Validation of a Next-Generation Sequencing Assay for Clinical
Total Page:16
File Type:pdf, Size:1020Kb
The Journal of Molecular Diagnostics, Vol. -, No. -, - 2013 1 63 2 64 3 65 4 66 5 67 6 jmd.amjpathol.org 68 7 69 8 70 9 71 10 72 11 73 12 Validation of a Next-Generation Sequencing Assay for 74 13 75 14 Clinical Molecular Oncology 76 15 77 y 16 Q32 Catherine E. Cottrell,* Hussam Al-Kateb,* Andrew J. Bredemeyer,* Eric J. Duncavage,* David H. Spencer,* Haley J. Abel, 78 17 y y y 79 Christina M. Lockwood,* Ian S. Hagemann,* Stephanie M. O’Guin, Lauren C. Burcea, Christopher S. Sawyer, 18 y y y y 80 Dayna M. Oschwald, Jennifer L. Stratman,* Dorie A. Sher,* Mark R. Johnson, Justin T. Brown, Paul F. Cliften, Bijoy George,* 19 y 81 20 Leslie D. McIntosh,* Savita Shrivastava,* TuDung T. Nguyen,* Jacqueline E. Payton,* Mark A. Watson,* Seth D. Crosby, 82 Richard D. Head,y Robi D. Mitra,y Rakesh Nagarajan,* Shashikant Kulkarni,*y Karen Seibert,* Herbert W. Virgin IV,* 21 y 83 22 Jeffrey Milbrandt, and John D. Pfeifer* 84 23 85 24 y 86 25 From the Departments of Pathology and Immunology,* and Genetics, Genomics and Pathology Services, Washington University School of Medicine, 87 St. Louis, Missouri 26 88 27 89 28 Accepted for publication 90 Currently, oncology testing includes molecular studies and cytogenetic analysis to detect genetic aberrations 29 October 1, 2013. 91 of clinical significance. Next-generation sequencing (NGS) allows rapid analysis of multiple genes for clin- 30 Address correspondence to 92 ically actionable somatic variants. The WUCaMP assay uses targeted capture for NGS analysis of 25 cancer- 31 Hussam Al-Kateb, Ph.D., 93 Q26 associated genes to detect mutations at actionable loci. We present clinical validation of the assay and a 32 Washington University School 94 detailed framework for design and validation of similar clinical assays. Deep sequencing of 78 tumor spec- 33 of Medicine, 660 S. Euclid 95 imens (1000Â average unique coverage across the capture region) achieved high sensitivity for detecting 34 Ave., Box 8118, St. Louis, 96 MO 63110. E-mail: halkateb@ somatic variants at low allele fraction (AF). Validation revealed sensitivities and specificities of 100% for 35 97 path.wustl.edu. detection of single-nucleotide variants (SNVs) within coding regions, compared with SNP array sequence 36 98 data (95% CI Z 83.4e100.0 for sensitivity and 94.2e100.0 for specificity) or whole-genome sequencing 37 99 (95% CI Z 89.1e100.0 for sensitivity and 99.9e100.0 for specificity) of HapMap samples. Sensitivity for 38 100 detecting variants atan observed 10%AF was 100%(95%CI Z 93.2e100.0) in HapMap mixes. Analysis of 15 39 101 masked specimens harboring clinically reported variants yielded concordant calls for 13/13 variants at AF of 40 102 15%. The WUCaMP assay is a robust and sensitive method to detect somatic variants of clinical significance 41 103 in molecular oncology laboratories, with reduced time and cost of genetic analysis allowing for strategic 42 104 patient management. (J Mol Diagn 2013, -:1e17; http://dx.doi.org/10.1016/j.jmoldx.2013.10.002) 43 105 44 106 45 107 46 Q1 Traditional approaches to the genetic characterization of Increased access to sequencing technology and a decrease 108 47 clinical oncology specimens include cytogenetic analysis, in the associated costs have made it possible for clinical 109 48 fluorescence in situ hybridization (FISH), and molecular laboratories to develop testing strategies using NGS. Clinical 110 49 111 studies of single genes. These methodologies are comple- tests may be targeted to a panel of genes relevant to a given 50 112 51 mentary to each other and generate information of diagnostic phenotype or disease, or may be more broad in scope (eg, 113 52 and prognostic relevance. However, as new insight is gained whole-exome or whole-genome analyses). To date, most 114 53 into the complexities of cancer at the molecular level, the NGS clinical testing has focused on the detection of consti- 115 54 need emerges to obtain a more detailed cancer genetic profile tutional rather than somatic sequence variation, such as that 116 55 for improved patient management. As illustrated by recent reported in neuromuscular disease, mitochondrial disorders, 117 56 studies, identifying DNA mutations in cancer may aid in familial cancer syndromes, cardiomyopathy, ciliopathies, 118 57 understanding clonal evolution,1 risk stratification,2 and 119 58 therapeutic strategies.3,4 With the advent of next-generation 120 59 Supported by Genomics and Pathology Services, Washington University 121 sequencing (NGS), a more complete biological character- School of Medicine, St. Louis, MO. Q25 60 5 122 ization of a tumor can be attained at the molecular level. C.E.C., H.A.