Portal de Periódicos da CAPES

Detection of Circulating Tumor DNA in Patients with Pancreatic Cancer Using Digital Next-Generation Sequencing

2020; Elsevier BV; Volume: 22; Issue: 6 Linguagem: Inglês

10.1016/j.jmoldx.2020.02.010

1943-7811

Anne Macgregor-Das, Jun Yu, Koji Tamura, Toshiya Abe, Masaya Suenaga, Koji Shindo, Michael Borges, Chiho Koi, Shiro Kohi, Yoshihiko Sadakari, Marco Dal Molin, Jose Alejandro Almario, Madeline Ford, Miguel Chuidian, Richard A. Burkhart, Jin He, Ralph H. Hruban, James R. Eshleman, Alison P. Klein, Christopher L. Wolfgang, Marcia I. Canto, Michael Goggins,

Renal cell carcinoma treatment

Circulating tumor DNA (ctDNA) measurements can be used to estimate tumor burden, but avoiding false-positive results is challenging. Herein, digital next-generation sequencing (NGS) is evaluated as a ctDNA detection method. Plasma KRAS and GNAS hotspot mutation levels were measured in 140 subjects, including 67 with pancreatic ductal adenocarcinoma and 73 healthy and disease controls. To limit chemical modifications of DNA that yield false-positive mutation calls, plasma DNA was enzymatically pretreated, after which DNA was aliquoted for digital detection of mutations (up to 384 aliquots/sample) by PCR and NGS. A digital NGS score of two SDs above the mean in controls was considered positive. Thirty-seven percent of patients with pancreatic cancer, including 31% of patients with stages I/II disease, had positive KRAS codon 12 ctDNA scores; only one patient had a positive GNAS mutation score. Two disease control patients had positive ctDNA scores. Low-normal–range digital NGS scores at mutation hotspots were found at similar levels in healthy and disease controls, usually at sites of cytosine deamination, and were likely the result of chemical modification of plasma DNA and NGS error rather than true mutations. Digital NGS detects mutated ctDNA in patients with pancreatic cancer with similar yield to other methods. Detection of low-level, true-positive ctDNA is limited by frequent low-level detection of false-positive mutation calls in plasma DNA from controls. Circulating tumor DNA (ctDNA) measurements can be used to estimate tumor burden, but avoiding false-positive results is challenging. Herein, digital next-generation sequencing (NGS) is evaluated as a ctDNA detection method. Plasma KRAS and GNAS hotspot mutation levels were measured in 140 subjects, including 67 with pancreatic ductal adenocarcinoma and 73 healthy and disease controls. To limit chemical modifications of DNA that yield false-positive mutation calls, plasma DNA was enzymatically pretreated, after which DNA was aliquoted for digital detection of mutations (up to 384 aliquots/sample) by PCR and NGS. A digital NGS score of two SDs above the mean in controls was considered positive. Thirty-seven percent of patients with pancreatic cancer, including 31% of patients with stages I/II disease, had positive KRAS codon 12 ctDNA scores; only one patient had a positive GNAS mutation score. Two disease control patients had positive ctDNA scores. Low-normal–range digital NGS scores at mutation hotspots were found at similar levels in healthy and disease controls, usually at sites of cytosine deamination, and were likely the result of chemical modification of plasma DNA and NGS error rather than true mutations. Digital NGS detects mutated ctDNA in patients with pancreatic cancer with similar yield to other methods. Detection of low-level, true-positive ctDNA is limited by frequent low-level detection of false-positive mutation calls in plasma DNA from controls. Pancreatic cancer incidence has been increasing in the United States and although survival rates have been improving slowly, for most patients it is still a deadly disease.1Rahib L. Smith B.D. Aizenberg R. Rosenzweig