Background: The objective of this study was to estimate the accuracy of transcriptome-based classifier in differential diagnosis of uterine leiomyoma and leiomyosarcoma. We manually selected 114 normal uterine tissue and 31 leiomyosarcoma samples from publicly available transcriptome data in UCSC Xena as training/validation sets. We developed pre-processing procedure and gene selection method to sensitively find genes of larger variance in leiomyosarcoma than normal uterine tissues. Through our method, 17 genes were selected to build transcriptome-based classifier. The prediction accuracies of deep feedforward neural network (DNN), support vector machine (SVM), random forest (RF), and gradient boosting (GB) models were examined. We interpret the biological functionality of selected genes via network-based analysis using GeneMANIA. To validate the performance of trained model, we additionally collected 35 clinical samples of leiomyosarcoma and leiomyoma as a test set (18 + 17 as 1st and 2nd test sets). Results: We discovered genes expressed in a highly variable way in leiomyosarcoma while these genes are expressed in a conserved way in normal uterine samples. These genes were mainly associated with DNA replication. As gene selection and model training were made in leiomyosarcoma and uterine normal tissue, proving discriminant of ability between leiomyosarcoma and leiomyoma is necessary. Thus, further validation of trained model was conducted in newly collected clinical samples of leiomyosarcoma and leiomyoma. The DNN classifier performed sensitivity 0.88, 0.77 (8/9, 7/9) while the specificity 1.0 (8/8, 8/8) in two test data set supporting that the selected genes in conjunction with DNN classifier are well discriminating the difference between leiomyosarcoma and leiomyoma in clinical sample. Conclusion: The transcriptome-based classifier accurately distinguished uterine leiomyosarcoma from leiomyoma. Our method can be helpful in clinical practice through the biopsy of sample in advance of surgery. Identification of leiomyosarcoma let the doctor avoid of laparoscopic surgery, thus it minimizes un-wanted tumor spread.
Background/Objectives: Ovarian cancer (OC) remains one of the most lethal gynecologic malignancies, largely due to the challenges of early detection. While next-generation sequencing (NGS) has been explored for screening, its high cost limits large-scale implementation. To develop a more accessible diagnostic solution, we designed a qPCR-based algorithm optimized for early OC detection, with a focus on high-grade serous ovarian cancer (HGSOC), the most aggressive subtype. Methods: Peripheral blood samples from 19 ovarian cancer patients, 37 benign tumor patients, and 34 asymptomatic controls were analyzed using RNA sequencing to identify splice junction-based biomarkers with minimal expression in benign samples but elevated in OC. Results: A final panel of 10 markers was validated via qPCR, demonstrating strong agreement with sequencing data (R2 = 0.44–0.98). The classification algorithm achieved 94.1% sensitivity and 94.4% specificity (AUC = 0.933). Conclusions: By leveraging platelet RNA profiling, this approach offers high specificity, accessibility, and potential for early OC detection. Future studies will focus on expanding histologic diversity and refining biomarker panels to further enhance diagnostic performance.
Early detection of ovarian cancer is crucial for successful treatment, yet most cases are diagnosed at advanced stages due to a lack of effective screening. Recent advancements in RNA technology from platelets aid in early tumor detection. Here, we proposed our two-step method for assessing the existence of pelvic mass either located at ovaries or uterus with more than 99% specificity by utilizing exon-exon junction features with a sampling invariant normalization technique; then next our model finds the malignancy of detected mass with more than 99% negative predictive value for ovarian cancer to practically assist clinicians’ further investigation via combined features of exon-exon junctions, and hematology parameters. We diverged from traditional methods by employing intron-spanning reads (ISR) counts rather than gene expression levels to use splice junctions as features in our models. If integrated with current screening methods, our algorithm holds promise for identifying ovarian or endometrial cancer in its early stages.
Background: Ovarian cancer is the most lethal gynecologic malignancy, and CA-125 has limited early-stage sensitivity. We developed and evaluated Foretell Ovary, an ensemble machine-learning classifier that integrates tumor-educated-platelet RNA markers and routine hematologic parameters for non-invasive ovarian cancer detection, and compared it head-to-head with CA-125. Methods: A multi-site prospective cohort study enrolled 460 South Korean participants (67 ovarian cancer, 15 borderline ovarian tumors, 197 benign gynecological conditions, and 181 asymptomatic controls) at six clinical sites between September 2024 and March 2026. The cohort was split into training (n = 279) and independent test (n = 181) sets under a pre-specified ACTB cycle threshold ≤ 28 per-protocol rule (test n = 169). The classifier blended a random forest and two calibrated logistic-regression sub-models built on eight platelet RNA markers, 24 complete blood count parameters, and 30 marker-by-hematology combination features. Performance was assessed at a pre-frozen cutoff of 0.44 and compared against CA-125 (35 U/mL cutoff) in a matched subset (n = 100). Results: Foretell Ovary achieved an area under the receiver operating characteristic curve (AUC) of 0.923, sensitivity of 87.5% (100.0% for Stage II–IV), specificity of 89.0%, and negative predictive value of 97.7% in the test set. In the matched subset, Foretell Ovary outperformed CA-125 on every metric (AUC 0.839 vs. 0.723; sensitivity 85.0% vs. 65.0%), with the largest gap in pre-menopausal participants (+25 percentage points in sensitivity). A three-tier risk-grading scheme retained no ovarian cancer in the low-risk grade. Conclusions: Foretell Ovary offers more accurate non-invasive detection of ovarian cancer than CA-125, particularly in pre-menopausal women, offering the flexibility to be used alone or combined with CA-125. Prospective validation studies are warranted.