-K., and A.J.B. contributed equally to this work. 61 123 62 Copyright ª 2013 American Society for Investigative Pathology 124 and the Association for Molecular Pathology. Published by Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.jmoldx.2013.10.002 FLA 5.2.0 DTD JMDI284_proof 14 November 2013 2:15 pm EO: JMD13_0016 Cottrell et al 125 and familial hypercholesterolemia.6e12 Nonetheless, there is control (QC) directly. The U.S. Food and Drug Administration 187 126 an increasing role for NGS testing to direct the management (FDA) is likewise aware of the need for increased oversight 188 127 of oncology patients. For example, KRAS mutations in co- of NGS as a laboratory-developed test (http://www.fda.gov/ 189 128 190 dons 12 and 13 are observed in approximately 40% of MedicalDevices/NewsEvents/WorkshopsConferences/ucm 129 191 colorectal cancer cases,13 and these correlate with a poor 255327.htm, last accessed October 30, 2013); to date, how- Q4 130 192 14 fi 131 response to anti-EGFR antibody therapy. Likewise, the ever, the FDA has issued no speci c regulatory guidance. 193 e 132 detection of an EGFR exon 19 or 21 mutation in non small Thus, although several organizations have recognized the 194 133 cell lung cancer is correlated with sensitivity to EGFR tyro- need for formalized guidelines under which clinical NGS 195 15 134 sine kinase inhibitors, including gefitinib and erlotinib. can be performed, there remains a gap between the general 196 135 Concomitant NGS analysis of a select set of genes with requirements that have been published and detailed and 197 136 relevance across a broad scope of cancers increases the focused information regarding how those requirements 198 137 likelihood of detecting rare but clinically actionable variants should be satisfied in routine practice. It is understandable 199 138 (such as KIT mutations, which are present in <10% of thymic that clinical NGS laboratories may be struggling with 200 139 carcinomas16), is an aid in selecting therapeutics for tumors questions of what constitutes an appropriate approach to 201 140 202 harboring multiple genetic changes (such as combination these requirements, including documentation of the analyt- 141 203 therapies used for synergistic suppression), and allows for a ical wet-bench process used to generate NGS data and of the 142 204 143 tailored treatment regimen and more personalized patient bioinformatics used to support the analysis, interpretation, as 205 17 fl 144 care. Unlike single-gene tests, applied chie y to cancer well as reporting of NGS-based results as required by the 206 18e20 145 types commonly harboring mutations in one gene, the use of various regulatory bodies. Different laboratories address 207 146 NGS-based testing of multiple oncology targets also allows the regulatory requirements in various ways, and the discrete 208 147 for detection of rare and at times unexpected genetic varia- validation paradigms pursued by each laboratory need to be 209 148 tion. Moreover, the additive cost of performing multiple disseminated, so that their strengths and weaknesses can be 210 149 single-gene assays quickly exceeds that of multiplex testing evaluated. To this end, here we report the validation per- 211 150 but provides less in terms of cumulative information. formed at Genomics and Pathology Services at Washington 212 151 The emerging use of NGS approaches in clinical labo- University for an NGS test designed to detect sequence 213 152 214 ratories has brought increased interest in the development of variation within a targeted panel of actionable oncology genes 153 215 guidelines to ensure that NGS testing to direct patient care is (Figure 1), a validation process designed to meet the regula- ½F1 154 216 155 performed to the same rigorous standards as other clinical tory guidelines that have thus far been published. 217 156 tests focused on the analysis of nucleic acids, such as DNA 218 157 sequence analysis by Sanger methodology, DNA copy Materials and Methods 219 158 number analysis by microarray analysis, and detection of 220 159 chromosome aberrations by interphase FISH. To that end, WUCaMP Assay Design 221 160 several organizations have promulgated guidelines for 222 161 clinical NGS analysis. The College of American Patholo- The Washington University Cancer Mutation Profiling 223 162 gists (CAP) has released a checklist covering NGS.18 (WUCaMP) gene set targets oncology genes containing 224 163 Although the checklist addresses both the technical and known, clinically important variants. Genes were selected 225 164 226 bioinformatics components of NGS, it is structured as a based on the presence of described mutations with an 165 227 166 series of requirements, with little guidance as to how the established role for targeted therapy (or that affect response 228 167 requirements should be met in routine clinical practice. The to targeted therapy), an established role in current treatment 229 e 168 Next Generation Sequencing Standardization of Clinical paradigms, and a record of reimbursement when sequenced 230 169 Testing (Nex-StoCT) working group facilitated by the U.S.