A.B. Fleshman J.M. Matrisian L.M. Projecting cancer incidence and deaths to 2030: the unexpected burden of thyroid, liver, and pancreas cancers in the United States.Cancer Res. 2014; 74: 2913-2921Crossref PubMed Scopus (4229) Google Scholar Recent improvements in early detection may be beginning to impact survival. For example, pancreatic imaging surveillance of individuals at sufficiently high risk of developing pancreatic cancer is associated with improved survival.2Canto M.I. Almario J.A. Schulick R.D. Yeo C.J. Klein A. Blackford A. Shin E.J. Sanyal A. Yenokyan G. Lennon A.M. Kamel I.R. Fishman E.K. Wolfgang C. Weiss M. Hruban R.H. Goggins M. Risk of neoplastic progression in individuals at high risk for pancreatic cancer undergoing long-term surveillance.Gastroenterology. 2018; 155: 740-751.e2Abstract Full Text Full Text PDF PubMed Scopus (198) Google Scholar,3Vasen H. Ibrahim I. Ponce C.G. Slater E.P. Matthai E. Carrato A. Earl J. Robbers K. van Mil A.M. Potjer T. Bonsing B.A. de Vos Tot Nederveen Cappel W.H. Bergman W. Wasser M. Morreau H. Kloppel G. Schicker C. Steinkamp M. Figiel J. Esposito I. Mocci E. Vazquez-Sequeiros E. Sanjuanbenito A. Munoz-Beltran M. Montans J. Langer P. Fendrich V. Bartsch D.K. Benefit of surveillance for pancreatic cancer in high-risk individuals: outcome of long-term prospective follow-up studies from three European expert centers.J Clin Oncol. 2016; 34: 2010-2019Crossref PubMed Scopus (217) Google Scholar Further improvements in the early detection of pancreatic cancer likely would occur if accurate blood tests were available to detect early stage disease. Many biomarker blood tests have been evaluated as candidate early detection tests for pancreatic cancer, including circulating tumor DNA (ctDNA). Molecular approaches for detecting ctDNA for patients with pancreatic cancer have focused on identifying KRAS hotspot mutations because more than 90% of pancreatic cancers have KRAS mutations and more than 95% of these mutations occur at codon 12. Circulating KRAS mutation concentrations in patients with pancreatic cancer are generally of low abundance and have been detected in only approximately 30% of patients with low-stage disease using existing methods.4Takai E. Totoki Y. Nakamura H. Morizane C. Nara S. Hama N. Suzuki M. Furukawa E. Kato M. Hayashi H. Kohno T. Ueno H. Shimada K. Okusaka T. Nakagama H. Shibata T. Yachida S. Clinical utility of circulating tumor DNA for molecular assessment in pancreatic cancer.Sci Rep. 2015; 5: 18425Crossref PubMed Scopus (128) Google Scholar, 5Berger A.W. Schwerdel D. Costa I.G. Hackert T. Strobel O. Lam S. Barth T.F. Schroppel B. Meining A. Buchler M.W. Zenke M. Hermann P.C. Seufferlein T. Kleger A. Detection of hot-spot mutations in circulating cell-free DNA from patients with intraductal papillary mucinous neoplasms of the pancreas.Gastroenterology. 2016; 151: 267-270Abstract Full Text Full Text PDF PubMed Scopus (61) Google Scholar, 6Brychta N. Krahn T. von Ahsen O. Detection of KRAS mutations in circulating tumor DNA by digital PCR in early stages of pancreatic cancer.Clin Chem. 2016; 62: 1482-1491Crossref PubMed Scopus (62) Google Scholar, 7Allenson K. Castillo J. San Lucas F.A. Scelo G. Kim D.U. Bernard V. Davis G. Kumar T. Katz M. Overman M.J. Foretova L. Fabianova E. Holcatova I. Janout V. Meric-Bernstam F. Gascoyne P. Wistuba I. Varadhachary G. Brennan P. Hanash S. Li D. Maitra A. Alvarez H. High prevalence of mutant KRAS in circulating exosome-derived DNA from early-stage pancreatic cancer patients.Ann Oncol. 