Association of circulating tumor DNA with patient prognosis in surgically resected renal cell carcinoma

Abstract

Background: Despite complete resection, 20%-50% of patients with localized renal cell carcinoma (RCC) experience recurrence within 5 years. Accurate assessment of prognosis in high-risk patients would aid in improving outcomes. Here we evaluate the use of circulating tumor DNA (ctDNA) in RCC using banked samples and clinical data from a single institution.

Methods: The cohort consisted of 45 RCC patients (≥pT1b) who underwent complete resection. The presence of ctDNA in plasma was determined using a personalized, tumor-informed ctDNA assay (Signatera RUO, Natera, Inc.). Relationships with outcomes and other relevant clinical variables were assessed. The median follow-up was 62 months.

Results: Plasma ctDNA was detected in 18 out of 36 patients (50%) pre-operatively and was associated with increased tumor size (mean 9.3 cm vs. 7.0 cm, P < .05) and high Fuhrman grade (60% grades III-IV vs 27% grade II, P = .07). The presence of ctDNA, either pre-operatively or at any time post-operatively, was associated with inferior relapse-free survival (HR = 2.70, P = .046; HR = 3.23, P = .003, respectively). Among patients who were ctDNA positive at any time point, the sensitivity of relapse prediction was 84% with a PPV of 90%. Of note, ctDNA positivity at a post-surgical time point revealed a PPV of 100% and NPV of 64%. The lack of ctDNA detection at both time points yielded an NPV of 80%.

Conclusions: Detection of plasma ctDNA using a personalized assay is prognostic of recurrence in patients with resected RCC. Herein, we describe a successful approach for its application and identify potential limitations to be addressed in future studies.

Keywords: circulating tumor DNA; liquid biopsy; recurrence; renal cell carcinoma.

Figure 1.

Genetic variants most frequently observed in the patient cohort (N = 48) with renal cell carcinoma. (A) Onco‐plot summarizing clinical and genomic features of 48 patients with measurable disease. The top bar graph shows tumor mutation burden (Mutations/Mb) for each patient. The bar graph on the right of the onco-plot represents the mutation frequency of mutation type for each gene in this cohort. Green represents nonsense mutations, red—in frame indels, yellow—frameshift indels, blue—missense mutations, orange—splice site, light blue—translation start site, and purple—silent variants. (B) Clinical characteristics (stage, grade, relapse, and ctDNA detection status pre- and post-surgery).

Figure 2.

ctDNA status and clinicopathological characteristics. The presence of ctDNA in pre-operative plasma is associated with (A) increased tumor size but not (B) high-grade disease. Chi-square test was used to analyze the association between ctDNA status and clinicopathological variables.

Figure 3.

ctDNA status and disease progression. The presence of ctDNA in pre- and post-operative plasma samples is associated with reduced relapse-free survival (A, C) and overall survival (B, D), respectively. (E, F) The presence of ctDNA in plasma samples either prior to primary resection or during surveillance is associated with reduced relapse-free survival and overall survival. Analysis was performed using a log-rank test, and the number of patients at risk at any given time point is indicated.