Portal de Periódicos da CAPES

Genomic Analysis of Circulating Tumor DNA Using a Melanoma-Specific UltraSEEK Oncogene Panel

2019; Elsevier BV; Volume: 21; Issue: 3 Linguagem: Inglês

10.1016/j.jmoldx.2018.12.001

1943-7811

Elin S. Gray, Tom Witkowski, Michelle R. Pereira, Leslie Calapre, Karl Herron, Darryl Irwin, Brett Chapman, Muhammad A. Khattak, Jeanette Raleigh, Athena Hatzimihalis, Jonathan Cebon, Shahneen Sandhu, Grant A. McArthur, Michael Millward, Mel Ziman, Alexander Dobrovic, Stephen Q. Wong,

Cancer Cells and Metastasis

The analysis of circulating tumor DNA provides a minimally invasive molecular interrogation that has the potential to guide treatment selection and disease monitoring. Here, the authors evaluated a custom UltraSEEK melanoma panel for the MassARRAY system, probing for 61 mutations over 13 genes. The analytical sensitivity and clinical accuracy of the UltraSEEK melanoma panel was compared with droplet digital PCR. The blinded analysis of 68 mutations detected in 48 plasma samples from stage IV melanoma patients revealed a concordance of 88% between the two platforms. Further comparison of both methods for the detection of BRAF V600E mutations in 77 plasma samples demonstrated a Cohen's κ of 0.826 (bias-corrected and accelerated 95% CI, 0.669–0.946). These results indicate that the UltraSEEK melanoma panel is as sensitive as droplet digital PCR for the detection of circulating tumor DNA in this cohort of patients but highlight the need for detected variants to be confirmed orthogonally to mitigate any false-positive results. The MassARRAY system enables rapid and sensitive genotyping for the detection of multiple melanoma–associated mutations in plasma. The analysis of circulating tumor DNA provides a minimally invasive molecular interrogation that has the potential to guide treatment selection and disease monitoring. Here, the authors evaluated a custom UltraSEEK melanoma panel for the MassARRAY system, probing for 61 mutations over 13 genes. The analytical sensitivity and clinical accuracy of the UltraSEEK melanoma panel was compared with droplet digital PCR. The blinded analysis of 68 mutations detected in 48 plasma samples from stage IV melanoma patients revealed a concordance of 88% between the two platforms. Further comparison of both methods for the detection of BRAF V600E mutations in 77 plasma samples demonstrated a Cohen's κ of 0.826 (bias-corrected and accelerated 95% CI, 0.669–0.946). These results indicate that the UltraSEEK melanoma panel is as sensitive as droplet digital PCR for the detection of circulating tumor DNA in this cohort of patients but highlight the need for detected variants to be confirmed orthogonally to mitigate any false-positive results. The MassARRAY system enables rapid and sensitive genotyping for the detection of multiple melanoma–associated mutations in plasma. Though comprising <2% of skin cancers, melanoma is responsible for the largest number of skin cancer–related deaths. Advances in targeted therapies and immune checkpoint inhibitors have revolutionized treatment in the metastatic setting.1Hamid O. Robert C. Daud A. Hodi F.S. Hwu W.J. Kefford R. Wolchok J.D. Hersey P. Joseph R.W. Weber J.S. Dronca R. Gangadhar T.C. Patnaik A. Zarour H. Joshua A.M. Gergich K. Elassaiss-Schaap J. Algazi A. Mateus C. Boasberg P. Tumeh P.C. Chmielowski B. Ebbinghaus S.W. Li X.N. Kang S.P. Ribas A. Safety and tumor responses with lambrolizumab (anti-PD-1) in melanoma.N Engl J Med. 2013; 369: 134-144Crossref PubMed Scopus (2715) Google Scholar, 2Hodi F.S. O'Day S.J. McDermott D.F. Weber R.W. Sosman J.A. Haanen J.B. Gonzalez R. Robert C. Schadendorf D. Hassel J.C. Akerley W. van den Eertwegh A.J. Lutzky J. Lorigan P. Vaubel J.M. Linette G.P. Hogg D. Ottensmeier C.H. Lebbe C. Peschel C. Quirt I. Clark J.I. Wolchok J.D. Weber J.S. Tian J. Yellin M.J. Nichol G.M. Hoos A. Urba W.J. Improved survival with ipilimumab in patients with metastatic melanoma.N Engl J Med. 2010; 363: 711-723Crossref PubMed Scopus (11189) Google Scholar, 3Homet B. Ribas A. New drug targets in metastatic melanoma.J Pathol. 