Before anti-EGFR therapy is given to patients with colorectal cancer, it is required to determine KRAS mutation status in tumor. When tumor tissue is not available, cell-free DNA (liquid biopsy) is commonly used as an alternative. Due to the low abundance of tumor-derived DNA in cell-free DNA samples, methods with high sensitivity were preferred, including digital polymerase chain reaction, amplification refractory mutation system and next-generation sequencing. The aim of this systemic review and meta-analysis was to investigate the accuracy of those methods in detecting KRAS mutation in cell-free DNA sample from patients with colorectal cancer.

Literature search was performed in Pubmed, Embase, and Cochrane Library. After removing duplicates from the 170 publications found by literature search, eligible studies were identified using pre-defined criteria. Quality of the publications and relevant data were assessed and extracted thereafter. Meta-DiSc and STATA softwares were used to pool the accuracy parameters from the extracted data.

A total of 33 eligible studies were identified for this systemic review and meta-analysis. After pooling, the overall sensitivity, specificity, and diagnostic odds ratio were 0.77 (95%CI: 0.74-0.79), 0.87 (95%CI: 0.85-0.89), and 23.96 (95%CI: 13.72-41.84), respectively. The overall positive and negative likelihood ratios were 5.55 (95%CI: 3.76-8.19) and 0.29 (95%CI: 0.21-0.38), respectively. Area under curve of the summarized ROC curve was 0.8992.

Digital polymerase chain reaction, amplification refractory mutation system, and next-generation sequencing had overall high accuracy in detecting KRAS mutation in cell-free DNA sample. Large prospective randomized clinical trials are needed to further convince the accuracy and usefulness of KRAS mutation detection using cfDNA/liquid biopsy samples in clinical practice.

PROSPERO CRD42020176682; https://www.crd.york.ac.uk/prospero/display_record.php?RecordID=176682.