Application of deep-learning to the automatic segmentation and classification of lateral lymph nodes on ultrasound images of papillary thyroid carcinoma.

PURPOSE: It is crucial to preoperatively diagnose lateral cervical lymph node (LN) metastases (LNMs) in papillary thyroid carcinoma (PTC) patients. This study aims to develop deep-learning models for the automatic segmentation and classification of LNM on original ultrasound images. METHODS: This study included 1000 lateral cervical LN ultrasound images (consisting of 512 benign and 558 metastatic LNs) collected from 728 patients at the Chongqing General Hospital between March 2022 and July 2023. Three instance segmentation models (MaskRCNN, SOLO and Mask2Former) were constructed to segment and classify ultrasound images of lateral cervical LNs by recognizing each object individually and in a pixel-by-pixel manner. The segmentation and classification results of the three models were compared with an experienced sonographer in the test set. RESULTS: Upon completion of a 200-epoch learning cycle, the loss among the three unique models became negligible. To evaluate the performance of the deep-learning models, the intersection over union threshold was set at 0.75. The mean average precision scores for MaskRCNN, SOLO and Mask2Former were 88.8%, 86.7% and 89.5%, respectively. The segmentation accuracies of the MaskRCNN, SOLO, Mask2Former models and sonographer were 85.6%, 88.0%, 89.5% and 82.3%, respectively. The classification AUCs of the MaskRCNN, SOLO, Mask2Former models and sonographer were 0.886, 0.869, 0.90.2 and 0.852 in the test set, respectively. CONCLUSIONS: The deep learning models could automatically segment and classify lateral cervical LNs with an AUC of 0.92. This approach may serve as a promising tool to assist sonographers in diagnosing lateral cervical LNMs among patients with PTC.

MALAT1 expression in granulosa cells in PCOS patients with different phenotypes.

Polycystic ovary syndrome (PCOS) is one of the most common reproductive endocrine metabolic disorders. The lncRNA metastasis-associated lung adenocarcinoma transcript 1 (MALAT1) reportedly can regulate the reproductive system. Despite studies, the alteration of MALAT1 expression in granulosa cells (GCs) from PCOS patients was inconsistent. To evaluate MALAT1 expression in GCs in different PCOS subgroups and its association with PCOS phenotypes, we collected GCs from 110 PCOS cases and 71 controls, and examined MALAT1 expression by quantitative PCR. The results showed MALAT1 expression was upregulated in PCOS cases, especially in insulin resistant (IR) PCOS subgroup, obese PCOS subgroup and non-hyperandrogenic (NHA) PCOS subgroup. MALAT1 expression was positively correlated with BMI and several metabolic parameters in controls. Interestingly, MALAT1 expression was notably associated with some critical endocrine indexes for PCOS, including E2, FSH, LH and LH/FSH ratio. In different PCOS subgroups, we found significant positive correlations with LH/FSH ratio in IR-PCOS and PCOS with normal weight, and with serum T and LH level in NHA-PCOS subgroup. Integrated analysis with lncRNA target databases and PCOS-related databases revealed MALAT1 could participate in PCOS by influencing immune response and lipids metabolism in GCs. In conclusion, MALAT1 was differently expressed in GCs in PCOS, especially in IR, obese and NHA PCOS subgroups. MALAT1 was likely involved in metabolism and immune response in GCs in PCOS. However, more studies are necessary to establish this concept.

Investigation of androgen receptor CAG repeats length in polycystic ovary syndrome diagnosed using the new international evidence-based guideline.

BACKGROUND: To study whether CAG repeat polymorphism of androgen receptor (AR) contributes to the risk of polycystic ovarian morphology (PCOM) with antral follicle count (AFC) ≥ 20 in the context of new international guideline of polycystic ovary syndrome (PCOS). METHODS: Blood of 109 PCOS cases and 61 controls were collected for the measurement of AR CAG repeats length by sequencing. The mean number and frequency distribution of CAG repeats length were observed. Detailed analysis was conducted by dividing PCOS cases into low AFC group (L-AFC, AFC < 20) and high AFC group (H-AFC, AFC ≥ 20) according to the new international evidence-based guideline. RESULTS: The portion of individuals with lower CAG repeats length in H-AFC group was significantly larger than those with higher CAG repeats length. Logistic model revealed individuals with lower CAG length tended to develop H-AFC. CONCLUSION: Lower CAG repeats length in the AR gene of PCOS cases increases risk of PCOM.

Deciphering the DNA methylome in women with PCOS diagnosed using the new international evidence-based guidelines.