2014; 232: 134-141Crossref PubMed Scopus (47) Google Scholar Despite significant improvement in overall survival, most patients on targeted therapies develop drug resistance within 12 months and immunotherapies are only effective in some patients.3Homet B. Ribas A. New drug targets in metastatic melanoma.J Pathol. 2014; 232: 134-141Crossref PubMed Scopus (47) Google Scholar Currently, radiological analysis and existing disease-monitoring biomarkers (mainly lactate dehydrogenase levels) are inadequate for guiding treatment selection, tracking response kinetics, and the detection of emerging treatment resistance. Melanoma patients would benefit from a sensitive personalized test to monitor disease that can complement current therapies to melanoma. Cell-free DNA (cfDNA) are fragments of DNA shed into the bloodstream during cellular turnover, and in the case of tumor cells, the released DNA is referred to as circulating tumor DNA (ctDNA), which can be distinguished from normal cfDNA by the detection of tumor-associated somatic mutations. Analysis of ctDNA offers the potential of a noninvasive method for identification of melanoma patients for molecularly based targeted therapies.4Santiago-Walker A. Gagnon R. Mazumdar J. Casey M. Long G.V. Schadendorf D. Flaherty K. Kefford R. Hauschild A. Hwu P. Haney P. O'Hagan A. Carver J. Goodman V. Legos J. Martin A.M. Correlation of BRAF mutation status in circulating-free DNA and tumor and association with clinical outcome across four BRAFi and MEKi clinical trials.Clin Cancer Res. 2016; 22: 567-574Crossref PubMed Scopus (149) Google Scholar, 5Ascierto P.A. Minor D. Ribas A. Lebbe C. O'Hagan A. Arya N. Guckert M. Schadendorf D. Kefford R.F. Grob J.J. Hamid O. Amaravadi R. Simeone E. Wilhelm T. Kim K.B. Long G.V. Martin A.M. Mazumdar J. Goodman V.L. Trefzer U. Phase II trial (BREAK-2) of the BRAF inhibitor dabrafenib (GSK2118436) in patients with metastatic melanoma.J Clin Oncol. 2013; 31: 3205-3211Crossref PubMed Scopus (343) Google Scholar In addition, ctDNA is emerging as a promising biomarker for early detection of disease status, particularly at times of treatment response or tumor regrowth.6Lee J.H. Long G.V. Boyd S. Lo S. Menzies A.M. Tembe V. Guminski A. Jakrot V. Scolyer R.A. Mann G.J. Kefford R.F. Carlino M.S. Rizos H. Circulating tumour DNA predicts response to anti-PD1 antibodies in metastatic melanoma.Ann Oncol. 2017; 28: 1130-1136Abstract Full Text Full Text PDF PubMed Scopus (198) Google Scholar, 7Gray E.S. Rizos H. Reid A.L. Boyd S.C. Pereira M.R. Lo J. Tembe V. Freeman J. Lee J.H. Scolyer R.A. Siew K. Lomma C. Cooper A. Khattak M.A. Meniawy T.M. Long G.V. Carlino M.S. Millward M. Ziman M. Circulating tumor DNA to monitor treatment response and detect acquired resistance in patients with metastatic melanoma.Oncotarget. 2015; 6: 42008-42018Crossref PubMed Scopus (232) Google Scholar, 8Wong S.Q. Raleigh J.M. Callahan J. Vergara I.A. Ftouni S. Hatzimihalis A. Colebatch A.J. Li J. Semple T. Doig K. Mintoff C. Sinha D. Yeh P. Silva M.J. Alsop K. Thorne H. Bowtell D.D. Gyorki D.E. Arnau G.M. Cullinane C. Kee D. Brady B. Kelleher F. Dawson M.A. Papenfuss A.T. Shackleton M. Hicks R.J. McArthur G.A. Sandhu S. Dawson S.-J. Circulating tumor DNA analysis and functional imaging provide complementary approaches for comprehensive disease monitoring in metastatic melanoma.JCO Precis Oncol. 2017; 2017: 1-14Google Scholar, 9Tsao S.C. Weiss J. Hudson C. Christophi C. Cebon J. Behren A. Dobrovic A. Monitoring response to therapy in melanoma by quantifying circulating tumour DNA with droplet digital PCR for BRAF and NRAS mutations.Sci Rep. 2015; 5: 11198Crossref PubMed Scopus (140) Google Scholar ctDNA levels in plasma have been found to be strongly associated with tumor burden8Wong S.Q. Raleigh J.M. Callahan J. Vergara I.A. Ftouni S. Hatzimihalis A. Colebatch A.J. Li J. Semple T. Doig K. Mintoff C. Sinha D. Yeh P. Silva M.J. Alsop K. Thorne H. Bowtell D.D. Gyorki D.E. Arnau G.M. Cullinane C. Kee D. Brady B. Kelleher F. Dawson M.A. Papenfuss A.T. Shackleton M. Hicks R.J. McArthur G.A. Sandhu S. Dawson S.