The effect of tumor purity on next generation sequencing of colorectal cancer.

Abstract

e15557 Background: Considerable advancements have been made in tumorigenesis, therapeutic target and diagnostic markers of colorectal cancer by next generation sequencing (NGS). However, almost tumor tissues contain the non-cancerous components, such as inflammatory cells and stromal cells. The DNA from these cells may cause the fraction of reads containing somatic mutations from tumor cell decreases. Therefore, it’s necessary to systematically clarify the relationship between tumor purity and NGS in colorectal cancer. Methods: Formalin fixed and paraffin embedded (FFPE) blocks were collected and cut into 20-40 8 um-thick sections used for genomic DNA extraction. Sections were evenly divided into two groups, purity group and non-purity group. In purity group, the sections were stained with Histogene (Life Technologies), a PCR-compatible tissue stain. And the tumor samples were accurately collected on the section under the microscope. No special treatment before sequencing in non-purity group. Next-generation sequencing (NGS) based on a 425-gene panel was performed in tumor samples and matched plasma samples. Results: 30 samples were included in this study. 687 non-synonymous mutations were found in purity group, 271 mutations were found in non-purity group, including 247 common mutations in both group, 440 purity-specific mutations (false negative mutations), and 24 non-purity-specific mutations (false positive mutations). The Variant allele fraction (VAF) of common mutations in purity group was significantly higher than that in non-purity group (Wilcoxon test: P< 0.05). The number of mutations in purity group was higher than that in non-purity group, but no significant difference between them (10.5[3, 154] vs 9[3-90], Wilcoxon test: P> 0.05). The false positive ratio (the number of false negative mutations/the number of common mutations) was negatively correlated with tumor purity (Spearman rank correlation test: r = 0.4280, P = 0.02596). When the tumor purity below 30%, 86% false negative mutations belong to high-frequency mutation (VAF > 10%) in purity group. Only 8% false negative mutations belong to high-frequency mutation when purity above 30%. The false positive ratio (the number of false positive mutations/the number of common mutations) wasn’t correlated with tumor purity (Spearman rank correlation test: r = 0.1387, P= 0.4648). Conclusions: Tumor purity is an importance role in precise NGS of colorectal cancer. Higher tumor purity can ensure the integrity of NGS.