2017; 28: 741-747Abstract Full Text Full Text PDF PubMed Scopus (285) Google Scholar, 8Pietrasz D. Pecuchet N. Garlan F. Didelot A. Dubreuil O. Doat S. Imbert-Bismut F. Karoui M. Vaillant J.C. Taly V. Laurent-Puig P. Bachet J.B. Plasma circulating tumor DNA in pancreatic cancer patients is a prognostic marker.Clin Cancer Res. 2017; 23: 116-123Crossref PubMed Scopus (158) Google Scholar, 9Sorber L. Zwaenepoel K. Deschoolmeester V. Roeyen G. Lardon F. Rolfo C. Pauwels P. A comparison of cell-free DNA isolation kits: isolation and quantification of cell-free DNA in plasma.J Mol Diagn. 2017; 19: 162-168Abstract Full Text Full Text PDF PubMed Scopus (128) Google Scholar, 10Kruger S. Heinemann V. Ross C. Diehl F. Nagel D. Ormanns S. Liebmann S. Prinz-Bravin I. Westphalen C.B. Haas M. Jung A. Kirchner T. von Bergwelt-Baildon M. Boeck S. Holdenrieder S. Repeated mutKRAS ctDNA measurements represent a novel and promising tool for early response prediction and therapy monitoring in advanced pancreatic cancer.Ann Oncol. 2018; 29: 2348-2355Abstract Full Text Full Text PDF PubMed Scopus (82) Google Scholar, 11Perets R. Greenberg O. Shentzer T. Semenisty V. Epelbaum R. Bick T. Sarji S. Ben-Izhak O. Sabo E. Hershkovitz D. Mutant KRAS circulating tumor DNA is an accurate tool for pancreatic cancer monitoring.Oncologist. 2018; 23: 566-572Crossref PubMed Scopus (54) Google Scholar, 12Hellwig S. Nix D.A. Gligorich K.M. O'Shea J.M. Thomas A. Fuertes C.L. Bhetariya P.J. Marth G.T. Bronner M.P. Underhill H.R. Automated size selection for short cell-free DNA fragments enriches for circulating tumor DNA and improves error correction during next generation sequencing.PLoS One. 2018; 13: e0197333Crossref PubMed Scopus (41) Google Scholar, 13Bernard V. Kim D.U. San Lucas F.A. Castillo J. Allenson K. Mulu F.C. Stephens B.M. Huang J. Semaan A. Guerrero P.A. Kamyabi N. Zhao J. Hurd M.W. Koay E.J. Taniguchi C.M. Herman J.M. Javle M. Wolff R. Katz M. Varadhachary G. Maitra A. Alvarez H.A. Circulating nucleic acids are associated with outcomes of patients with pancreatic cancer.Gastroenterology. 2019; 156: 108-118.e4Abstract Full Text Full Text PDF PubMed Scopus (203) Google Scholar, 14Cohen J.D. Javed A.A. Thoburn C. Wong F. Tie J. Gibbs P. et al.Combined circulating tumor DNA and protein biomarker-based liquid biopsy for the earlier detection of pancreatic cancers.Proc Natl Acad Sci U S A. 2017; 114: 10202-10207Crossref PubMed Scopus (332) Google Scholar, 15Pishvaian M.J. Joseph Bender R. Matrisian L.M. Rahib L. Hendifar A. Hoos W.A. Mikhail S. Chung V. Picozzi V. Heartwell C. Mason K. Varieur K. Aberra M. Madhavan S. Petricoin 3rd, E. Brody J.R. A pilot study evaluating concordance between blood-based and patient-matched tumor molecular testing within pancreatic cancer patients participating in the Know Your Tumor (KYT) initiative.Oncotarget. 2017; 8: 83446-83456Crossref PubMed Scopus (48) Google Scholar Methods to detect ctDNA include droplet digital PCR, safe-sequencing system (Safeseqs),16Kinde I. Wu J. Papadopoulos N. Kinzler K.W. Vogelstein B. Detection and quantification of rare mutations with massively parallel sequencing.Proc Natl Acad Sci U S A. 2011; 108: 9530-9535Crossref PubMed Scopus (810) Google Scholar targeted error correction sequencing (TEC-Seq),17Phallen J. Sausen M. Adleff V. Leal A. Hruban C. White J. et al.Direct detection of early-stage cancers using circulating tumor DNA.Sci Transl Med. 2017; 9Crossref PubMed Scopus (595) Google Scholar simple, multiplexed, PCR-based barcoding of DNA for sensitive mutation detection using sequencing (SiMSen-seq),18Stahlberg A. Krzyzanowski P.M. Jackson J.B. Egyud M. Stein L. Godfrey T.E. Simple, multiplexed, PCR-based barcoding of DNA enables sensitive mutation detection in liquid biopsies using sequencing.Nucleic Acids Res. 2016; 44: e105Crossref PubMed Scopus (86) Google Scholar and high multiplex amplicon barcoding PCR,19Peng Q. Vijaya Satya R. Lewis M. Randad P. Wang Y. Reducing amplification artifacts in high multiplex amplicon sequencing by using molecular barcodes.BMC Genomics. 