A.V. Asaturova

Eosinophilic cells associated with BRAF mutation in borderline serous ovarian tumors

Objective. To define the diagnostic value of eosinophilic cells for the detection of BRAF-mutated serous borderline ovarian tumors. Material and methods. The study included 42 cases of serous borderline ovarian tumor, each of which was analyzed by 3 pathologists for the presence of eosinophilic cells. Genetic profiling using Sanger sequencing was performed to identify the BRAFV600E mutation. Comparisons between two groups were performed using the Mann-Whitney test, Fisher’s exact test. Fleiss’s kappa was used to assess the interobserver agreement. To assess the diagnostic value of eosinophilic cells, sensitivity and specificity were assessed. Results. According to the results of a genetic study, the BRAFV600Emutation was found in 19 of 42 tumors. When analyzing interobserver agreement, the Fleiss’s kappa values allowed us to determine the reliability of the test as sufficient (ϰ=0.7). The sensitivity and specificity for predicting BRAFV600Emutation for eosinophilic cells were 78.9% and 91.3%, respectively. Patients with the BRAFV600E mutation were significantly younger than patients without it. Thus, the average age of patients in the group with the BRAFV600E mutation was 33.6±15.6 years, while in the group of tumors without the mutation the average age of patients was 43.9±12.7 years (p=0.002). Non-invasive implants were less frequently found in tumors with the BRAFV600E mutation compared to tumors without the mutation: 11.76% (2/17) versus 33.3% (6/18), respectively, but these differences were not statistically significant (p=0.228). Conclusion. Eosinophilic cells in ovarian serous borderline tumors may sufficiently reflect the BRAFV600E mutation, thereby correlating with disease prognosis (low risk of progression to low-grade serous carcinoma).

Sex cord tumor with annular tubules

Sex cord tumor with annular tubules is a rare type of ovarian tumor, which seems to arise from Sertoli cells. Most tumors are associated with Peutz—Jeghers syndrome and have a benign course and favorable prognosis; however, about 20% of sporadic sex cord tumors with annular tubules have a high risk of developing metastases. We present a case report about 18-year-old patient who presented with irregular menstruation and abdominal pain. Based on the histological features and immunohistochemical results, the diagnosis of a sex cord tumor with annular tubules was confirmed. The above-described ovarian tumor is rare in childhood and adolescence compared to the general population, which, combined with nonspecific clinical manifestations, can result certain diagnostic difficulties in this age group of patients.

Evaluation of general-purpose large language models as diagnostic support tools in cervical cytology

The application of general-purpose large language models (LLMs) in cytopathology remains largely unexplored. This study aims to evaluate the accuracy and consistency of a custom version of ChatGPT-4 (GPT), ChatGPT o3, and Gemini 2.5 Pro as diagnostic support tools for cervical cytology. A total of 200 Papanicolaou-stained cervical cytology images were acquired at 40x magnification, each measuring 384 × 384 pixels. These images consisted of 100 cases classified as negative for intraepithelial lesion or malignancy (NILM) and 100 cases across various abnormal categories: 20 low-grade squamous intraepithelial lesion (LSIL), 20 high-grade squamous intraepithelial lesion (HSIL), 20 squamous cell carcinoma (SCC), 20 adenocarcinoma in situ (AIS), and 20 adenocarcinoma (ADC). Diagnostic accuracy and consistency were evaluated by submitting each image to a GPT, ChatGPT o3, and Gemini 2.5 Pro 5-10 times. When distinguishing normal from abnormal cytology, LLMs showed mean sensitivity between 85.4 % and 100 %, and specificity between 67.2 % and 92.7 %. ChatGPT o3 was more accurate in identifying NILM (mean 89.2 % vs. 67.2 %) but less accurate in detecting LSIL (34 % vs. 85 %), HSIL (6 % vs. 63 %), and ADC (28 % vs. 91 %). Chain-of-thought prompting and submitting multiple images of the same diagnosis to ChatGPT o3 and Gemini 2.5 Pro did not significantly improve accuracy. Both models also performed poorly in identifying cervicovaginal infections. ChatGPT o3 and Gemini 2.5 Pro demonstrated complementary strengths in cervical cytology. Due to their low accuracy and inconsistency in abnormal cytology, general-purpose LLMs are not recommended as diagnostic support tools in cervical cytology.

