Circulating tumor DNA analysis to reveal genomic profiles for epithelial ovarian cancer.

Abstract

5543 Background: Circulating tumor DNA (ctDNA) analysis in epithelial ovarian cancer (EOC) was previously reported, however with limited samples or limited genes. Here, we reported an analysis of ctDNA in EOC cohort using targeted sequencing with a 1021-gene panel. Methods: Patients with EOC were enrolled, and treatment-naïve tumor tissues and blood samples were collected. We utilized a 1021-gene NGS panel in matched tissue DNA and ctDNA to identify somatic mutations with white blood cell DNA as a germline control. Results: Mutations were identified in all of the 65 tissues and in 53 (81.5%) ctDNA. The median ctDNA mutation allelic frequency was 2.5%, ranging from 0.1% to 36.2%. A median of 66.7% (12.5%-100.0%) of tissue derived mutations were observed in ctDNA. Besides, there were 91 ctDNA private mutations, including TP53 gene mutations. The most frequently mutated genes were TP53 (55.4%), PIK3CA (13.8%) and ARID1A (12.3%) in ctDNA analysis, which were consistent with tissue analysis (60.0%, 26.2% and 20.0% of tissues with TP53, PIK3CA and ARID1A mutations, respectively). Mutations of TP53 (37/42) in high-grade serous ovarian carcinoma (HGSOC), PIK3CA (10/11) and ARID1A (8/11) in ovarian clear cell carcinoma, BRAF (4/5) in low-grade serous ovarian carcinoma and PIK3CA (3/5), ARID1A (2/5) and PTEN (2/5) in endometrioid carcinoma were observed as the most commonly genetic aberrations in ctDNA in different sub-types of EOC, which located in different signal pathways and suggested different pathogenesis. In total, 90.5% (38/42) of HGSOC were ctDNA positive, comparing with 65.2% (15/23) of other EOC subtypes (p = 0.012). In addition, 56.5% (13/23) of stage I~II EOC were ctDNA positive, comparing with 94.7% (36/38) of stage III (p = 0.002). No association between ctDNA positivity and other clinic characteristics was observed, including pathological differentiation, CA125, lesion density (solid vs. cystic-solid and cystic). Multivariable analysis suggested FIGO stage III (p = 0.008) as an independent predictor of ctDNA detection. Conclusions: In summary, genomic characterization of EOC may offer insights into tumorigenesis and identify potential therapeutic targets in this disease.