Abstract 4417
Background
Precision medicine is dependent on identifying actionable mutations in tumors. Accurate detection of mutations is often problematic in formalin-fixed paraffin-embedded (FFPE) tissues, as it causes DNA damage such as fragmentation and cytosine deamination. These Sequence artifacts can be difficult to distinguish from true mutations, and are an increasing interpretive issue. Understanding of the characteristics of these sequence artifacts in FFPE tissues is critical to improve the accurate detection of actionable mutations.
Methods
We reviewed the clinical courses of 156 small cell lung cancer (SCLC) patients who had undergone surgery at 17 institutions in Japan between January 2003 and January 2013. In these patients, we obtained the FFPE tissues of 79 cases which were histopathologically confirmed as SCLC and fitted for sequencing analysis with suitable DNA quality. Targeted amplicon sequence was conducted with MiSeq and TruSight panel (Illumina) which is a dual stranded amplicon kit for detecting cytosine deamination. We evaluated the characteristics of deamination bias and the relations with institutions and age of the tissue block.
Results
We could evaluate sequence of 73 samples data from 14 institutions. Target region of the sequencing was 26 genes, total 14686 bp. The total discordant single nucleotide variant (SNV) between forward and reverse strand were 690 cases, 16.4 cases per sample. The highest number of discordant SNV was 132 per sample. The most part of discordant SNV was the deamination change (C>T/G>A), 589 (85.4%) of 690 cases. The highest discordant SNV frequency was 0.25 with read depth 1876 in deamination change pattern, and 0.10 with read depth 4196 in the others. The frequency of the deamination change was different by institutions more than age of the tissue block.
Conclusions
Cytosine deamination from formalin fixation can be a major issue in diagnostic test of genome DNA for cancer samples. Procedures that assess, minimize or remove formalin-induced influences is important in the interpretation of genomic DNA analysis leading to better practice.
Clinical trial identification
Legal entity responsible for the study
Toraji Amano
Funding
None
Disclosure
All authors have declared no conflicts of interest.