2015; 16: 589Crossref PubMed Scopus (83) Google Scholar as well as in silico error correction strategies20Salk J.J. Schmitt M.W. Loeb L.A. Enhancing the accuracy of next-generation sequencing for detecting rare and subclonal mutations.Nat Rev Genet. 2018; 19: 269-285Crossref PubMed Scopus (249) Google Scholar and the detection of allelic imbalance after size selection.21Mouliere F. Chandrananda D. Piskorz A.M. Moore E.K. Morris J. Ahlborn L.B. et al.Enhanced detection of circulating tumor DNA by fragment size analysis.Sci Transl Med. 2018; 10Crossref PubMed Scopus (417) Google Scholar One method developed to detect low-abundance mutations is digital NGS.22Yu J. Sadakari Y. Shindo K. Suenaga M. Brant A. Almario J.A.N. Borges M. Barkley T. Fesharakizadeh S. Ford M. Hruban R.H. Shin E.J. Lennon A.M. Canto M.I. Goggins M. Digital next-generation sequencing identifies low-abundance mutations in pancreatic juice samples collected from the duodenum of patients with pancreatic cancer and intraductal papillary mucinous neoplasms.Gut. 2017; 66: 1677-1687Crossref PubMed Scopus (100) Google Scholar Digital NGS involves undertaking discrete NGS analyses on many (eg, 384) individual aliquots of DNA from a single biological sample in which each aliquot contains only a limited number of genome equivalents of DNA so that each aliquot has either zero or one mutation-containing DNA templates at each nucleotide of interest in addition to wild-type templates. The identification of recurrent mutations in more than one aliquot favors a true mutation over NGS-induced error. Although KRAS mutations are the major target of ctDNA detection strategies for pancreatic cancer, some studies have evaluated GNAS mutations as a potential ctDNA target because GNAS mutations commonly arise early in the development of the pancreatic cystic precursor neoplasm known as intraductal papillary mucinous neoplasm (IPMN). Hotspot mutations in GNAS (codon 201) are present in the neoplastic tissue of approximately 60% or more of IPMNs,23Wu J. Matthaei H. Maitra A. Dal Molin M. Wood L.D. Eshleman J.R. Goggins M. Canto M.I. Schulick R.D. Edil B.H. Wolfgang C.L. Klein A.P. Diaz Jr., L.A. Allen P.J. Schmidt C.M. Kinzler K.W. Papadopoulos N. Hruban R.H. Vogelstein B. Recurrent GNAS mutations define an unexpected pathway for pancreatic cyst development.Sci Transl Med. 2011; 3: 92ra66Crossref PubMed Scopus (616) Google Scholar,24Suenaga M. Yu J. Shindo K. Tamura K. Almario J.A. Zaykoski C. Witmer P.D. Fesharakizadeh S. Borges M. Lennon A.M. Shin E.J. Canto M.I. Goggins M. Pancreatic juice mutation concentrations can help predict the grade of dysplasia in patients undergoing pancreatic surveillance.Clin Cancer Res. 2018; 24: 2963-2974Crossref PubMed Scopus (47) Google Scholar as well as in pancreatic cyst fluid23Wu J. Matthaei H. Maitra A. Dal Molin M. Wood L.D. Eshleman J.R. Goggins M. Canto M.I. Schulick R.D. Edil B.H. Wolfgang C.L. Klein A.P. Diaz Jr., L.A. Allen P.J. Schmidt C.M. Kinzler K.W. Papadopoulos N. Hruban R.H. Vogelstein B. Recurrent GNAS mutations define an unexpected pathway for pancreatic cyst development.Sci Transl Med. 2011; 3: 92ra66Crossref PubMed Scopus (616) Google Scholar and pancreatic juice from patients with IPMN,25Kanda M. Knight S. Topazian M. Syngal S. Farrell J.J. Lee J. Kamel I. Lennon A.M. Borges M. Young A. Fujiwara S. Seike J. Eshleman J. Hruban R.H. Canto M.I. Goggins M. Mutant GNAS detected in duodenal collections of secretin-stimulated pancreatic juice indicates the presence or emergence of pancreatic cysts.Gut. 2013; 62: 1024-1033Crossref PubMed Scopus (133) Google Scholar raising the possibility that the detection of circulating GNAS mutations could have a role in the early detection of pancreatic cancers that arise from IPMNs. Methods used to detect ctDNA mutations