Genomic characterization of mutation hotspots related to chemotherapy responses in esophageal squamous-cell carcinoma with archived formalin-fixed paraffin-embedded tissues

Abstract

Introduction: Currently neoadjuvant chemoradiation therapy (CRT) prior to surgery is the most common treatment for Esophageal Squamous-cell Carcinoma (ESCC) patients. Given the fact that traditional TNM staging of the patients before treatment could not reflect their overall survival, a recent study conducted in Hong Kong demonstrated that the percentage of viable tumor cells remaining in the primary tumor after the CRT therapy was a prognostic marker for overall survival. Clinicians defined patients with no viable tumor cells (0%) as good responders, while the others having greater than 50% tumor cells remaining (50-100%) were considered poor responders. They observed that good responders show significantly better five-year survival rate. However, little is known about the mutation profiles between good and poor responders stratified in this way. Methods: Taking advantage of archived formalin-fixed paraffin-embedded (FFPE) tissues with comprehensive clinical outcomes, I apply a novel approach to systematically screen for possible somatic mutations that are associated with patients’ responses to CRT in ESCC. I obtained 100 archived FFPE blocks from Queen Mary Hospital, Hong Kong, 50 each representing either good responders or poor responders. DNA libraries were constructed to target 131 cancer-related genes with TOMA Os-seq, which adopted single-stranded ligation to minimize the loss of input FFPE DNA. Then, libraries were sequenced on Illumina HiSeq X Ten. The final list of somatic mutations was filtered with in-house pipeline, excluding both germline variants from public databases and in-house Hong Kong database. Copy number variations (CNVs) were also identified with a pre-designed control sample in the same sequencing batch. Results: One test run on samples with good quality DNA showed high sensitivity and specificity up to 100% for pre-designed mutations. For CNV, the sensitivity was 87.5% and the specificity was 97.5%. Trials of sequencing with several ESCC FFPE samples revealed deleterious mutations from genes including TP53, BRCA2, and those mutations were successfully validated by Sanger sequencing. Groups of critical protein-altered mutations and CNVs that distinguish good and poor responders of ESCC will be presented in detail. Conclusion: In all, FFPE blocks are valuable sources to uncover the biomarkers for CRT response in ESCC. I expect those biomarkers can aid future evaluation of newly diagnostic patients to avoid unnecessary efforts in treating patients that are unsuitable for the traditional therapy and thus facilitate effective treatments. Acknowledgements: Theme-based Research Scheme grant provided by the Hong Kong Research Grants Council (T12-701/17-R to MLL).