Rigorous Validation Of A Clinical Circulating Tumor Dna Assay For Cancer Molecular Profiling

Background: Profiling cell-free circulating tumor DNA (ctDNA) for genomic alterations which drive oncogenesis in patients with cancer promises to provide information important for understanding cancer biology, informing therapy selection when conventional FFPE biopsies are unobtainable and potentially monitoring response to therapy. To allow routine use of blood-based ctDNA molecular profiling with clinical samples we developed and performed analytic validation of an accurate, targeted NGS-based assay. The analytic validation included over 400 samples demonstrating ≥99% sensitivity and ≥99% positive predictive value for base substitutions, indels and rearrangements with limit-of-detection below 1%. Methods: To ensure robust performance, the ctDNA assay was developed as part of an integrated workflow including sample collection, storage and transport, and ctDNA purification, followed by optimized construction of adaptor-ligated sequencing libraries and enrichment by solution hybridization and then sequencing to high depth (Illumina HiSeq). Computational methodologies were developed to enable sensitive and specific detection of base substitutions, indels, genomic rearrangements and high-level amplifications from ctDNA. Accuracy and reproducibility were analytically validated in a CLIA-certified laboratory using reference samples with known alterations (117 cell-line mixtures and synthetic constructs) and 268 clinical ctDNA samples. Many alterations found in clinical ctDNA samples were validated with orthogonal reference methods including a CLIA-validated NGS assay, droplet digital PCR and break-point PCR. Results: The ctDNA assay validation demonstrated ≥99% sensitivity and ≥99% positive predictive value for base substitutions, indels and rearrangements with a limit-of-detection below 1% and robust detection of high-level, focal amplifications when present at adequate tumor fraction. In addition, the assay accurately reports the allele frequency of alterations in the sample. In 48 clinical ctDNA samples, 95 alterations of all classes were 100% confirmed by orthogonal testing. As part of our extensive clinical utility study, we report results comparing alterations from patient-matched ctDNA and FFPE biopsies across more than 200 lung, breast and other cancer samples. Conclusions: Accurate clinical profiling of ctDNA enables detection of genomic alterations in patient plasma samples to provide rationale targeted therapeutic options. Our rigorous analytic validation study demonstrates high-sensitivity detection of alterations present in blood at low frequency with a very low rate of false positives, realizing the potential of ctDNA-based molecular profiling for the management of patients with cancer. This validated assay allows us to embark upon a rigorous investigation of clinical best-practices based on tumor-type specific assessment of matched ctDNA and solid biopsy specimens. Citation Format: Travis A. Clark, Mark Kennedy, Jie He, Geneva Young, Mandy Zhao, Mike Coyne, Virginia Breese, Lauren Young, Shan Zhong, Mark Bailey, Bernard Fendler, Erica Schleifman, Eric Peters, Phil J. Stephens, Geoff A. Otto, Doron Lipson. Rigorous validation of a clinical circulating tumor DNA assay for cancer molecular profiling. [abstract]. In: Proceedings of the 107th Annual Meeting of the American Association for Cancer Research; 2016 Apr 16-20; New Orleans, LA. Philadelphia (PA): AACR; Cancer Res 2016;76(14 Suppl):Abstract nr 3965.