Circulating tumor DNA (ctDNA) and correlations with clinical prognostic factors in patients with metastatic castration-resistant prostate cancer (mCRPC).

Abstract

186 Background: Sequencing of ctDNA is a minimally invasive method to study somatic DNA alterations. Quantitation and characterization of ctDNA may correlate with tumor burden and poor risk clinical features, but this is unproven in mCRPC. Methods: We performed deep targeted sequencing of 73 mCRPC-related genes in plasma ctDNA from 136 treatment naïve mCRPC patients. Fraction of ctDNA was determined by quantifying somatic mutant allele frequency and the deviation from heterozygosity for germline single nucleotide polymorphisms at regions of somatic copy number alterations. Baseline clinical factors were correlated with ctDNA fraction and genomic aberrations. Results: ctDNA could be quantified in 86/136 (63%) patients. Increasing fraction correlated with elevated ALP (chi2, p < 0.001), LDH (p < 0.001) and presence of liver metastasis (p = 0.01). Patients with any one of these factors were more likely to have a ctDNA fraction ≥ 40% (67% vs 36% of patients with neither factor, p = 0.001) and less likely to have unquantifiable ctDNA (26% vs 49%, p = 0.006). A decreased ctDNA fraction was associated with metastases only to lymph nodes (p = 0.012). Among the most frequently aberrant genes were Androgen Receptor (AR), p53, RB1, and PTEN, present in 38%, 30%, 23% and 22% of patients, respectively. Aberrations were associated with poor prognosis factors, including: AR and p53 alterations, which both correlated with presence of liver metastasis (p = 0.020 and p = 0.004 respectively), elevated LDH (p = 0.001, p < 0.001), and presence of ≥ 10 bone metastasis (p = 0.018, p = 0.017); PTEN deletion which correlated with elevated ALP (p < 0.001), elevated LDH (p = 0.001), ≥ 10 bone metastasis (p = 0.009) and time to CPRC of < 12 months (p = 0.021); and RB1 correlated with ≥ 10 bone metastasis (p = 0.009). Conclusions: We found an association between poor prognostic factors and increasing ctDNA fraction in patients with mCRPC. Genomic aberrations, particularly alteration in AR, p53 and PTEN, correlated with poor prognostic factors. These data show that clinical factors may help predict ctDNA yield in patients with mCRPC, and also provide insight into the role of deleterious genetic alterations.