STUDY QUESTION: Is there any methylome alteration in women with PCOS who were diagnosed using the new international evidence-based guidelines? SUMMARY ANSWER: A total of 264 differentially methylated probes (DMPs) and 53 differentially methylated regions (DMRs) were identified in patients with PCOS and healthy controls. WHAT IS KNOWN ALREADY: PCOS is a common endocrine disorder among women of reproductive age and polycystic ovarian morphology (PCOM) is one of the main features of the disease. Owing to the availability of more sensitive ultrasound machines, the traditional diagnosis of PCOM according to the Rotterdam criteria (≥12 antral follicles per ovary) is currently debated as there is a risk of overdiagnosis. The new international evidence-based guidelines set the threshold for PCOM as ≥20 antral follicles per ovary when using endovaginal ultrasound transducers with a frequency bandwidth that includes 8 MHz. However, current DNA methylation studies in PCOS are still based on the Rotterdam criteria. This study aimed to explore aberrant DNA methylation in patients diagnosed with PCOS according to the new evidence-based guidelines. STUDY DESIGN, SIZE, DURATION: This cross-sectional case-control study included 34 PCOS cases diagnosed using new international evidence-based guidelines and 36 controls. PARTICIPANTS/MATERIALS, SETTING, METHODS: A total of 70 women, including 34 PCOS cases and 36 controls, were recruited. DNA extracted from whole blood samples of participants were profiled using array technology. Data quality control, preprocessing, annotation, and statistical analyses were performed. Least absolute shrinkage and selection operator (LASSO) regression were used to build a PCOS diagnosis model with DNA methylation sites. MAIN RESULTS AND THE ROLE OF CHANCE: We identified 264 DMPs between PCOS cases and controls, which were mainly located in intergenic regions or gene bodies of the genome, CpG open sea sites, and heterochromatin of functional elements. Pathway enrichment analysis showed that DMPs were significantly enriched in biological processes involved in triglyceride regulation. Three of these DMPs overlapped with the PCOS susceptibility genes thyroid adenoma-associated protein (THADA), aminopeptidase O (AOPEP), and tripartite motif family-like protein 2 (TRIML2). Fifty-three DMRs were identified and their annotated genes were largely enriched in allograft rejection, thyroid hormone production, and peripheral downstream signaling effects. Two DMRs were closely related to the PCOS susceptibility genes, potassium voltage-gated channel subfamily A member 4 (KCNA4) and farnesyl-diphosphate farnesyltransferase 1 (FDFT1). Finally, based on LASSO regression, we built a methylation marker model with high accuracy for PCOS diagnosis (AUC=0.952). LIMITATIONS, REASONS FOR CAUTION: The study cohort was single-center and the sample size was relatively limited. Further analyses with a larger number of participants are required. WIDER IMPLICATIONS OF THE FINDINGS: This is the first study to identify DNA methylation alterations in women with PCOS diagnosed using the new international evidence-based guideline, and it provided new molecular insight into the application of the new guidelines. STUDY FUNDING/COMPETING INTEREST(S): This study was supported by the National Key Research and Development Program of China (2021YFC2700400), Basic Science Center Program of NSFC (31988101), CAMS Innovation Fund for Medical Sciences (2021-I2M-5-001), National Natural Science Foundation of China (32370916, 82071606, 82101707, 82192874, and 31871509), Shandong Provincial Key Research and Development Program (2020ZLYS02), Taishan Scholars Program of Shandong Province (ts20190988), and Fundamental Research Funds of Shandong University. The authors declare no conflicts of interest. TRIAL REGISTRATION NUMBER: N/A.

Systematical assessment of digit ratio in a female masculinization disease: polycystic ovary syndrome.

Background: In recent years, the right ratio of 2nd and 4th digit length (2D:4D) is regarded as an anatomical marker of prenatal testosterone exposure. Polycystic ovary syndrome (PCOS) is a female masculinized disease and is determined by prenatal testosterone exposure. Whether the ratio in the right hand of PCOS women is reduced or not compared with non-PCOS women is under debate. To further investigate the relationship between PCOS and digit ratio, we systematically measured all the digit ratios. Methods: We recruited 34 non-PCOS women, 116 PCOS women, and 40 men and systematically measured all the ratios of digit length (2D:3D, 2D:4D, 2D:5D, 3D:4D, 3D:5D, and 4D:5D) of right hands and left hands. Results: Left 2D:3D, 2D:4D, and 2D:5D in men were significantly lower than those in non-PCOS women. Significantly lower digit ratios of left 2D:3D and 2D:4D were observed in PCOS compared with non-PCOS women. In the subgroup analysis, the left ratio of digit length in 2D:3D and 2D:5D of the hyperandrogenism subgroup was lower than that of the non-hyperandrogenism subgroup without statistical significance. The logistic regression model of PCOS revealed that 2D:3D, 2D:4D, 2D:5D, and 3D:4D of left hands were statistically related to the diagnosis of PCOS among all the digit ratios. Conclusion: Not only 2D:4D but also other digit ratios, such as 2D:3D and 2D:5D, are a marker of prenatal testosterone exposure and may be an anatomical marker of PCOS. The majority of these significant differences included left 2D, with the following order: non-PCOS women > PCOS women > men.