rely on PCR and sequencing; these methods generate rare sequencing errors such as those related to nucleotide incorporation by DNA polymerases that need to be accounted for in the detection of low-abundance mutations. Another source of error with ctDNA detection methods can arise from ex vivo chemical modification of DNA such as by cytosine deamination to uracil.26Chen G. Mosier S. Gocke C.D. Lin M.T. Eshleman J.R. Cytosine deamination is a major cause of baseline noise in next-generation sequencing.Mol Diagn Ther. 2014; 18: 587-593Crossref PubMed Scopus (87) Google Scholar,27Lin M.T. Mosier S.L. Thiess M. Beierl K.F. Debeljak M. Tseng L.H. Chen G. Yegnasubramanian S. Ho H. Cope L. Wheelan S.J. Gocke C.D. Eshleman J.R. Clinical validation of KRAS, BRAF, and EGFR mutation detection using next-generation sequencing.Am J Clin Pathol. 2014; 141: 856-866Crossref PubMed Scopus (119) Google Scholar Enzymatic pretreatment of DNA has been used to try to limit the errors generated by cytosine deamination and other chemical modifications of DNA in formalin-fixed tissues and in forensic samples,28Ambers A. Turnbough M. Benjamin R. King J. Budowle B. Assessment of the role of DNA repair in damaged forensic samples.Int J Legal Med. 2014; 128: 913-921Crossref PubMed Scopus (19) Google Scholar but it has not been evaluated extensively for its potential to reduce errors when testing plasma DNA for mutations. In this study, enzymatic pretreatment of plasma DNA followed by digital NGS was used to detect hotspot mutations in KRAS and GNAS in patients with pancreatic cancer, including patients with an IPMN-associated pancreatic cancer, and in healthy and disease controls. Plasma samples were obtained either from patients who were enrolled in the Cancer of the Pancreas Screening studies or from patients undergoing a pancreatic resection at the Johns Hopkins Hospital.2Canto M.I. Almario J.A. Schulick R.D. Yeo C.J. Klein A. Blackford A. Shin E.J. Sanyal A. Yenokyan G. Lennon A.M. Kamel I.R. Fishman E.K. Wolfgang C. Weiss M. Hruban R.H. Goggins M. Risk of neoplastic progression in individuals at high risk for pancreatic cancer undergoing long-term surveillance.Gastroenterology. 2018; 155: 740-751.e2Abstract Full Text Full Text PDF PubMed Scopus (198) Google Scholar,29Canto M.I. Hruban R.H. Fishman E.K. Kamel I.R. Schulick R. Zhang Z. Topazian M. Takahashi N. Fletcher J. Petersen G. Klein A.P. Axilbund J. Griffin C. Syngal S. Saltzman J.R. Mortele K.J. Lee J. Tamm E. Vikram R. Bhosale P. Margolis D. Farrell J. Goggins M. Frequent detection of pancreatic lesions in asymptomatic high-risk individuals.Gastroenterology. 2012; 142 (quiz e14–e15): 796-804Abstract Full Text Full Text PDF PubMed Scopus (462) Google Scholar Patients included those with pancreatic ductal adenocarcinoma (n = 67), including 10 cases with a co-occurring IPMN lesion, 9 of which were thought to be arising pathologically from the IPMN; and controls (n = 73), including healthy laboratory employees (n = 19) and disease controls (n = 54). The disease controls included patients evaluated by endoscopic ultrasound or endoscopic retrograde cholangiopancreatography for nonpancreatic indications such as abdominal pain or benign biliary disease and found to have normal pancreata (n = 21), patients with known or suspected IPMN (n = 13), and/or familial/inherited susceptibility to pancreatic cancer (n = 9), patients with serous cystadenomas (n = 8), patients with acute/chronic pancreatitis (n = 6), and pancreatic neuroendocrine tumors (n = 2). Disease controls had a similar sex profile and age range to patients with pancreatic cancer, although the average age of the disease controls was somewhat younger (means ± SD/range, 57.9 ± 13.2/27 to 84 years versus 67.2 ± 10.5/44 to 89 years; P < 0.0001, unpaired t-test). Overall