Evaluation of TP53 mutations in Pap test and blood DNA as novel biomarkers of ovarian cancer risk

Abstract

<h3>Objectives:</h3> The use of Pap tests as a molecular diagnostic for high-grade serous ovarian carcinoma (HGSC) is a promising concept, but studies to detect tumor DNA report limited sensitivity. ‘Cancer-like' somatic mutations from clonal expansions in non-neoplastic tissue can be detected in Pap tests and blood, and could potentially be utilized for cancer risk-assessment. We aimed to determine whether <i>TP53</i> somatic mutations in blood and Pap test DNA are associated with HGSC or with inherited ovarian cancer susceptibility. <h3>Methods:</h3> Peripheral blood and Pap samples were collected from 26 women with HGSC (10 with germline <i>BRCA1</i> or <i>BRCA2</i> [<i>BRCA</i>] mutations) and 35 women without cancer (17 with germline <i>BRCA</i> mutations), all of whom underwent gynecologic surgery at a single academic institution. Blood samples alone were available for an additional four women. DNA was extracted from the blood buffy coat layer and Pap samples. The coding region of <i>TP53</i> was deeply sequenced (mean depth ~3800x) using duplex sequencing, an ultra-accurate error-correction sequencing approach. Mutations frequently found in the UMD <i>TP53</i> Cancer Database were termed ‘cancer-like'. Mutation frequencies (total coding mutations divided by nucleotides sequenced, MF) were compared between groups using t-tests, restricted to those between the ages of 40 and 60 for improved age-matching. Age-related trends were explored with fitted linear regression models. <h3>Results:</h3> 572 coding-region <i>TP53</i> mutations were identified in DNA from Pap tests from 26 women with HGSC (300 mutations) and 35 women without cancer (272 mutations). 287 mutations were identified in the DNA from blood buffy coat from 29 women with HGSC (161 mutations) and 36 without cancer (226 mutations). DNA from Pap tests demonstrated a higher MF in patients with HGSC than those without cancer (2.5x10^-6 vs. 1.6x10^-6, p=0.03). Cancer-like MF was also significantly higher in patients with HGSC (8.8x10^-7 vs. 5.0x10^-7, p=0.01). There were no significant differences in MF amongst cases without cancer with or without germline <i>BRCA</i> mutations, or amongst cases without cancer in those with <i>BRCA1</i> vs <i>BRCA2</i> germline mutations. Cancer-like mutations increased with age in Pap test samples from women without cancer (p<0.01, R2 0.34) but not in patients with HGSC (p=0.58, R2 0.01). In DNA from blood, there were no significant differences in average MF between HGSC and non-cancer cases nor between <i>BRCA</i> wildtype and mutant cases. No age-related correlations were found. Among twelve cases where the tumor mutation was known, duplex sequencing identified the mutation in five Pap samples (42%) and zero blood samples. <h3>Conclusions:</h3> Identifying the tumor-specific mutation in Pap tests, even with highly accurate DNA sequencing methods, does not reliably detect ovarian cancer. However, Pap tests carry abundant low frequency cancer-like <i>TP53</i> mutations, which are found in higher rates in women with HGSC than without. In patients without cancer, these <i>TP53</i> mutations increase linearly with age. Somatic <i>TP53</i> mutation frequency using age-related thresholds should be further explored as a biomarker for ovarian cancer risk.