-J. Circulating tumor DNA analysis and functional imaging provide complementary approaches for comprehensive disease monitoring in metastatic melanoma.JCO Precis Oncol. 2017; 2017: 1-14Google Scholar and in particular with metabolic tumor burden,10McEvoy A.C. Warburton L. Al-Ogaili Z. Celliers L. Calapre L. Pereira M.R. Khattak M.A. Meniawy T.M. Millward M. Ziman M. Gray E.S. Correlation between circulating tumour DNA and metabolic tumour burden in metastatic melanoma patients.BMC Cancer. 2018; 18: 726Crossref PubMed Scopus (59) Google Scholar and low pretreatment ctDNA levels are associated with better overall response rates and longer progression-free survival.4Santiago-Walker A. Gagnon R. Mazumdar J. Casey M. Long G.V. Schadendorf D. Flaherty K. Kefford R. Hauschild A. Hwu P. Haney P. O'Hagan A. Carver J. Goodman V. Legos J. Martin A.M. Correlation of BRAF mutation status in circulating-free DNA and tumor and association with clinical outcome across four BRAFi and MEKi clinical trials.Clin Cancer Res. 2016; 22: 567-574Crossref PubMed Scopus (149) Google Scholar, 7Gray E.S. Rizos H. Reid A.L. Boyd S.C. Pereira M.R. Lo J. Tembe V. Freeman J. Lee J.H. Scolyer R.A. Siew K. Lomma C. Cooper A. Khattak M.A. Meniawy T.M. Long G.V. Carlino M.S. Millward M. Ziman M. Circulating tumor DNA to monitor treatment response and detect acquired resistance in patients with metastatic melanoma.Oncotarget. 2015; 6: 42008-42018Crossref PubMed Scopus (232) Google Scholar, 11Sanmamed M.F. Fernandez-Landazuri S. Rodriguez C. Zarate R. Lozano M.D. Zubiri L. Gracia J.L. Martin-Algarra S. Gonzalez A. Quantitative cell-free circulating BRAFV600E mutation analysis by use of droplet digital PCR in the follow-up of patients with melanoma being treated with BRAF inhibitors.Clin Chem. 2015; 61: 297-304Crossref PubMed Scopus (193) Google Scholar Droplet digital PCR (ddPCR) has emerged as one of the most cost-effective and sensitive methods for the analysis of rare copies of mutant ctDNA. However, ddPCR at present only allows for the detection of one mutation at a time with some multiplex assays available targeting a limited number of specific hotspot mutations. Although next-generation sequencing can detect mutations from a broad range of genes from plasma DNA, it is relatively costly, has a slow turnaround time, and requires high input material and complex bioinformatics platform analysis. Therefore, there is a requirement for a rapid, sensitive, and cost-effective assay that comprehensively screens for multiple commonly occurring mutations in melanoma. Comprehensive genetic studies of melanomas have provided insights into the mutational landscape of melanoma, providing potentially important implications for prognosis and therapy.12Hodis E. Watson I.R. Kryukov G.V. Arold S.T. Imielinski M. Theurillat J.P. et al.A landscape of driver mutations in melanoma.Cell. 2012; 150: 251-263Abstract Full Text Full Text PDF PubMed Scopus (1843) Google Scholar, 13Krauthammer M. Kong Y. Ha B.H. Evans P. Bacchiocchi A. McCusker J.P. Cheng E. Davis M.J. Goh G. Choi M. Ariyan S. Narayan D. Dutton-Regester K. Capatana A. Holman E.C. Bosenberg M. Sznol M. Kluger H.M. Brash D.E. Stern D.F. Materin M.A. Lo R.S. Mane S. Ma S. Kidd K.K. Hayward N.K. Lifton R.P. Schlessinger J. Boggon T.J. Halaban R. Exome sequencing identifies recurrent somatic RAC1 mutations in melanoma.Nat Genet. 