMS: This study investigated the molecular mechanisms that relay the hypoxia signal into the epithelial-mesenchymal transition and metastasis. METHODS: Morphology analysis and tumor cell migration and invasion assays were performed to detect phenotypic changes of pancreatic cancer cells under normoxic and hypoxic conditions after lentiviral HIF-1α shRNA transfection. Quantitative reverse transcription polymerase chain reaction, western blot, and immunohistochemistry were used to detect gene expression in pancreatic cancer cell lines and tissues or normal pancreatic tissues. Luciferase, gel shift, and ChIP assays were used to assess gene regulation. RESULTS: Under hypoxic conditions, these tumor cells underwent typical morphological and molecular changes to epithelial-mesenchymal transition. Moreover, Snail expression was induced by hypoxic conditions and was regulated by HIF-1α expression at the transcriptional level through HIF-1α-binding to the second site of hypoxia-responsive elements of the Snail gene promoter. In addition, Snail expression was associated with HIF-1α expression in pancreatic cancer tissues, and expression of both was associated with tumor metastasis and poor patient survival. CONCLUSIONS: Our study provides key evidence that HIF-1α and Snail are responsible for hypoxia-induced metastasis phenotypes in pancreatic cancer and that HIF-1α and Snail expression can be used as biomarkers to predict tumor metastasis and patient survival.

Expression and prognostic significance of CD151, c-Met, and integrin alpha3/alpha6 in pancreatic ductal adenocarcinoma.

BACKGROUND: CD151, c-Met, and integrin alpha3/alpha6 are all involved in the hepatocyte growth factor (HGF)/c-Met signal pathway, which plays an important role in the malignant progression of tumors. AIMS: The purpose of this study was to explore the expression and prognostic significance of these proteins in pancreatic ductal adenocarcinoma (PDAC). METHODS: We used immunohistochemical methods to investigate the expression patterns of CD151, c-Met, and integrin alpha3/alpha6proteins in 71 patients with PDAC and in ten samples of normal pancreatic tissue. We also assessed correlations between these proteins and clinicopathological parameters and survival of PDAC patients using various statistical methods. RESULTS: CD151, c-Met, and integrin alpha3/alpha6 were all overexpressed in PDAC. CD151 and c-Met overexpressions were significantly associated with TNM stage (p=0.001 and p=0.038, respectively) and lymph node invasion (p=0.000, p=0.012, respectively). A significant positive linear correlation was found between CD151 and c-Met (r=0.583; p=0.000), integrin alpha3 (r=0.457; p=0.000), and integrin alpha6 (r=0.671; p=0.000). Overexpression of CD151, c-Met, integrin alpha3, or integrin alpha6 was related to poor survival of PDAC patients (p=0.000, p=0.000, p=0.005, and p=0.003, respectively), and CD151 and c-Met were independent factors in prognosis of PDAC. CONCLUSIONS: CD151, c-Met, and integrin alpha3/alpha6 were all overexpressed in PDAC. CD151 and c-Met might be new molecular markers to predict the prognosis of PDAC patients.

Follicle-stimulating hormone inhibits apoptosis in ovarian cancer cells by regulating the OCT4 stem cell signaling pathway.

OCT4, a stem cell marker, is overexpressed in several types of human cancer and can induce resistance to chemotherapy and inhibition of apoptosis. We previously demonstrated that human follicle stimulating hormone (FSH) can inhibit ovarian cancer cell apoptosis. However, the role of OCT4 in FSH-induced inhibition of apoptosis has not been reported in detail. Here, we profiled OCT4 protein expression in ovarian epithelial cancer (OEC) with benign cystadenoma, borderline tumor and carcinoma tissues as well as different ovarian cancer cell lines and normal ovarian epithelial cells. Furthermore, the effects of FSH on OCT4 expression and related signaling pathways were evaluated. The overexpression of OCT4 in ovarian carcinoma and OEC cell lines suggest that OCT4 plays a critical role in OEC carcinogenesis. Moreover, FSH-induced apoptosis inhibition was confirmed and FSH stimulation induced the expansion of CD44+CD117+ cells with a stem cell-like phenotype. Re-expression of OCT4 enhanced the expression of Notch, Sox2 and Nanog molecules that play critical roles in cancer stem cell proliferation and differentiation. FSH upregulated the expression of Notch, Sox2 and Nanog and these effects were abolished by knocking down OCT4, suggesting that several cancer stem cell pathways are involved in FSH regulation. We also examined OCT4 expression in surgical specimens of ovarian cancer. Immunohistostaining revealed that OCT4 expression was increased in ovarian carcinoma compared with benign cystadenomas and borderline tumors, and OCT4 expression was significantly correlated with histological grade. Staining for OCT4 was increased in serous cystadenocarcinoma, when compared with clear cell carcinoma. In summary, the OCT4 cancer stem cell signaling pathway may mediate FSH-induced inhibition of apoptosis and could provide a target for treatment of ovarian cancer.