survival from pancreatic cancer was determined from the date of surgery (blood was drawn just before the surgery). Further description of the patient population is provided in Table 1 and Supplemental Tables S1 and S2.Table 1Characteristics of Cases Included in This StudyCharacteristicTotal (n = 140)Controls (n = 73)PDAC (n = 67)Sex, n (%) Male58 (41.4)24 (32.9)34 (50.7) Female63 (45.0)30 (41.1)33 (49.3) Unknown19 (13.6)19 (26.0)0 (0.0)Age, years, mean (range)∗Age and race information were not collected from healthy controls.63 (27–89)58 (27–84)67 (44–89)Race/ethnicity, n (%)∗Age and race information were not collected from healthy controls. Caucasian100 (71.4)42 (57.5)58 (86.6) African American12 (8.6)7 (9.6)5 (7.5) Hispanic/Latino2 (1.4)1 (1.4)1 (1.5) Asian2 (1.4)2 (2.7)0 (0.0) Other/unknown24 (17.1)21 (28.8)3 (4.5)AJCC stage, n (%) IA5NA5 (7.5) IB9NA9 (13.4) IIA1NA1 (1.5) IIB22NA22 (32.8) III11NA11 (16.4) IV19NA19 (28.4)Neoadjuvant chemotherapy, n (%)34NA34 (50.7)The controls include both healthy (n = 19) and disease controls (n = 54).AJCC, American Joint Committee on Cancer; IPMN, intraductal papillary mucinous neoplasm; NA, not applicable; PDAC, pancreatic ductal adenocarcinoma.∗ Age and race information were not collected from healthy controls. Open table in a new tab The controls include both healthy (n = 19) and disease controls (n = 54). AJCC, American Joint Committee on Cancer; IPMN, intraductal papillary mucinous neoplasm; NA, not applicable; PDAC, pancreatic ductal adenocarcinoma. All peripheral blood samples were collected in 10-mL EDTA vacutainers (BD Biosciences, San Jose, CA) and processed within 2 hours of collection. Plasma tubes were first spun at 1200 × g for 10 minutes. The plasma layer then was transferred to a new collection tube and spun at 1500 × g for 5 minutes. Plasma was aliquoted and stored at −80°C until DNA extractions were performed. Genomic DNA was extracted from approximately 3 mL plasma using the QIAamp Circulating Nucleic Acid Kit (Qiagen, Hilden, Germany), according to the manufacturer's instructions. When available, DNA extracted from archived pancreatic cancer tissues was analyzed to compare mutations found in the plasma with those in the cancer. To isolate tumor tissue, formalin-fixed, paraffin-embedded tumor tissues were cut onto membrane slides for laser capture microdissection to enrich for tumor cellularity, as previously described.30Hata T. Suenaga M. Marchionni L. Macgregor-Das A. Yu J. Shindo K. Tamura K. Hruban R.H. Goggins M. Genome-wide somatic copy number alterations and mutations in high-grade pancreatic intraepithelial neoplasia.Am J Pathol. 2018; 188: 1723-1733Abstract Full Text Full Text PDF PubMed Scopus (20) Google Scholar Genomic DNA was processed using the DNeasy Blood and Tissue Kit (Qiagen). Extracted DNA was quantified using the Quantifiler Human DNA Quantification Kit (Applied Biosystems, Foster City, CA). The Johns Hopkins Institutional Review Board approved all elements of this study, and written informed consent was obtained from all patients. The Ion Torrent PGM (Thermo Fisher Scientific, Waltham, MA) platform was used to to perform targeted NGS for KRAS (codons 12 and 13) and GNAS (codon 201). PreCR Repair Mix (New England Biolabs, MA) was used to treat DNA before library preparation as recommended by the manufacturer's protocol to limit false-positive detection of mutations arising from chemically modified DNA. PCR was performed using the Platinum SuperFi DNA Polymerase after either the Ion Torrent Ion Amplicon Fusion library preparation method or the AmpliSeq library preparation protocol (all from Thermo Fisher Scientific). These two methods were directly compared using 6 reference samples (four disease control and two pancreatic ductal adenocarcinoma patient plasma DNA samples). No