2012; 44: 1006-1014Crossref PubMed Scopus (859) Google Scholar, 14Cancer Genome Atlas NetworkGenomic classification of cutaneous melanoma.Cell. 2015; 161: 1681-1696Abstract Full Text Full Text PDF PubMed Scopus (1914) Google Scholar In particular, the Cancer Genome Atlas annotation of melanomas define four genetic subclasses: BRAF mutant, NRAS mutant, NF1 mutant, and triple wild-type.14Cancer Genome Atlas NetworkGenomic classification of cutaneous melanoma.Cell. 2015; 161: 1681-1696Abstract Full Text Full Text PDF PubMed Scopus (1914) Google Scholar Although most melanomas carry a mutation in BRAF codon V600 (approximately 50%) or NRAS codons Q61 or G12/13 (approximately 20%), a number of variants need to be assessed for each position. NF1 mutations are distributed across the whole gene with no defined hotspot mutations, making it difficult for targeted screening for somatic mutations.13Krauthammer M. Kong Y. Ha B.H. Evans P. Bacchiocchi A. McCusker J.P. Cheng E. Davis M.J. Goh G. Choi M. Ariyan S. Narayan D. Dutton-Regester K. Capatana A. Holman E.C. Bosenberg M. Sznol M. Kluger H.M. Brash D.E. Stern D.F. Materin M.A. Lo R.S. Mane S. Ma S. Kidd K.K. Hayward N.K. Lifton R.P. Schlessinger J. Boggon T.J. Halaban R. Exome sequencing identifies recurrent somatic RAC1 mutations in melanoma.Nat Genet. 2012; 44: 1006-1014Crossref PubMed Scopus (859) Google Scholar Thus, other commonly mutated sites need to be targeted for ctDNA monitoring of NF1 and triple–wild-type melanomas. For example, other melanoma-associated mutations such as those in the DPH3 promoter,15Denisova E. Heidenreich B. Nagore E. Rachakonda P.S. Hosen I. Akrap I. Traves V. Garcia-Casado Z. Lopez-Guerrero J.A. Requena C. Sanmartin O. Serra-Guillen C. Llombart B. Guillen C. Ferrando J. Gimeno E. Nordheim A. Hemminki K. Kumar R. Frequent DPH3 promoter mutations in skin cancers.Oncotarget. 2015; 6: 35922-35930Crossref PubMed Scopus (41) Google Scholar TERT promoter,16McEvoy A.C. Calapre L. Pereira M.R. Giardina T. Robinson C. Khattak M.A. Meniawy T.M. Pritchard A.L. Hayward N.K. Amanuel B. Millward M. Ziman M. Gray E.S. Sensitive droplet digital PCR method for detection of TERT promoter mutations in cell free DNA from patients with metastatic melanoma.Oncotarget. 2017; 8: 78890-78900Crossref PubMed Scopus (34) Google Scholar RPS27 UTR,17Dutton-Regester K. Gartner J.J. Emmanuel R. Qutob N. Davies M.A. Gershenwald J.E. Robinson W. Robinson S. Rosenberg S.A. Scolyer R.A. Mann G.J. Thompson J.F. Hayward N.K. Samuels Y. A highly recurrent RPS27 5′UTR mutation in melanoma.Oncotarget. 2014; 5: 2912-2917Crossref PubMed Scopus (40) Google Scholar and RAC1,18Watson I.R. Li L. Cabeceiras P.K. Mahdavi M. Gutschner T. Genovese G. Wang G. Fang Z. Tepper J.M. Stemke-Hale K. Tsai K.Y. Davies M.A. Mills G.B. Chin L. The RAC1 P29S hotspot mutation in melanoma confers resistance to pharmacological inhibition of RAF.Cancer Res. 2014; 74: 4845-4852Crossref PubMed Scopus (111) Google Scholar, 19Mar V.J. Wong S.Q. Logan A. Nguyen T. Cebon J. Kelly J.W. Wolfe R. Dobrovic A. McLean C. McArthur G.A. Clinical and pathological associations of the activating RAC1 P29S mutation in primary cutaneous melanoma.Pigment Cell Melanoma Res. 2014; 27: 1117-1125Crossref PubMed Scopus (39) Google Scholar among others, provide alternatives for ctDNA monitoring in BRAF/NRAS wild-type melanomas. Here, 48 plasma samples from metastatic melanoma patients were evaluated for mutations using a custom UltraSEEK melanoma panel on the MassARRAY system (Agena Bioscience, San Diego, CA). This test allows analysis of 61 mutations over 13 genes within a single reaction. To determine the accuracy of the assay, results were compared with mutations identified in the same plasma samples by ddPCR. Blood samples were collected from stage IV melanoma patients enrolled at the Sir Charles Gairdner Hospital and Fiona Stanley Hospital in Perth, WA, the Olivia Newton-John Cancer Wellness & Research Centre, and the Peter MacCallum Cancer Centre in Melbourne, Victoria. Written informed consent was obtained from all patients under approved Human Research Ethics Committee protocols from Edith Cowan University (No. 11543), Sir Charles Gairdner Hospital (No. 2007-123), Peter MacCallum Cancer Centre (PMCC: 11/105), and Austin Hospital (HREC/14/Austin/425), with all methods performed in accordance with the relevant ethical guidelines and regulations of the Australian National Health and Medical Research Council. All tumor and plasma samples from the Olivia Newton-John Cancer Wellness & Research Centre were collected from patients as part of the Melbourne Melanoma Project. All tumor and plasma samples from the Peter MacCallum Cancer Centre were collected from patients as part of the Melanoma Biomarkers Study.8Wong S.Q. Raleigh J.M. Callahan J. Vergara I.A. Ftouni S. Hatzimihalis A. Colebatch A.J. Li J. Semple T. Doig K. Mintoff C. Sinha D. Yeh P. Silva M.J. Alsop K. Thorne H. Bowtell D.D. Gyorki D.E. Arnau G.M. Cullinane C. Kee D. Brady B. Kelleher F. Dawson M.A. Papenfuss A.T. Shackleton M. Hicks R.J. McArthur G.A. Sandhu S. Dawson S.-J. Circulating tumor DNA analysis and functional imaging provide complementary approaches for comprehensive disease monitoring in metastatic melanoma.JCO Precis Oncol. 2017; 2017: 1-14Google Scholar Blood was collected into EDTA vacutainer tubes or BCT blood collection tubes (Streck, La Vista, NE) and stored at 4°C until processing. Plasma was separated within 24 hours by centrifugation at 1600 × g for 10 minutes, followed by a second centrifugation at 2000 × g for 10 minutes, and then stored at −80°C until extraction. cfDNA was isolated from between 2 and 5 mL of plasma using the QIAamp Circulating Nucleic Acid Kit (Qiagen, Hilden, Germany) as per the manufacturer's instructions and stored at −80°C until ctDNA quantification. PCR was performed using 10 ng of cfDNA in a single PCR reaction according to the manufacturer's instructions (Agena Bioscience). Reactions were incubated initially at 94°C for 2 minutes. Forty-five cycles of PCR were performed at 94°C for 30 seconds, 56°C for 30 seconds, and 72°C for 1 minute. The PCR was completed with a final incubation of 5 minutes at 72°C. Thermocycling and incubation were performed in a Veriti Thermal Cycler (Thermo Fisher Scientific, Waltham, MA). Amplified products (70 μL) were treated with shrimp alkaline phosphatase for 40 minutes at 37°C, followed by denaturation for 5 minutes at 85°C. Single-base extension with biotinylated chain terminator nucleotides specific to the mutant allele was performed at 94°C for 30 seconds, followed by 40 cycles at 94°C for 5 seconds with five nested cycles of 52°C for 5 seconds, then 80°C for 5 seconds and incubation at 72°C for 3 minutes. Streptavidin-coated magnetic beads were used to capture the single-base–extended oligonucleotides. Beads with captured products were pelleted using a magnet, suspended with 13 mL of biotin competition solution, and then incubated at 95°C for 5 minutes. Eluted products were conditioned with 2 μL (2 mg) of anion exchange resin slurry. Finally, the analyte was dispensed onto a SpectroCHIP Array solid support using a MassARRAY RS1000 Nano-dispenser. Data were acquired via matrix-assisted laser desorption/ionization time-of-flight mass spectrometry using the MassARRAY Analyzer. Data analysis was performed using Typer software version 4.0.26.74 (Agena Bioscience). Normalized intensity was calculated for the signal intensity of the mutant allele, which had been normalized against the capture control peaks found in the spectrum. A value of one means the peak intensity of the observed mutant allele is equal to the peak intensity of the average of the 5 capture control peaks found in the spectrum. The capture control peaks are biotin-labeled, nonreactive oligos, which are added to the extension reaction and used as an internal control for the streptavidin-bead capture and elution of the mutant extension product steps. The UltraSEEK Melanoma Panel assay validation used a model system developed to simulate samples harboring low-frequency somatic mutations. Wild-type DNA (Coriell Cell