Cell-free DNA copy number variations predict efficacy of immune checkpoint inhibitor-based therapy in hepatobiliary cancers

Abstract

Background: This study was designed to screen potential biomarkers in plasma cell-free DNA (cfDNA) for predicting the clinical outcome of immune checkpoint inhibitor (ICI)-based therapy in advanced hepatobiliary cancers.

Methods: Three cohorts including 187 patients with hepatobiliary cancers were recruited from clinical trials at the Peking Union Medical College Hospital. Forty-three patients received combination therapy of programmed cell death protein 1 (PD-1) inhibitor with lenvatinib (ICI cohort 1), 108 patients received ICI-based therapy (ICI cohort 2) and 36 patients received non-ICI therapy (non-ICI cohort). The plasma cfDNA and blood cell DNA mutation profiles were assessed to identify efficacy biomarkers by a cancer gene-targeted next-generation sequencing panel.

Results: Based on the copy number variations (CNVs) in plasma cfDNA, the CNV risk score model was constructed to predict survival by using the least absolute shrinkage and selection operator Cox regression methods. The results of the two independent ICI-based therapy cohorts showed that patients with lower CNV risk scores had longer overall survival (OS) and progression-free survival (PFS) than those with high CNV risk scores (log-rank p<0.01). In the non-ICI cohort, the CNV risk score was not associated with PFS or OS. Furthermore, the results indicated that 53% of patients with low CNV risk scores achieved durable clinical benefit; in contrast, 88% of patients with high CNV risk scores could not benefit from combination therapy (p<0.05).

Conclusions: The CNVs in plasma cfDNA could predict the clinical outcome of the combination therapy of PD-1 inhibitor with lenvatinib and other ICI-based therapies in hepatobiliary cancers.

Keywords: immunotherapy; liver neoplasms; tumor biomarkers.

Conflict of interest statement

Competing interests: None declared.

© Author(s) (or their employer(s)) 2021. Re-use permitted under CC BY. Published by BMJ.

Figures

Figure 1

Flow chart of the study design and the data selection process. A total of 471 patients with hepatobiliary cancers were screened. After excluding 284 patients with reasons, three cohorts including 187 patients with hepatobiliary cancers were recruited from clinical trials at the Peking Union Medical College Hospital. Forty-three patients received combination therapy of PD-1 inhibitor with lenvatinib (ICI cohort 1), 108 patients received ICI therapy (ICI cohort 2) and 36 patients received non-ICI therapy (non-ICI cohort). cfDNA, cell-free DNA; CNV, copy number variation; ICI, immune checkpoint inhibitor; LASSO, least absolute shrinkage and selection operator; PD-1, programmed cell death protein 1.

Figure 2

Construction and assessment of the copy number variation (CNV) risk score for hepatobiliary cancers. (A, E) The CNV risk scores of the patients in the discovery and validation cohorts sorted in ascending order. (B, F) Distributions of vital status for each patient according to the CNV risk score levels. (C, G) The copy number Z-scores and level of CNV risk scores of CALR, NR4A3, IDH2, IGF1R, ETV6, STAT3, NF2 and CTCF are shown in the heatmap. (D, H) The area under the curve (AUC) of the time-dependent receiver operating characteristic (ROC) curve was 0.881 in the discovery cohort and 0.907 in the validation cohort for the CNV risk score.

Figure 3

Association between the plasma cell-free DNA (cfDNA) copy number variation (CNV) risk score and the response to immune checkpoint inhibitor (ICI) therapy. (A and B) Kaplan-Meier curves for the overall survival (OS) and progression-free survival (PFS) of patients with hepatobiliary cancer in ICI cohort 1 stratified into high CNV risk and low CNV risk score groups. The PFS log-rank test showed p=0.0095; low versus high CNV risk, median PFS: 6.17 months vs 2.60 months, HR=0.045. The OS log-rank test showed p

Figure 4

Effect of the copy number variation (CNV) risk score on overall survival after combination therapy of programmed cell death protein 1 (PD-1) inhibitor with lenvatinib, by clinical characteristics. Forest plot for overall survival in subgroups. Estimates are based on a Cox proportional hazards model. AFP, alpha-fetoprotein; ECOG, Eastern Cooperative Oncology Group; PD-L1, programmed death-ligand 1; TBS, tumor burden score; TNM, tumor, node, metastases.

Figure 5

Association between the copy number variation (CNV) risk score of plasma cell-free DNA (cfDNA) and the response to immune checkpoint inhibitor (ICI)-based therapy. (A) CNV risk score levels of patients in the durable clinical benefit (DCB) and no durable benefit (NDB) groups from ICI cohort 1. (B) Proportional representation of DCB and NDB in ICI cohort 1 based on the level of the CNV risk score. (C) The CNV risk score levels of patients in the partial response (PR), stable disease (SD) and progressive disease (PD) groups from ICI cohort 1. (D) Proportional representation of the objective response rate (ORR) in ICI cohort 1 based on level of the CNV risk score. (E) CNV risk score levels of patients in the DCB and NDB groups from ICI cohort 2. (F) Proportional representation of DCB and NDB in ICI cohort 2 based on the level of the CNV risk score. (G) CNV risk score levels of patients in the PR, SD and PD groups from ICI cohort 2. (H) Proportional representation of the ORR in ICI cohort 2 based on the level of the CNV risk score.