To date, ctDNA has been exhibiting its unprecedented translational potential in cancer care. However, accurate identification of comprehensive genomic alternations is rigorously needed for its clinical utility.
Three biologically relevant reference materials with allele frequencies expected at 0.1%, 1% and 5% were used to analytically evaluate the accuracy and reproducibility. Clinically, 43 ctDNA samples from lung, liver, colorectal, breast and gastric cancers were genomically profiled and compared to the known alternations in their matched solid tumors, in terms of single base substitution, insertions/deletions, copy number variations and rearrangement.
The analytical validation demonstrated unprecedented accuracy: near 100% specificity (99.6%, 99.9% and 100%) and 95.8%, 100% and 100% sensitivity for 3 reference materials, respectively. The actual detection limit was as low as 0.05%. The reproducibility was assessed as 0.998 (jaccard index) by sequencing 2 replicates of each reference. In clinical validation, compared to matched FFPE results, this ctDNA assay showed overall 99.9% specificity and 89% sensitivity, with > 90% sensitivity when only drug-gable hotspots were concerned. Eight events of gene rearrangements involving known targeted genes of ALK (n = 4), ROS1 (n = 3) and MET (n = 1) were detected from seven patients with 100% sensitivity and 100% specificity, confirmed either by IHC or panel sequencing (depth > 1,000X) over their matched FFPE biopsies. Due to its typically low abundance, CNV from ctDNA was detectable only for those highly amplified genes ( > =8 copies) with > =4 exons, demonstrating 72.2% sensitivity and 99.5% specificity.
Stringent criteria for both analytical and clinical validations are required for clinical utility of ctDNA. Our ctDNA assay has demonstrated high accuracy and reliability in comprehensively genomic profiling of ctDNA, especially in regard to druggable targets, which assures its translational utility in optimizing and monitoring targeted therapies in cancer management.
Clinical trial identification
Legal entity responsible for the study
W. Liu, S. Mu, J. Yao, H. Chen, Z. Hu, J. Hu, G. Chirn, H. Kang, K. Wang, M. Yao: Employee of OrigiMed Inc.