Repositories, Camden, NJ) was spiked with different amounts of characterized cell lines (Horizon Diagnostics, Cambridge, UK) harboring engineered mutations (Supplemental Table S1). The mixtures represented a 0.1%, 0.2%, 0.5%, and 1% mutant allele frequency, while keeping the total number of DNA molecules constant. Each dilution was analyzed in four replicates. Each cell line harboring a mutation for a specific assay was considered wild type for all other assays in the multiplex. PCR reactions were performed in a 20-μL reaction containing 1× droplet PCR supermix (no dUTP) (Bio-Rad Laboratories, Hercules, CA), and each probe at a concentration of 250 nmol/L, primers at a concentration of 900 nmol/L, and 8 μL of cfDNA. Commercially available and/or customized probes were used to analyze ctDNA. Previously described custom primer and probe sets were used for detection of mutation in BRAF7Gray E.S. Rizos H. Reid A.L. Boyd S.C. Pereira M.R. Lo J. Tembe V. Freeman J. Lee J.H. Scolyer R.A. Siew K. Lomma C. Cooper A. Khattak M.A. Meniawy T.M. Long G.V. Carlino M.S. Millward M. Ziman M. Circulating tumor DNA to monitor treatment response and detect acquired resistance in patients with metastatic melanoma.Oncotarget. 2015; 6: 42008-42018Crossref PubMed Scopus (232) Google Scholar, 20Reid A.L. Freeman J.B. Millward M. Ziman M. Gray E.S. Detection of BRAF-V600E and V600K in melanoma circulating tumour cells by droplet digital PCR.Clin Biochem. 2015; 48: 999-1002Crossref PubMed Scopus (87) Google Scholar and DPH3 promoter.21Calapre L. Giardina T. Robinson C. Reid A.L. Al-Ogaili Z. Pereira M.R. McEvoy A.C. Warburton L. Hayward N.K. Khattak M.A. Meniawy T.M. Millward M. Amanuel B. Ziman M. Gray ES: Locus-specific concordance of genomic alterations between tissue and plasma circulating tumor DNA in metastatic melanoma.Mol Oncol. 2019; 13: 171-184Crossref PubMed Scopus (32) Google Scholar Droplets were generated and analyzed using the QX200 system (Bio-Rad Laboratories). ddPCR absolute quantification of mutant alleles and wild-type alleles was estimated by modeling as a Poisson distribution using the QuantaSoft analysis software version 1.6.6 (Bio-Rad). Thresholds were defined based on the signal from empty droplets, wild-type DNA controls, and mutant positive controls, as described in the Droplet Digital Application Guide (Bio-Rad). The absolute number of mutant allele per milliliter of blood and mutant allele frequency were calculated from the QuantaSoft analysis software version 1.6.6 outputs as follows: Copies/mL plasma = (copies per mL of reaction as per QuantaSoft analysis software version 1.6.6) × (volume of ddPCR reaction) ×([volume eluted/volume of DNA used in reaction]/4 mL of plasma) and Mutant allele frequency = mutant copies/mL of plasma/(mutant copies/mL of plasma+wild−type copies/mL of plasma). Cohen's kappa (κ) coefficient was used to assess agreement between ddPCR and UltraSEEK regarding the identification of BRAF V600E status (detected versus undetected), using SPSS software version 24 (IBM SPSS Statistics, Armonk, NY). Bias-corrected and accelerated 95% CI for Cohen's κ coefficient were constructed by bootstrapping, using 1000 bootstrap replicates. According to Landis and Koch,22Landis J.R. Koch G.G. The measurement of observer agreement for categorical data.Biometrics. 1977; 33: 159-174Crossref PubMed Scopus (51055) Google Scholar the following κ values were used for interpretation: poor-to-fair (≤0.4), moderate (0.41 to 0.60), substantial (0.61 to 0.80), and almost perfect agreement (0.81 to 1.00). A custom UltraSEEK melanoma panel was devised containing 86 assays targeting 61 melanoma-associated mutations over 13 genes (Supplemental Table S1). The presence of somatic mutations was analyzed in 48 plasma samples from stage IV melanoma patients recruited into liquid biopsy studies at four different hospitals across Australia. The selected plasma samples were previously screened for mutations by