1162P - Analytical performance of a new liquid biopsy mutation panel for detection of clinically actionable variants

Date 10 October 2016
Event ESMO 2016 Congress
Session Poster display
Presenter Christer Svedman
Citation Annals of Oncology (2016) 27 (6): 401-406. 10.1093/annonc/mdw380
Authors C. Svedman1, G. Alexander2, A. Bergamaschi2, J. Han2, P.B. Harrington2, C. Ku2, Y. Ma2, W. Gibb2, A. Dei Rossi2, L. Shen2, A.D. Goddard2, D.A. Eberhard2, K.M. Clark-Langone2
  • 1Research And Development, Genomic Health Inc, 94063 - Redwood City/US
  • 2Research And Development, Genomic Health Inc, Redwood City/US

Abstract

Background

Assessing genomic alterations in circulating tumor DNA (ctDNA) from liquid biopsies may better reflect tumor heterogeneity, facilitate monitoring of tumor evolution and overcome the challenges of obtaining tissue biopsies. Analytic performance of such assays should be established on a per sample basis, using clinically relevant variants at levels representative of ctDNA. Here we report the analytic performance of a 17-gene panel (Oncotype SEQ™, Liquid Select) and illustrate its ability to detect ctDNA in cancer patients.

Methods

Analytical specificity and sensitivity were characterized through determination of Limit of Blank (LoB) and Limit of Detection (LoD) respectively. 73 cell-free DNA (cfDNA) samples from 60 healthy donors were used to determine the LoB and set detection thresholds. A model system using cell line DNA harboring clinically actionable variants was then used to determine the allele fraction (AF)/copy number (CN) required for a 95% rate of detection (LoD), using 30-50ng DNA input. Repeatability and reproducibility was assessed using pools of cfDNA from cancer patients. Finally, liquid biopsies from 15 stage II-IV cancer patients (on or after therapy) were assayed for genomic alterations.

Results

Detection thresholds were set above the LoB corresponding to >99% per sample specificity. LoD was calculated using 105 samples for each variant tested. Mean LoDs were as follows; single nucleotide variants (SNVs), 0.56% AF; insertions/deletions (indels), 0.19% AF; fusions, 0.37% AF, and CN gain, 2.7 copies. Accuracy was verified using additional variant positive and variant negative standards. In the repeatability and reproducibility study using cfDNA pools, on average >95% of expected variants were detected in each run. 10 SNVs and 2 indels were found in the 15 patient plasma samples, ranging from

Conclusions

The 17-gene panel (Oncotype SEQ™, Liquid Select) provides high sensitivity, detecting ctDNA at

Clinical trial identification

Legal entity responsible for the study

Genomic Health

Funding

Genomic Health

Disclosure

C. Svedman: Employee of Genomic Health. G. Alexander, A. Bergamaschi, J. Han, P.B. Harrington, Y. Ma, A. Dei Rossi, L. Shen, A.D. Goddard, D.A. Eberhard, K.M. Clark-Langone: Stock ownership and employee of Genomic Health. C-J. Ku: Chin-Jen Ku. W. Gibb: Stock ownership and employee of Genomic Health.