Background: Hysteroscopy is a commonly used technique for diagnosing endometrial lesions. It is essential to develop an objective model to aid clinicians in lesion diagnosis, as each type of lesion has a distinct treatment, and judgments of hysteroscopists are relatively subjective. This study constructs a convolutional neural network model that can automatically classify endometrial lesions using hysteroscopic images as input. Methods: All histopathologically confirmed endometrial lesion images were obtained from the Shengjing Hospital of China Medical University, including endometrial hyperplasia without atypia, atypical hyperplasia, endometrial cancer, endometrial polyps, and submucous myomas. The study included 1851 images from 454 patients. After the images were preprocessed (histogram equalization, addition of noise, rotations, and flips), a training set of 6478 images was input into a tuned VGGNet-16 model; 250 images were used as the test set to evaluate the model’s performance. Thereafter, we compared the model’s results with the diagnosis of gynecologists. Results: The overall accuracy of the VGGNet-16 model in classifying endometrial lesions is 80.8%. Its sensitivity to endometrial hyperplasia without atypia, atypical hyperplasia, endometrial cancer, endometrial polyp, and submucous myoma is 84.0%, 68.0%, 78.0%, 94.0%, and 80.0%, respectively; for these diagnoses, the model’s specificity is 92.5%, 95.5%, 96.5%, 95.0%, and 96.5%, respectively. When classifying lesions as benign or as premalignant/malignant, the VGGNet-16 model’s accuracy, sensitivity, and specificity are 90.8%, 83.0%, and 96.0%, respectively. The diagnostic performance of the VGGNet-16 model is slightly better than that of the three gynecologists in both classification tasks. With the aid of the model, the overall accuracy of the diagnosis of endometrial lesions by gynecologists can be improved. Conclusions: The VGGNet-16 model performs well in classifying endometrial lesions from hysteroscopic images and can provide objective diagnostic evidence for hysteroscopists.

Background: Endometrial cancer (EC) is one of the three major gynecological malignancies. Numerous biomarkers that may be associated with survival and prognosis have been identified through database mining in previous studies. However, the predictive ability of single-gene biomarkers is not sufficiently specific. Genetic signatures may be an improved option for prediction. This study aimed to explore data from The Cancer Genome Atlas (TCGA) to identify a new genetic signature for predicting the prognosis of EC. Methods: mRNA expression profiling was performed in a group of patients with EC (n = 548) from TCGA. Gene set enrichment analysis was performed to identify gene sets that were significantly different between EC tissues and normal tissues. Cox proportional hazards regression models were used to identify genes significantly associated with overall survival. Quantitative real-time-PCR was used to verify the reliability of the expression of selected mRNAs. Subsequent multivariate Cox regression analysis was used to establish a prognostic risk parameter formula. Kaplan–Meier survival estimates and the log‐rank test were used to validate the significance of risk parameters for prognosis prediction. Result: Nine genes associated with glycolysis (CLDN9, B4GALT1, GMPPB, B4GALT4, AK4, CHST6, PC, GPC1, and SRD5A3) were found to be significantly related to overall survival. The results of mRNA expression analysis by PCR were consistent with those of bioinformatics analysis. Based on the nine-gene signature, the 548 patients with EC were divided into high/low-risk subgroups. The prognostic ability of the nine-gene signature was not affected by other factors. Conclusion: A nine-gene signature associated with cellular glycolysis for predicting the survival of patients with EC was developed. The findings provide insight into the mechanisms of cellular glycolysis and identification of patients with poor prognosis in EC.

Background: The preoperative peripheral blood neutrophil-lymphocyte ratio (NLR), platelet-lymphocyte ratio (PLR) and monocyte-lymphocyte ratio (MLR) have been reported to be associated with the prognosis of various cancers but are always discussed separately. The aim of this study is to bring the combination of NLR, PLR and MLR into the prognostic assessment system of endometrial cancer (EC) and establish a nomogram to provide an objective prediction model for clinical decisions. Methods: A total of 1111 patients with EC who had accepted surgical treatment during 2013–2017 were involved in the analysis. Their NLR, PLR, and MLR levels were obtained from a routine blood examination within 2 weeks before operation. Receiver operating characteristic curve (ROC) analysis was performed to determine optimal cutoffs. Chi-square tests analysed the associations of the ratios with other clinicopathological variables. The prognostic value was indicated by overall survival (OS) via Cox proportional hazards models and Kaplan-Meier analysis. R software was used to establish the nomogram based on the combination of NLR, PLR, MLR and other clinicopathological factors. Results: The median follow-up period was 40 months, and the median age was 56. The enrolled patients were stratified by cutoffs of 2.14 for NLR, 131.82 for PLR and 0.22 for MLR. Multivariate analyses demonstrated that high NLR over 2.14 (HR = 2.71, 95%CI = 1.83–4.02, P<0.001), high PLR over 131.82 (HR = 2.75, 95%CI = 1.90–3.97, P<0.001), and high MLR over 0.22 (HR = 1.72, 95%CI = 1.20–2.45, P = 0.003) were significantly associated with worse OS. The combined indicator, high NLR + high PLR + high MLR (HR = 4.34, 95%CI = 2.54–7.42, P<0.001), showed the highest prognostic value. The Harrell’s concordance index of the nomogram was 0.847 (95% CI = 0.804–0.890), showing good discrimination and calibration of this model. Conclusion: The combination of NLR, PLR, and MLR is a superior prognostic factor of EC. The nomogram involving the combination of NLR, PLR, MLR and other clinicopathological factors is recommended to predict OS for EC patients clinically.

