Utility of ctDNA in predicting outcome and pathological complete response in patients with bladder cancer as a guide for selective bladder preservation strategies.

Abstract

563 Background: Muscle-invasive bladder cancer (MIBC) accounts for ~25–30% of all bladder cancer diagnoses. With neoadjuvant chemotherapy (NAC) followed by radical cystectomy (RC) as standard-of-care, the 5-year survival rate ranges from 40%–60%. Bladder-sparing protocols (BSP) have emerged as a feasible alternative to RC for MIBC treatment, however better tools are needed. In this study, we evaluated the prognostic value of circulating tumor DNA (ctDNA) in predicting recurrence in patients who achieved pathological complete response (pCR). Methods: We analyzed a previously described cohort of 68 patients (656 plasma samples; Christensen et al., JCO 2019) with MIBC who received NAC prior to cystectomy. Patients had an updated median follow-up of 58.94 months (range: 7.19-81.77) post-cystectomy. ctDNA was analyzed at baseline (before NAC; N=64), and prior to cystectomy (N=65) using a commercially available assay (Signatera, Natera, Inc.). Additionally, exploratory RNA-Seq was performed on tumors from 59 patients (samples with >5M total counts were utilized). Pathway analysis was used to compare ctDNA-positive and ctDNA-negative patients who failed to achieve pCR. Results: Of the 64 patients with ctDNA results available at baseline, 59.4% (38/64) tested ctDNA-negative, and of these 84.2% (32/38) achieved pCR. Furthermore, 40.6% (26/64) tested ctDNA-positive, and only 34.6% (9/26) achieved pCR. Likewise, prior to cystectomy, 83.9% (52/62) of patients were ctDNA-negative, and 80.7% (42/52) achieved pCR, while none of the ctDNA-positive patients achieved pCR (positive predictive value 100%; negative predictive value 80.8%). Based on both ctDNA timepoints, the probability of ctDNA-negative patients to achieve pCR was significantly higher than ctDNA-positive patients ( p<0.0001). Notably, ctDNA-positive patients without pCR demonstrated significantly poorer RFS and OS compared to the ctDNA-negative patients, at both timepoints (Baseline: RFS; HR=8.2, p=0.017, OS; HR=8.4, p=0.015, prior to cystectomy: RFS; HR=5.2, p=0.0078, OS; HR=4.8, p=0.012). Furthermore, ctDNA status at baseline and before cystectomy was a better predictor of RFS compared to pCR (HR=8.5, p<0.0001, HR=14, p<0.0001, respectively). Transcriptomic pathway analysis of patients who did not achieve pCR showed an enrichment of oncogenic pathways, namely EMT and angiogenesis, in tumors from ctDNA positive patients, whereas tumors from ctDNA negative patients showed an enrichment of anti-tumor immune signatures, including IFNα and IFNγ. Conclusions: Absence of ctDNA was significantly associated with pCR both at baseline and prior to cystectomy, identifying patients who may benefit from BSP. Larger cohorts are warranted to test the prognostic value of ctDNA combined with transcriptomic profiling in informing patient selection for avoiding cystectomy.