Liquid Biopsy Analysis of FGFR3 and PIK3CA Hotspot Mutations for Disease Surveillance in Bladder Cancer

Abstract

Background

Disease surveillance in patients with bladder cancer is important for early diagnosis of progression and metastasis and for optimised treatment.

Objective

To develop urine and plasma assays for disease surveillance for patients with FGFR3 and PIK3CA tumour mutations.

Design, setting, and participants

Droplet digital polymerase chain reaction (ddPCR) assays were developed and tumour DNA from two patient cohorts was screened for FGFR3 and PIK3CA hotspot mutations. One cohort included 363 patients with non–muscle-invasive bladder cancer (NMIBC). The other cohort included 468 patients with bladder cancer undergoing radical cystectomy (Cx). Urine supernatants (NMIBC n = 216, Cx n = 27) and plasma samples (NMIBC n = 39, Cx n = 27) from patients harbouring mutations were subsequently screened using ddPCR assays.

Outcome measurements and statistical analysis

Progression-free survival, recurrence-free survival, and overall survival were measured. Fisher's exact test, the Wilcoxon rank-sum test and Cox regression analysis were applied.

Results and limitations

In total, 36% of the NMIBC patients (129/363) and 11% of the Cx patients (44/403) harboured at least one FGFR3 or PIK3CA mutation. Screening of DNA from serial urine supernatants from the NMIBC cohort revealed that high levels of tumour DNA (tDNA) were associated with later disease progression in NMIBC (p = 0.003). Furthermore, high levels of tDNA in plasma samples were associated with recurrence in the Cx cohort (p = 0.016). A positive correlation between tDNA levels in urine and plasma was observed (correlation coefficient 0.6). The retrospective study design and low volumes of plasma available for analysis were limitations of the study.

Conclusions

Increased levels of FGFR3 and PIK3CA mutated DNA in urine and plasma are indicative of later progression and metastasis in bladder cancer.

Patient summary

Urine and plasma from patients with bladder cancer may be monitored for diagnosis of progression and metastasis using mutation assays.

Introduction

Bladder cancer is a common malignancy, with an estimated 430 000 new cases and 165 000 deaths worldwide every year [1]. Approximately 75% of newly diagnosed patients present with non–muscle-invasive bladder cancer (NMIBC), which is characterised by a high recurrence rate and 5-yr survival of ∼90% [2]. Progression to muscle-invasive bladder cancer (MIBC) is observed in 10–30% of patients [3]. The 5-yr survival rate for patients with MIBC is ∼50% [4], while ∼50% develop metastasis [5], which is associated with a 5-yr survival of only ∼5% [6]. Patients are monitored over many years to detect disease progression and metastatic disease, with high costs for health care systems and massive discomfort for the patients [2], [7]. Clinical and histopathologic information is used for risk stratification of patients to follow-up programs, and for more accurate identification of disease aggressiveness. Numerous promising biomarkers have been identified, including genomic alterations and transcriptional subtypes [8], [9], [10]. Frequent activating (hotspot) mutations have been identified in the TERT promoter and in FGFR3 and PIK3CA [11], [12]. Multiple studies have focused on hotspot mutations and methylation markers for disease surveillance using urine sediments, mainly from NMIBC patients [13], [14]. However, no molecular tests have yet been implemented in clinic practice.

Cell-free tumour DNA in liquid biopsies (urine supernatant and plasma) may have huge biomarker potential. Tumour DNA (tDNA) is continually shed into the circulation along with DNA from normal cells. Even though tDNA constitutes a small fraction of the total DNA in the circulation, its short half-life makes it a promising marker of the tumour burden [15]. Cell-free tDNA in urine supernatant may originate from lysed tumour cells, but may also originate from renal clearance of cell-free DNA in plasma [16]. In a recent study, we showed that tumour-specific structural variants can be detected in both plasma and urine supernatant, even in patients with NMIBC, and at high levels in patients before progression to MIBC [17]. Studies of breast and colon cancer, for example, have primarily focused on plasma analyses and have shown that tDNA in plasma effectively predicts disease recurrence [18], [19]. However, these studies have collectively depended on designing patient-specific assays or applying costly next-generation sequencing approaches.

Here we developed droplet digital polymerase chain reaction (ddPCR) assays for hotspot FGFR3 and PIK3CA mutation detection in liquid biopsies. Longitudinally collected urine supernatants from patients with NMIBC were analysed to predict disease progression to MIBC, and preoperative plasma and urine samples were analysed to predict disease recurrence following cystectomy. We demonstrate that prediction of aggressiveness and surveillance of relapse in patients with hotspot FGFR3 or PIK3CA mutations may be possible using urine and plasma.