Abstract 4919: Evaluating the Gene Expression of Frozen Tissue-Derived Prognostic Signatures in FFPE Colorectal Cancer Samples
Jing Zhu,Natasha G. Deane,Keeli B. Lewis,Chandrasekhar Padmanabhan,Mary K. Washington,Kristen K. Ciombor,Cynthia Timmers,Richard M. Goldberg,R. Daniel Beauchamp,Xi Chen
DOI: https://doi.org/10.1158/1538-7445.am2016-4919
IF: 11.2
2016-01-01
Cancer Research
Abstract:Abstract OBJECTIVE: Defining molecular features that can predict the recurrence of colorectal cancer (CRC) for stage II and stage III patients remains a challenging problem in cancer research. Most available clinical samples are Formalin Fixed and Paraffin Embedded (FFPE). Nanostring nCounter® and Affymetrix GeneChip® Human Transcriptome Array 2.0 (HTA) are the two platforms marketed for high throughput measurement of mRNA expression from FFPE tissue samples. In this study, to identify an optimal platform for the gene expression profiling of FFPE CRC samples, we compared the expression of genes that make up published frozen tissue-derived prognostic signatures measured by these two platforms. METHODS: FFPE primary tumor tissue blocks were identified from 500 patients with stage II or III CRC and complete pathological information. We removed samples with neoadjuvant therapy, inadequate clinical follow-up or tumor specimens, samples with failure of RNA extraction and highly degraded samples. We identified 194 eligible FFPE-derived CRC primary tumors with a 7.4-year mean follow up. To measure the gene expression of the 194 samples, we designed a custom nCounter codeset using 536 gene elements from multiple published frozen tissue-derived prognostic signatures for CRC. We also performed gene expression profiling using the HTA platform on a subset of 84 of the 194 samples. For the nCounter data, samples with low total average count or with more than 20% genes having lower average count than the synthetic negative control genes were omitted to ensure adequate data quality. RESULTS: In total, for CRC samples from 42 patients, the gene expression data of 516 common genes measured by both platforms were of sufficient quality for comparative analysis. Based on HTA platform-derived data, we found that (1) 36 and 97 of the 516 common genes are significantly associated with overall survival (OS) and disease-free survival (DFS) at the single gene level (FDR < 0.05), respectively; and (2) two of the nine reported multi-gene signatures, for which sufficient information from the original papers enabled such evaluation, can divide patients into high-risk and low-risk groups with significantly different DFS (p ≤ 0.05). Based on nCounter platform-derived data, no individual gene was found to be significantly associated with survival at the single gene level (FDR < 0.05), but one of the nine published multi-gene signatures can divide patients into two groups with significantly different OS (p ≤ 0.05). Our results also showed moderate correlations between paired FFPE samples measured by nCounter and HTA platforms (the median correlation coefficient is 0.52). CONCLUSION: Our data demonstrate that while both platforms may identify gene or gene set expression differences associated with survival outcomes, the HTA appears to provide a more robust gene expression analysis dataset when using genes selected from published gene signatures. Citation Format: Jing Zhu, Natasha G. Deane, Keeli B. Lewis, Chandrasekhar Padmanabhan, Mary K. Washington, Kristen K. Ciombor, Cynthia Timmers, Richard M. Goldberg, R. Daniel Beauchamp, Xi Chen. Evaluating the gene expression of frozen tissue-derived prognostic signatures in FFPE colorectal cancer samples. [abstract]. In: Proceedings of the 107th Annual Meeting of the American Association for Cancer Research; 2016 Apr 16-20; New Orleans, LA. Philadelphia (PA): AACR; Cancer Res 2016;76(14 Suppl):Abstract nr 4919.