Blood Plasma Small Non-Coding RNAs as Diagnostic Molecules for the Progesterone-Receptor-Negative Phenotype of Serous Ovarian Tumors

The expression level of the progesterone receptor (PGR) plays a crucial role in determining the biological characteristics of serous ovarian carcinoma. Low PGR expression is associated with chemoresistance and a poorer outcome. In this study, our objective was to explore the relationship between tumor progesterone receptor levels and RNA profiles (miRNAs, piwiRNAs, and mRNAs) to understand their biological characteristics and behavior. To achieve this, we employed next-generation sequencing of small non-coding RNAs, quantitative RT-PCR, and immunohistochemistry to analyze both FFPE and frozen tumor samples, as well as blood plasma from patients with benign cystadenoma (BSC), serous borderline tumor (SBT), low-grade serous ovarian carcinoma (LGSOC), and high-grade serous ovarian carcinoma (HGSOC). Our findings revealed significant upregulation of MMP7 and MUC16, along with downregulation of PGR, in LGSOC and HGSOC compared to BSC. We observed significant correlations of PGR expression levels in tumor tissue with the contents of miR-199a-5p, miR-214-3p, miR-424-3p, miR-424-5p, and miR-125b-5p, which potentially target MUC16, MMP7, and MMP9, as well as with the tissue content of miR-16-5p, miR-17-5p, miR-20a-5p, and miR-93-5p, which are associated with the epithelial–mesenchymal transition (EMT) of cells. The levels of EMT-associated miRNAs were significantly correlated with the content of hsa_piR_022437, hsa_piR_009295, hsa_piR_020813, hsa_piR_004307, and hsa_piR_019914 in tumor tissues. We developed two optimal logistic regression models using the quantitation of hsa_piR_020813, miR-16-5p, and hsa_piR_022437 or hsa_piR_004307, hsa_piR_019914, and miR-93-5p in the tumor tissue, which exhibited a significant ability to diagnose the PGR-negative tumor phenotype with 93% sensitivity. Of particular interest, the blood plasma levels of miR-16-5p and hsa_piR_022437 could be used to diagnose the PGR-negative tumor phenotype with 86% sensitivity even before surgery and chemotherapy. This knowledge can help in choosing the most effective treatment strategy for this aggressive type of ovarian cancer, such as neoadjuvant chemotherapy followed by cytoreduction in combination with hyperthermic intraperitoneal chemotherapy and targeted therapy, thus enhancing the treatment’s effectiveness and the patient’s longevity.

The association of rs25487 of the <i>XRCC1</i> gene and rs13181 of the <i>ERCC2 </i>gene polymorphisms with the ovarian cancer risk

Ovarian cancer (OC) is the most lethal gynecological cancer worldwide. DNA damage plays an important role in cancer development, and the proteins encoded by XRCC1 and ERCC2 are important components of the DNA repair system. This study aimed to examine the relationship between the rs25487 XRCC1 and rs13181 ERCC2 polymorphisms and the risk of OC development in women from the Moscow region. DNA was isolated from the blood of 129 healthy donors and tissues and blood samples from 125 patients with OC and studied using real-time PCR. An increase in odds ratios (OR) was obtained for OC tissue and blood for both T (OR = 1.46, 95% confidence interval [CI] = 1.22–1.76, P = 0.00005), and for T/T of rs25487 XRCC1. The most significant OR values were found for the T/T genotype using the codominant model (OR = 2.11, 95% CI = 1.44–3.07, P = 0.00006) and dominant model (OR = 3.13, 95% CI = 1.44–6.79, P = 0.0025) for the pooled blood and tissue groups. For rs13181 ERCC2, differences were observed for the T/G genotype in OC tissues (OR = 0.69, 95% CI = 0.51–0.92, P = 0.011) in the codominant model. In this study, the association of allele T and genotypes of rs25487 XRCC1 and T/G of rs13181 ERCC2 with OC was shown. Our results indicate that these polymorphisms may be involved in the pathogenesis of OC and are promising for further studies on therapeutic applications in OC.