ddPCR or targeted sequencing. Characteristics of tumor and plasma samples are described in Supplemental Table S2. UltraSEEK analysis identified 80 mutations in these samples. Of those, 68 mutations could be compared with ddPCR results revealing concordant results for 60 (88%) of the mutations (Table 1). There was a significant correlation between the ddPCR mutational frequency abundance and the UltraSEEK normalized intensity (Pearson's r = 0.7056, P < 0.0001) (Figure 1). Overall, UltraSEEK was able to detect mutations across a broad range of copies/mL of plasma (range, 1.4 to 212,160) and fractional abundances (range, 0.1 to 97.4) as defined by ddPCR analysis (Figure 2).Table 1Comparison of Mutations Identified by UltraSEEK and ddPCR AnalysisPatientMutationddPCRUltraSEEKCopies/mLFANormIntP1BRAF V600K∗Mutation identified in formalin-fixed, paraffin-embedded tumor tissue.6518.873.81.21P3BRAF V600K∗Mutation identified in formalin-fixed, paraffin-embedded tumor tissue.5024.045.91.09IDH1 R132CNTNT0.16P4BRAF V600E∗Mutation identified in formalin-fixed, paraffin-embedded tumor tissue.63.02.40.33P5BRAF V600E∗Mutation identified in formalin-fixed, paraffin-embedded tumor tissue.268.03.41.13P7BRAF V600R∗Mutation identified in formalin-fixed, paraffin-embedded tumor tissue.2506.075.00.26IDH1 R132CNTNT0.27YAE1D1 c39605969G>ANTNT1.28P9BRAF V600K∗Mutation identified in formalin-fixed, paraffin-embedded tumor tissue.95.51.6NDBRAF V600E0.00.02.59P10BRAF V600E∗Mutation identified in formalin-fixed, paraffin-embedded tumor tissue.1527.016.32.59CDKN2A R80X1020.018.82.48P11BRAF V600K∗Mutation identified in formalin-fixed, paraffin-embedded tumor tissue.206.011.50.39CDKN2A R80X176.014.32.29SDHD c111957523C>TNTNT1.21P12CTNNB1 S45P0.00.00.44P15NRAS Q61K∗Mutation identified in formalin-fixed, paraffin-embedded tumor tissue.99.013.01.23DPH3 c16306504C>T128.01.70.99IDH1 R132CNTNT0.22P16BRAF V600E∗Mutation identified in formalin-fixed, paraffin-embedded tumor tissue.4265.328.93.39NRAS Q61K57.41.58P33BRAF K601E∗Mutation identified in formalin-fixed, paraffin-embedded tumor tissue.224.00.50.40P35NRAS G13R∗Mutation identified in formalin-fixed, paraffin-embedded tumor tissue.63.00.50.24P37BRAF V600K∗Mutation identified in formalin-fixed, paraffin-embedded tumor tissue.67.00.80.20P38BRAF V600K∗Mutation identified in formalin-fixed, paraffin-embedded tumor tissue.1018.05.30.30P39BRAF V600E∗Mutation identified in formalin-fixed, paraffin-embedded tumor tissue.0.00.00.29RQCD1 P131LNTNT0.29P40NRAS Q61K∗Mutation identified in formalin-fixed, paraffin-embedded tumor tissue.127.03.10.69E1BRAF V600E∗Mutation identified in formalin-fixed, paraffin-embedded tumor tissue.5.40.50.10BRAF K601ENTNT0.47E2BRAF V600E∗Mutation identified in formalin-fixed, paraffin-embedded tumor tissue.126.03.10.83E3BRAF V600E∗Mutation identified in formalin-fixed, paraffin-embedded tumor tissue.100.02.10.62BRAF K601ENTNT0.51E4BRAF V600E∗Mutation identified in formalin-fixed, paraffin-embedded tumor tissue.4.80.30.14E5BRAF V600E∗Mutation identified in formalin-fixed, paraffin-embedded tumor tissue.24.00.40.26E6NRAS Q61K∗Mutation identified in formalin-fixed, paraffin-embedded tumor tissue.318.010.01.37DPH3 c16306504C>T26.01.71.51E7BRAF V600E∗Mutation identified in formalin-fixed, paraffin-embedded tumor tissue.11.00.40.29E8BRAF V600E∗Mutation identified in formalin-fixed, paraffin-embedded tumor tissue.2380.031.53.38DPH3 c16306504C>T260.018.12.81E9BRAF V600E∗Mutation identified in formalin-fixed, paraffin-embedded tumor tissue.202.03.11.00E10BRAF V600R∗Mutation identified in formalin-fixed, paraffin-embedded tumor tissue.212160.097.40.45CDKN2A R80XNTNT2.76DPH3 c16306504C>T1960.032.82.19MAP2K1 P124S∗Mutation identified in formalin-fixed, paraffin-embedded tumor tissue.31300.012.31.52RAC1 P29S∗Mutation identified in formalin-fixed, paraffin-embedded tumor tissue.17420.062.22.62E11BRAF V600E2∗Mutation identified in formalin-fixed, paraffin-embedded tumor