Introduction & Objectives
Circulating tumor DNA (ctDNA) is a substantial prognostic biomarker in many cancers. In urothelial carcinoma, its use is limited due to tumoral genetic heterogeneity. Methylation of tumor suppressor genes promotors being an early event in carcinogenesis, we aimed to identify a methylation signature that allows plasmatic and urinary ctDNA detection in invasive urothelial tumors, to help riskassessment and disease-monitoring.
Materials & Methods
Identification of targets of interest was realized by a bioinformatic analysis of methylation data sets from TCGA. A digital droplet PCR assay (ddPCR) was developed to detect these targets with high sensitivity in cell-free DNA extracted from tissue, plasma and urine samples from a monocentric observational cohort of patients with localized or locally advanced invasive urothelial carcinoma.
Results
We developed a ddPCR test detecting three targets that are differentially hypermethylated in urothelial carcinoma. The test was considered positive if at least one of the targets was detected, allowing 100% sensitivity and 77% specificity on tumor tissue (compared to adjacent non-tumor tissue). The cohort included 93 patients. 78% were male, average age 73. 80% were treated for bladder cancer, 40% had a neoadjuvant chemotherapy. In plasma samples, ctDNA was detected in 36% of chemotherapy-naive patients. ctDNA detection before surgery was associated with lymph node invasion, and poor prognostic in progression-free survival (HR 3.26 [1.64 – 6.50]) and in overall survival (HR 2.50 [1.26 – 4.97]). In the group receiving NAC, prognostic association was even stronger (PFS: HR 18.2, OS: HR 6.04). ctDNA detection after surgery was associated with metastatic relapse. Urinary ctDNA (UtDNA) was detected in 90% of urine samples at diagnosis. UtDNA detection and concentration levels were associated with tumor stage and with histological response to chemotherapy.
Conclusions
We developed a simple, sensitive ddPCR test that allows ctDNA detection in plasma and urine samples of patients treated for invasive urothelial carcinoma. This test might permit to use ctDNA as a prognostic and predictive tool in everyday practice. The limit of this study is the lack of regularity of sampling, as the cohort was observational. A validation cohort with patients included in an interventional study is in progress and will help to confirm our results.