significant differences in the detection of mutations between the two methods was observed (Supplemental Table S3). The uracil N-glycosylase enzyme (Thermo Fisher Scientific), which removes uracil residues and results in abasic polynucleotides that are degraded with heat, was added to Platinum SuperFi Taq PCR buffers to limit the carryover of deaminated templates during PCR (uracil N-glycosylase inhibited AmpliSeq PCR and therefore was not used). For digital NGS, multiple (384 of each plasma DNA sample) aliquots were made (100 pg/well) and sequenced separately. If the DNA yield isolated from plasma was not sufficient for 384 NGS reactions (approximately 40 ng total), then fewer aliquots were sequenced (a minimum of 96 NGS reactions were undertaken per sample). After library preparation, the libraries were cleaned using the Select-a-size DNA clean and concentrator kit (Zymo Research, Irvine, CA), quantified, and loaded into the OneTouch2 (Thermo Fisher Scientific) for emulsion PCR. After enrichment, samples were sequenced on the Ion Torrent PGM using 314v2 chips per the manufacturer's protocol. Analysis after sequencing included alignment to the human genome assembly 19 human reference genome and variant calling, and was performed using NextGENe software version 2.4 (SoftGenetics, State College, PA). All mutations were verified visually using the Integrative Genomics Viewer version 2.3 (Broad Institute, Cambridge, MA). A digital NGS score was determined for each hotspot mutation for each sample, with a score of 1 for each independent NGS aliquot with a mutation (up to 384). Background errors tend to arise in proportion to the number of DNA molecules sampled and vary by nucleotide, with more positives at sites of cytosine deamination.31Suenaga M. Sadakari Y. Almario J.A. Borges M. Lennon A.M. Shin E.J. Canto M.I. Goggins M. Using an endoscopic distal cap to collect pancreatic fluid from the ampulla (with video).Gastrointest Endosc. 2017; 86: 1152-1156.e2Abstract Full Text Full Text PDF PubMed Scopus (10) Google Scholar The limit of detection for KRAS and GNAS hotspot mutations for the digital NGS assay was calculated by performing eight replicate digital NGS analyses of a wild-type DNA sample (approximately 5000 genome equivalents) on the AmpliSeq platform; the limit of detection (means plus 2 SDs) for the major hotspot mutations ranged from a digital NGS score of 0 for KRAS G12R and G12A to a score of 2 of 384 for G12D (approximately 1/5000 genome equivalents), 3 of 384 for G12V and G12C, and for GNAS 201C and 5 for GNAS 201H. A diagnostic cut-off value for each KRAS and GNAS hotspot mutation was determined by the means plus 2 SDs of the mean score in the disease control group. Based on this criterion, a positive ctDNA score for KRAS G12D was 5 positive wells per 384 tested; for KRAS G12V it was 3 positive wells per 384 tested; for G12R it was 2 positive wells per 384 tested; and for GNAS 201C and 201H it was 10 positive wells per 384 tested. A few samples were classified as borderline or indeterminate because the mutation score for the sample was just below the cut-off value set for a 384 digital NGS assay, but fewer NGS reactions were performed (range, 96 to 288). For example, a positive digital NGS ctDNA score for KRAS G12D was 5 or more aliquots with mutations detected per 384 tested; an indeterminate result would be 4 positive wells of 288 analyzed. Preoperative cancer antigen 19-9 (CA19-9) levels were obtained from the patient's medical record when available (n = 12), or were measured in duplicate serum from blood samples obtained before pancreatic resection (n = 33) using the CA19-9 enzyme-linked immunosorbent assay kit (DRG International, Springfield Township, NJ). A cut-off value of 37 U/mL was considered increased. Categoric variables were