Using postmortem formalin fixed paraffin-embedded tissues for molecular testing of sudden cardiac death: A cautionary tale of utility and limitations.

For archived cases of previously young healthy individuals where cause of sudden death remains undetermined, formalin fixed paraffin-embedded tissues (FFPE) samples are often the only biological resource available for molecular testing. We aim to ascertain the validity of postmortem molecular analysis of 95 cardiac genes using the FFPE samples routinely processed in the offices of medical examiners - typical fixation time in formalin ranges from days to months. The study was conducted in the College of American Pathologists accredited Molecular Genetics Laboratory within the City of New York Office of Chief Medical Examiner. Twelve cases, with FFPE samples and corresponding non-formalin fixed samples (RNAlater-preserved tissues or bloodstain card), were chosen for testing results comparison. The methods of extracting DNA from FFPE samples using Covaris, Qiagen, and Promega products showed comparable results. The quality of the extracted DNA, the target-enriched DNA libraries of 95 cardiac genes using HaloPlex Target Enrichment system by Agilent Technologies, and sequencing results using Illumina Miseq instrument were evaluated. Compared to the sequencing results of the nonfixed samples, the FFPE samples were categorized into three groups: 1) Group 1 samples fixed in formalin 2-6 days, had greater than 55 % sequencing regions ≥30x and 94%-100% variant concordance. 2) Group 2 samples fixed in formalin for 8 days, showed intra-sample sequencing variations: the surface tissues showed 25%-27% extra variants (false positive) and 8.1%-9.7% missing variants (false negative), whereas the repeated core tissues showed reduced extra variants to 1.6 % and the false negative error was unchanged. 3) Group 3 samples fixed in formalin 29-136 days, had 2-55 % sequencing regions ≥30x, up to 52.2 % missed variants and up to 6.3 % extra variants. All reportable variants (pathogenic, likely pathogenic or variant of uncertain significance) identified in the nonfixed samples were also identified in FFPE, albeit three variants had low confidence variant calling. In summary, our study showed that postmortem molecular diagnostic testing using FFPE samples routinely processed by the medical examiners should be cautioned, as they are replete with false positive and negative results, particularly when sample fixation time is longer than 8 days. Saving non-formalin fixed samples for high fidelity molecular analysis is strongly encouraged.