Comprehensive genomic transcriptomic tumor-normal gene panel analysis for enhanced precision in patients with lung cancer
Metrics: PDF 1510 views | HTML 1572 views | ?
Shahrooz Rabizadeh1, Chad Garner1, John Zachary Sanborn1, Stephen C. Benz1, Sandeep Reddy2 and Patrick Soon-Shiong1,2
1NantOmics, LLC, Culver City, CA, USA
2NantHealth, Inc., Culver City, CA, USA
Shahrooz Rabizadeh, email: firstname.lastname@example.org
Keywords: precision medicine; tumor-normal sequencing; lung cancer; somatic variants; tumor sequencing
Received: November 30, 2017 Accepted: March 15, 2018 Published: April 10, 2018
A CMS approved test for lung cancer is based on tumor-only analysis of a targeted 35 gene panel, specifically excluding the use of the patient’s normal germline tissue. However, this tumor-only approach increases the risk of mistakenly identifying germline single nucleotide polymorphisms (SNPs) as somatically-derived cancer driver mutations (false positives). 621 patients with 30 different cancer types, including lung cancer, were studied to compare the precision of tumor somatic variant calling in 35 genes using tumor-only DNA sequencing versus tumor-normal DNA plus RNA sequencing. When sequencing of lung cancer was performed using tumor genomes alone without normal germline controls, 94% of variants identified were SNPs and thus false positives. Filtering for common SNPs still resulted in as high as 48% false positive variant calling. With tumor-only sequencing, 29% of lung cancer patients had a false positive variant call in at least one of twelve genes with directly targetable drugs. RNA analysis showed 18% of true somatic variants were not expressed. Thus, sequencing and analysis of both normal germline and tumor genomes is necessary for accurate identification of molecular targets. Treatment decisions based on tumor-only analysis may result in the administration of ineffective therapies while also increasing the risk of negative drug-related side effects.
All site content, except where otherwise noted, is licensed under a Creative Commons Attribution 3.0 License.