summarized as frequencies (%) and compared with the χ2 test or the Fisher exact test as appropriate. Sensitivity and specificity were calculated using 2 × 2 contingency tables. Statistical analysis was performed using GraphPad Prism software version 7 (GraphPad Software, San Diego, CA) and JMP 23 software version 14 (SAS Institute, Cary, NC). P < 0.05 was considered statistically significant. Reference DNA samples developed by the National Institute for Standards and Technology (NIST) that contained KRAS (p.G12D) and GNAS (p.R201C) mutations spiked into wild-type genomic DNA at a variant allele frequency of either 2% or 0.5% were used to assess the reproducibility of this digital NGS assay.32He H.J. Stein E.V. Konigshofer Y. Forbes T. Tomson F.L. Garlick R. Yamada E. Godfrey T. Abe T. Tamura T. Borges M. Goggins M. Elmore S. Gulley M.L. Larson J.L. Ringel L. Haynes B.C. Karlovich C. Williams P.M. Garnett A. Ståhlberg A. Filges S. Sorbara S. Young M.R. Srivastava S. Cole K.D. Multi-laboratory assessment of a new reference material for quality assurance of cell-free tumor DNA measurements.J Mol Diagn. 2019; 21: 658-676Abstract Full Text Full Text PDF PubMed Scopus (10) Google Scholar For each sample, 1536 aliquots across six independent NGS experiments were separately sequenced. KRAS and GNAS mutations were detected at the expected variant concentrations (means ± SD, 2.0% ± 0.3% and 2.08% ± 0.43%, respectively, for the 2% variant allele frequency NIST sample and 0.60% ± 0.24% and 0.54% ± 0.17%, respectively, for the 0.5% variant allele frequency NIST sample). To address whether DNA pretreatment could interfere with the detection of real mutations, NIST reference DNA samples were pretreated with the DNA repair cocktail (PreCR Repair Mix; New England Biolabs) in three independent experiments, and the variant concentrations for KRAS (p.G12D) and GNAS (p.R201C) with and without pretreatment were compared. A slight, but not significant, difference in variant allele frequencies for either KRAS (means ± SD, 2.12% ± 0.37% versus 1.84% ± 0.53%, untreated versus treated, P = 0.40, paired t-test) or GNAS (means ± SD, 1.91% ± 0.55% versus 1.71% ± 0.28%, untreated versus treated, P = 0.60, paired t-test) was observed using the 2% mutant DNA reference sample from NIST or the 0.5% sample (KRAS: means ± SD, 0.67% ± 0.36% versus 0.55% ± 0.12%, untreated versus treated, P = 0.50, paired t-test; GNAS: means ± SD, 0.55% ± 0.06% versus 0.50% ± 0.15, untreated versus treated, P = 0.74, paired t-test). The current study also examined whether pretreatment with the PreCR repair cocktail could induce mutations. A control patient DNA sample was pretreated with the DNA repair cocktail in three independent experiments and observed no induction of either KRAS or GNAS mutations when compared with the untreated DNA. Plasma DNA from 67 patients with pancreatic cancer and from 73 controls was analyzed using digital NGS for KRAS and GNAS mutations. The cancer cases included 10 with pancreatic cancer associated with an IPMN. Positive digital NGS scores indicating the presence of circulating KRAS mutations were detected in the plasma of 23 of 63 (36.5%) patients with pancreatic cancer (Table 2 and Supplemental Table S2). There was no significant difference in the likelihood of having detectable mutant KRAS ctDNA among patients who had (14 of 32) versus did not have (9 of 31) neoadjuvant therapy (P = 0.3). In addition, four patients with pancreatic cancer had mutation scores that were borderline and therefore classified as indeterminate. Digital NGS scores ranged from 3 to 94, corresponding to a 0.03% to 0.75% mutant concentration range in patients with a positive KRAS mutation. Three patients with pancreatic ductal adenocarcinoma had multiple KRAS mutations detected in their plasma; all three