Long noncoding RNAs (lncRNAs) have been implicated in tumorigenesis and cancer progression and are tightly associated with the phenotypes of numerous cancers. However, the functional roles underlying these effects are unknown. The expression levels of LINC01016, miR-302a-3p, miR-3130-3p, NFYA, and SATB1 were evaluated by quantitative real-time polymerase chain reaction (qRT-PCR) in 33 endometrial cancer tissues and 20 normal tissues. Bioinformatics analyses, luciferase reporter analyses, chromatin immunoprecipitation (ChIP) assays, and qRT-PCR assays were performed to verify potential binding sites. The qRT-PCR and western blot were used to identify the regulatory mechanisms of LINC01016 in cell biological behavior, which were also examined by cell counting kit -8 (CCK-8), 5-ethynyl-2′-deoxyuridine (EdU) assays, flow cytometry, wound healing assays, and transwell assays. LINC01016 was substantially upregulated in endometrial cancer tissues, and LINC01016 silencing abolished the malignant behavior of endometrial cancer cells. LINC01016 positively rescued the downstream gene nuclear factor YA (NFYA) by competitively “sponging” miR-302a-3p and miR-3130-3p. In turn, these two miRNAs could inhibit LINC01016 transcription, thus forming two reciprocal repression cycles, which influenced the biological behavior of endometrial cancer cells. MiR-302a-3p and miR-3130-3p could specifically bind with the 3′-UTR regions of NFYA, and NFYA could upregulate the expression of special AT-rich sequence-binding protein 1 (SATB1) as a transcriptional factor. This study was the first to show that the LINC01016–miR-302a-3p/miR-3130-3p/NFYA/SATB1 axis played a crucial role in the occurrence of endometrial cancer. These findings may provide relevant insights into the diagnosis and therapy of endometrial cancer.

Metastasis is one of the main reasons for treatment failure in endometrial cancer. Notably, high mobility group AT-hook 2 (HMGA2) has been recognized as a driving factor of tumour metastasis. microRNAs (miRNAs) are powerful posttranscriptional regulators of HMGA2. The binding sites of miR-302a-5p and miR-367-3p on HMGA2 mRNA were identified using bioinformatics prediction software and were validated via luciferase assay. The expression levels of miR-302a-5p and miR-367-3p were detected using quantitative real-time PCR and in situ hybridization. Western blotting and immunohistochemistry were used to detect the levels of HMGA2 and epithelial-mesenchymal transition pathway-related proteins. Co-immunoprecipitation was used to detect protein interactions. The roles of miR-302a-5p and miR-367-3p in the regulation of HMGA2 during the progression of endometrial cancer were investigated using both in vitro and in vivo assays. In the present study, high HMGA2 expression was correlated with poor clinical outcomes in endometrial cancer. The binding sites of miRNAs on HMGA2 mRNA were identified using bioinformatics prediction software and were validated via luciferase assay. In the endometrial cancer cell lines Ishikawa and HEC-1A, the overexpression of miR-302a-5p/367-3p significantly inhibited the expression of HMGA2 mRNA. In endometrial cancer tissues, we showed that miR-302a-5p and miR-367-3p were significantly downregulated and thus inversely correlated with HMGA2. The miR-302a-5p and miR-367-3p expression levels were closely correlated with FIGO stage and lymph node metastasis. High expression of miR-302a-5p/367-3p was correlated with high survival rates in endometrial cancer. In addition, miR-302a-5p/367-3p suppressed the malignant behaviour of endometrial carcinoma cells via the inhibition of HMGA2 expression. Our findings indicate that miR-302a-5p/367-3p-mediated expression of HMGA2 regulates the malignant behaviour of endometrial carcinoma cells, which suggests that the miR-302a-5p/367-3p-HMGA2 axis may be a predictive biomarker of endometrial cancer metastasis and patient survival and a potential therapeutic target in metastatic endometrial cancer.