Purpose: This study aimed to investigate the inflammatory-immune cells in the peripheral blood of women with polycystic ovary syndrome (PCOS) and assessed the potential correlation between inflammatory-immune cells and infertility in PCOS women. Materials and Methods: In this case–control study, the profiles of lymphocyte subsets were analyzed by flow cytometry. White blood cells (WBC), neutrophils (Neu), lymphocytes, Ferriman–Gallwey (F–G) score, testosterone, prolactin, follicle-stimulating hormone, luteinizing hormone, fasting blood glucose, and fasting plasma insulin were measured, together with body mass index. Association between inflammatory-immune cells and PCOS was evaluated. Moreover, inflammatory-immune cells of the PCOS women with infertility were evaluated, and the relative operating characteristic (ROC) curve and cutoff values were calculated. Results: The number of WBC, Neu, and lymphocytes was higher in PCOS women than controls (P< 0.05). The percentages of total T lymphocytes, CD4+T, and NK were significantly increased in the PCOS group (P< 0.001). The CD4/CD8 ratio was obviously elevated for increasing CD4+T (P< 0.05). Consequently, T%, CD4+T%, and NK% were found to be the independent risk factors of PCOS by ROC curve and multivariate logistic regression analysis. Furthermore, only NK% was significantly higher in PCOS women with infertility than those who had PCOS without infertility (P< 0.001). To diagnose infertility in PCOS, the cutoff value of NK% was calculated as 16.43%. Conclusion: These findings suggest that the pathogenesis of PCOS is related to immune cells including T, CD4+T, and NK cells. NK cells are likely to be a potential predictive factor for PCOS women with infertility.

Sun P, et al. J Gynecol Oncol. 2017 Mar;28(5):e50. https://doi.org/10.3802/jgo.2017.28.e50

Purpose: Currently, the associations between type-specific high-risk human papillomavirus (HR-HPV) viral loads and cervical lesions are still inconsistent. We aimed to assess the type-specific HR-HPV viral load as a risk triage indicator for development of high-grade squamous intraepithelial lesion or worse (≥HSIL). Patients and methods: A total of 19,446 women who underwent primary screening for cervical cancer using Cervista^® HR-HPV and cytology assays were enrolled. The viral loads of 1,396 HR-HPV-positive specimens confirmed by Cervista^® assay were detected by BioPerfectus Multiplex Real-Time PCR assay. The correlation between viral loads and cervical lesions was analyzed. The optimal cutoffs of individual HR-HPV viral loads used to predict ≥HSIL were determined from the receiver operating characteristic curve. A logistic regression model was used to analyze the relationship between covariates and the probability of ≥HSIL. Results: The viral loads of HPV-16, -31, -33, -52, and -58 were positively correlated with the severity of the cervical lesion, which was significantly elevated in patients with ≥HSIL, whereas those of HPV-18, -45, -56, -59, and other types were not. The optimal cutoffs of the log₁₀-transformed viral loads for HPV-16, -31, -33, -52, and -58 in identifying ≥HSIL were 4.26, 4.46, 4.48, 4.36, and 4.26 copies per 10,000 cells, respectively. Furthermore, multivariate analysis indicated that type-specific viral loads of HPV-16, -31, -33, -52, and -58 exceeding the cutoffs could be independent risk factors for the incidence of ≥HSIL. Conclusion: The BioPerfectus Multiplex Real-Time PCR viral load assay provides viable triage for ≥HSIL when using appropriate cutoff levels.

Colposcopy is an essential medical examination that mainly through visual inspection for cervical epithelial tissue to preventing cervical cancer. However, screening by artificial vision has the problems of missed diagnosis, low efficiency, and diversity. This paper proposes a deep learning-based method using multi-CNN decision feature integration for classification and diagnosis of cervical lesions. The proposed method first uses the k-means algorithm to aggregate training data into specific classes in data preprocessing and then trained in cross-validation to improve the generalization ability of the model. Subsequently, two kinds of CNN are separately fine-tuned based on the transfer learning, and then XGBoost algorithm is used to integrate the different CNN decision results for optimizing the final prediction. Two integration ways for decision features (named inner-model integration and outer-model integration) are proposed to embody multi-feature integration from the data level and the model structural level. Simultaneously, two specific integration strategies (inner-to-outer and outer-to-inner) are designed for as the final output based on the combination of the two ways. Experimental results show that the K-means data preprocessing method can improve the training of neural networks, and the proposed multi-decision feature fusion strategy can better obtain the results of computer-aided diagnosis that meet the needs of clinical diagnosis.