Development and validation of a ultra-high depth FFPE targeted exome sequencing platform for routine cancer patient care. K. Chen, F. Meric-Bernstam, H. Zhao, Q. Zhang, N. Ezzeddine, L. Tang, P. Song, Y. Qi, Y. Mao, T. Chen, Z. Chong, W. Zhou, X. Zheng, A. Johnson, S. Kopetz, M. Davies, J. DeGroot, S. Moulder, K. Aldape, M. Routbort, R. Luthra, K. Shaw, J. Mendelsohn, G. Mills, A. Eterovic The University of Texas MD Anderson cancer center, Houston, TX.

   Although recent studies have indicated the potential of revolutionizing cancer patient care based on routine genomic sequencing, the community is still at an early stage of developing an optimized, practical approach that can deliver highly sensitive detection of actionable genomic aberrations at a reasonable cost. We established an ultra-deep targeted sequencing platform for 201 cancer genes to identify DNA alterations that would inform on disease status or could be potentially actionable in clinical tumor samples. Here we emphasize the significance and the direct correlation between depth of sequencing, input DNA and the ability to call rare mutational events with confidence and accuracy based on results in test samples. We then assayed 515 tumor samples and matched germline (blood) from 12 disease sites in 475 patients. With a mean haploid coverage of over 900x, we identified 4793 non-synonymous mutations, at a false discovery rate < 2.2%. Our results were highly concordant with test results obtained from a CLIA compliant hotspot panel, but identified alterations in potentially clinically actionable genes in twice as many patients. About 15.2% mutations are in low (< 10%) allele frequency with a landscape similar to that of high frequency mutations, which indicated their potential relevance for clinical decision making. In conclusion, our results indicated that an ultra-deep targeted sequencing approach can potentially impact both research and patient care through robust identification of low-level mutations that are potentially relevant for clinical decision making and that could be missed by other sequencing approaches that do not allow high depth of sequencing.

You may contact the first author (during and after the meeting) at