Next generation sequencing (NGS) has emerged as a primary tool for “precision” medicine, especially in the rapidly evolving field of cancer care. However NGS-based somatic variant calling in a clinical oncology setting is susceptible to numerous sources of technical error impeding optimal patient management. One important source of error in variant reporting is formalin fixation and paraffin embedding (FFPE). Of note, FFPE-derived errors are commonly present at clinically actionable loci and can occur at allelic frequencies (5-65%) that meet treatment criteria. Compounding this problem, degree of FFPE error is highly susceptible to inter-specimen and inter-laboratory variations. In this context, static variant calling thresholds to control for “general FFPE-effect” are likely inadequate; resulting in both False-Negatives (ignoring true somatic variation) and False-Positives (reporting technical error as truth). Despite the importance of fresh tissue for molecular studies, routine tissue histology (derived from FFPE) continues to be the gold standard for cancer staging/grading. And, as emphasis on smaller sample sizes becomes the norm, routine histology will be prioritized in many scenarios. Thus, the need to understand how molecular testing performs in FFPE tissue is a driving imperative for optimal cancer care.

In this study, our aim was to develop and evaluate “best” quality control practices that would minimize False-Positive or -Negative somatic variant calling in FFPE specimens. Reference samples (paired breast carcinoma and “normal” lymphoblast cell-lines) from the FDA-led Sequencing Quality Control Phase II (SEQC2) consortium were used. These specimens were subjected to varying FFPE conditions (n=8) mimicking the pathology laboratory setting. For each condition, multiple replicates of whole genome and exome sequencing (WGS/WES) were performed with ~100X coverage across the entire human genome. Observed somatic variation frequencies for each genomic position for each FFPE condition were compared to ground truth and quality control measures derived from matched fresh materials using numerous statistical models.

Applying rigorous analytical strategies, we aim to identify genomic regions that serve as markers (internal control regions) for the full range of FFPE-effect, develop QC measures and statistics to define lower limit of detection for variant calling for each genomic position, and provide general guidance for the community on how to accurately report mutations in FFPE processed clinical cancer specimens.

Citation Format: Thomas M. Blomquist, Wenming Xiao, Somatic Mutation Working Group of the SEQC2 Consortium. Comprehensive investigation of false mutation discoveries in FFPE samples [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2018; 2018 Apr 14-18; Chicago, IL. Philadelphia (PA): AACR; Cancer Res 2018;78(13 Suppl):Abstract nr 1427.