Abstract 7073: Development and validation of a high-specificity PCR algorithm for ovarian cancer diagnosis

Abstract Ovarian cancer (OC) remains one of the most lethal gynecologic malignancies, largely due to difficulties in achieving early diagnosis. To address the limitations of next-generation sequencing (NGS) platforms, particularly their high cost, we developed a qPCR-based method tailored for the early detection of ovarian cancer. Building on our previous research, we systematically selected and evaluated a refined panel of biomarkers optimized for qPCR detection. This study focuses on the development and validation of a highly specific PCR-based diagnostic algorithm, with particular emphasis on detecting high-grade serous ovarian cancer (HGSOC), the most common and aggressive subtype. A total of 30 samples were utilized to train the diagnostic model, supplemented by additional validation using QC-passed (5 benign and 1 cancer samples) and QC-failed samples (1 benign and 9 cancer samples). Biomarker selection prioritized transcripts with low expression in normal and benign tumor samples but significant outliers in ovarian cancer cases. A panel of 12 diagnostic markers and 1 reference marker was finalized based on statistical parameters such as variance, quantiles, and coverage of ovarian cancer-specific aberrations. PCR data were normalized using ACTB as the reference marker. Marker-specific cutoffs were determined from the training data, and a scoring system was devised to effectively differentiate ovarian cancer from non-OC samples. The algorithm achieved 100% specificity within the training set, missing one cancer sample out of 14, resulting in a sensitivity of 92.9%. Validation using QC-passed samples demonstrated a reasonable separation between OC and non-OC groups. The algorithm's robustness was further evaluated using QC-failed samples with challenges such as gDNA contamination and low RNA integrity (RIN), where performance for key markers was largely preserved. In the validation dataset, the algorithm missed one benign sample from the QC-passed group and one cancer sample from the QC-failed group, yielding an overall specificity of 83.3% and sensitivity of 90.0%. To address variability and refine score thresholds, synthetic datasets were generated, enabling the definition of "grey zone" ranges to manage overlapping distributions between OC and non-OC samples. The developed PCR-based diagnostic algorithm achieved exceptional specificity for ovarian cancer detection. This platform shows promise for clinical application in early ovarian cancer diagnosis, particularly for HGSOC, through precise molecular profiling of platelet RNA. Future work will focus on expanding sample diversity and exploring integration with broader biomarker panels. Citation Format: Eunyong Ahn, Se Ik Kim, Sungmin Park, Sarah Kim, Hyunjung Kim, Hyejin lee, Heeyeon Kim, TaeJin Ahn, Yong-Sang Song. Development and validation of a high-specificity PCR algorithm for ovarian cancer diagnosis [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2025; Part 1 (Regular Abstracts); 2025 Apr 25-30; Chicago, IL. Philadelphia (PA): AACR; Cancer Res 2025;85(8_Suppl_1):Abstract nr 7073.