Novel liquid biopsy assay to distinguish benign from malignant pelvic masses.

Abstract

e17563 Background: Despite not ranking among the top ten most common cancers in women in the United States, ovarian cancer stands as the fifth leading cause of cancer-related deaths in this population. According to the American Cancer Society, only 20% of ovarian cancers are found at stages I-II, when 94% of patients are expected to live longer than five years after diagnosis. Instead, the majority of ovarian cancers are detected at stages III-IV, when long term survival is 20% or less. Early identification of ovarian cancer is challenging due to the wide range of symptoms with which a patient may present, most of which lack specificity for ovarian cancer. Even in cases where a physician may want to rule in or out ovarian cancer, the available methods for screening have a low positive predictive value. These methods include bimanual exams, transvaginal ultrasounds, and CA-125 blood tests. In a screening capacity, the sensitivity for a bimanual exam finding ovarian cancer is only about 44%, transvaginal ultrasound is approximately 85%, and CA-125 testing registers at 77%. The need for better tools to detect ovarian cancer is clear. Genece Health has developed an assay utilizing a proprietary algorithm that leverages deep learning to analyze cfDNA fragment size, end motif, and coverage patterns, along with other genomic features, to detect the presence of ctDNA in blood originating from ovarian cancer. The data presented herein demonstrate that the Genece Health assay has the ability to detect ctDNA in a single tube of blood using low pass whole genome sequencing in a cost-effective way to distinguish between malignant and benign pelvic masses. Methods: Blood was prospectively collected from 56 females with pelvic masses in advance of removal and histopathology (approximately 32 cancer and 24 benign masses) in Streck cfDNA BCT Devices. Additionally, healthy female donor blood from 50 individuals procured through the San Diego Blood Bank were collected in Streck cfDNA BCT Devices. cfDNA was extracted from double-spun plasma and low-pass whole genome sequencing (2X – 5X coverage) was performed. Data were analyzed with the Genece Health Algorithm. Results: The Genece Health Algorithm demonstrated the ability to distinguish between pelvic masses that were deemed benign versus malignant via histopathology. There was a significant (P<0.01) difference between the classifier assigned to the malignant and benign and presumed normal groups. Conclusions: The investigational assay described herein shows progress toward the ability to use non-invasive molecular testing to identify ovarian cancer in a blood sample. A blood test to differentiate between a benign and malignant pelvic mass will enhance the ability of physicians to detect ovarian cancer at earlier stage, improving clinical outcomes for women. The cost-effective low-pass whole genome sequencing approach taken by Genece will be further explored in expanded cohorts.