A higher ctDNA fraction decreases survival in regorafenib-treated metastatic colorectal cancer patients. Results from the regorafenib's liquid biopsy translational biomarker phase II pilot study

Abstract

The predictive effect of circulating tumor DNA (ctDNA) in colorectal cancer (CRC) treatment is still highly discussed. The primary objective of our study was to investigate a possible prognostic/predictive value of ctDNA under regorafenib treatment. This prospective multicenter translational biomarker phase II pilot study enrolled 30 metastatic CRC patients (67% men, 33% women) treated with regorafenib. ctDNA was assessed in plasma before treatment start and at defined time points during administration. Measurement of tumor fraction as well as mutation and copy number analysis of CRC driver genes were performed by next-generation sequencing approaches. Multivariate analyses for survival and treatment efficacy were adjusted to age, gender and Eastern Cooperative Oncology Group. Disease control rate was 30%. Median tumor fraction at baseline was 18.5% (0-49.9). Mutations in CRC driver genes or genes involved in angiogenesis were identified in 25 patients (83.3%). KRAS mutations were detected in 13 of 14 KRAS-positive tumors; in three patients without KRAS mutation in the respective tumors, acquired mutations as a consequence of prior anti-EGFR treatment were detected. In a subset of patients, novel occurring mutations or focal amplifications were detected. A tumor fraction of 5% and higher at baseline was significantly associated with a decreased OS (P = .022; hazard ratio 3.110 (95% confidence interval: 1.2-8.2). ctDNA is detectable in a high proportion of mCRC patients. Higher ctDNA levels are associated with survival among regorafenib treatment. Moreover, our data highlight the benefit of a combined evaluation of mutations and somatic copy number alterations in advanced cancer patients.

Trial registration: ClinicalTrials.gov NCT01983098.

Keywords: circulating tumor DNA; metastatic colorectal cancer; prospective pilot study; regorafenib; sequencing.

Conflict of interest statement

All authors except G.P. and E.H. declare that there is no conflict of interest. G.P. received honoraria for advisory boards and speaker's fee from Taiko, Servier, Bayer, Terumo, Roche, Merck, Amgen, Sanofi, Lilly, BMS and MSD. E.H. received honoraria from Roche for advisory boards, not related to our study.

© 2020 The Authors. International Journal of Cancer published by John Wiley & Sons Ltd on behalf of Union for International Cancer Control.

Figures

FIGURE 1

Patient characteristics and mutation profile of the cohort. A, A total of 30 metastatic CRC treated with regorafenib were included in the study. Shown, in descending order, are sex, tumor localization and response to regorafenib based on the RECIST criteria (end of treatment [EOT] before computed tomography [CT], treatment was stopped due to side effects or progression before CT scan was scheduled), the RAS mutation status in tissue and plasma, the ichorCNA‐derived tumor traction (iTF) in %, the number of mutations identified with the SureSelect panel in the baseline samples, the presence of novel occurring mutation in the EOT samples and the ctDNA from baseline to EOT. B, Oncoprint of the most frequently mutated genes (top 10) in our cohort. Shown is an overview of genomic alterations (legend) in particular genes (rows) affecting individual samples (columns) [Color figure can be viewed at wileyonlinelibrary.com]

FIGURE 2

ctDNA levels at baseline. A, Distribution of various proxies for ctDNA levels prior to treatment initiation. iTF, tumor fraction calculated from ichorCNA; hVAF, highest variant allele frequency (VAF) identified with the SureSelect panel; aVAF, average VAF identified with the SureSelect panel; gw z‐score, genome‐wide z‐score calculated from mFAST‐SeqS. B, Linear regression of ctDNA levels calculated from ichorCAN compared to aVAF, hVAF and gw z‐scores [Color figure can be viewed at wileyonlinelibrary.com]

FIGURE 3

Kaplan–Meier survival analysis according to ichorCNA‐based tumor fraction (iTF) in metastatic CRC patients treated with regorafenib. Kaplan–Meier curves show that patients with iTF of (A) 5% and (B) 10% have significantly worse overall survival [Color figure can be viewed at wileyonlinelibrary.com]

FIGURE 4

ctDNA levels at baseline and end of treatment (EOT). Upper panel: Distribution of various proxies for ctDNA levels prior to treatment initiation (baseline) and at EOT including (A) iTF, tumor fraction calculated from ichorCNA; (B) gw z‐score, genome‐wide z‐score calculated from mFAST‐SeqS; (C) aVAF, average variant allele frequency (VAF) identified with the SureSelect panel. Lower panel: Changing levels of ctDNA reflected as (A) iTF, (B) gw z‐score and (C) aVAF [Color figure can be viewed at wileyonlinelibrary.com]