tissue.20.00.80.11DPH3 c16306504C>T0.00.00.12RAC1 P29S6.80.20.16E12BRAF V600E∗Mutation identified in formalin-fixed, paraffin-embedded tumor tissue.6.80.60.10IDH1 R132H1.40.10.31E13BRAF V600E∗Mutation identified in formalin-fixed, paraffin-embedded tumor tissue.56.08.31.86E14BRAF V600E∗Mutation identified in formalin-fixed, paraffin-embedded tumor tissue.18.00.70.31BRAF V600E2 TG>AA0.00.01.46E15BRAF V600E∗Mutation identified in formalin-fixed, paraffin-embedded tumor tissue.16.00.50.06E16BRAF V600R∗Mutation identified in formalin-fixed, paraffin-embedded tumor tissue.602.05.71.03E17BRAF V600E∗Mutation identified in formalin-fixed, paraffin-embedded tumor tissue.971.123.12.95E18KIT L576P342.012.43.28E19KIT V559A∗Mutation identified in formalin-fixed, paraffin-embedded tumor tissue.13.01.20.30E21DPH3 c16306504C>T0.00.00.14E23NRAS Q61K∗Mutation identified in formalin-fixed, paraffin-embedded tumor tissue.3825.40.98DPH3 c16306504C>T36.02.01.11E24NRAS Q61K∗Mutation identified in formalin-fixed, paraffin-embedded tumor tissue.9900403.02CDKN2A R80X1396.014.03.82BRAF V600E0.00.00.22O2BRAF V600E∗Mutation identified in formalin-fixed, paraffin-embedded tumor tissue.26.01.3NDDPH3 c16306504C>TNTNT0.50O3CTNNB1 S45PNTNT0.18DPH3 c16306504C>T0.00.01.17RAC1 P29S1943.08.51.60O4CTNNB1 S45FNTNT0.54DPH3 c16306504C>T1.30.10.30RPS27 c238CtoT11.01.40.54O6NRAS G12D538.01.71.15O9BRAF V600E∗Mutation identified in formalin-fixed, paraffin-embedded tumor tissue.882.00.20DPH3 c16306504C>T251.30.20O10BRAF V600E∗Mutation identified in formalin-fixed, paraffin-embedded tumor tissue.381.80.20DPH3 c16306504C>T00.01.20O11BRAF V600E∗Mutation identified in formalin-fixed, paraffin-embedded tumor tissue.2768.40.90DPH3 c16306504C>T52.91.50O12BRAF V600E∗Mutation identified in formalin-fixed, paraffin-embedded tumor tissue.39115.22.80DPH3 c16306504C>T32.51.60O13BRAF V600E∗Mutation identified in formalin-fixed, paraffin-embedded tumor tissue.912.50.40DPH3 c16306504C>T20.70.40O14BRAF V600E∗Mutation identified in formalin-fixed, paraffin-embedded tumor tissue.331.3NDDPH3 c16306504C>T00.00.40O15BRAF V600E∗Mutation identified in formalin-fixed, paraffin-embedded tumor tissue.261.3NDDPH3 c16306504C>T00.00.30O16BRAF V600E∗Mutation identified in formalin-fixed, paraffin-embedded tumor tissue.2195.60.60DPH3 c16306504C>T81.21.00Bold text indicates discordant results.FA, frequency abundance of mutant copies relative to wild-type DNA; ND, not detected; NormInt, normalized intensity; NT, not tested.∗ Mutation identified in formalin-fixed, paraffin-embedded tumor tissue. Open table in a new tab Figure 2Range of copies/mL and frequency abundance by ddPCR of mutations in the plasmas analyzed using MassARRAY UltraSEEK. Asterisks indicate samples not detected by UltraSEEK; horizontal dashed lines, 1% and 0.5% fractional abundance.View Large Image Figure ViewerDownload Hi-res image Download (PPT) Bold text indicates discordant results. FA, frequency abundance of mutant copies relative to wild-type DNA; ND, not detected; NormInt, normalized intensity; NT, not tested. Three BRAF mutations previously identified by ddPCR in these samples were not detected by the UltraSEEK panel. This may suggest a limit of detection for some assays in the UltraSEEK panel. However, when plotted according to their frequency abundance, the undetected mutations were neither at the lowest concentrations nor lowest frequency abundances (Figure 2). In one case, P9, a BRAF V600E mutation was identified by the UltraSEEK panel, whereas ddPCR indicated the presence of a BRAF V600K mutation (Table 1 and Figure 2). The tumor from this patient also contained a BRAF V600K mutation, confirming a false-positive call by UltraSEEK. Ten mutations detected by UltraSEEK, but not confirmed by ddPCR, were the DPH3 promoter 8C>T (n = 5), CTNNB1 S45P (n = 1), and BRAF V600E (n = 4) (Table 1). The latter may suggest low-level false positives or cross contamination with the UltraSEEK assays, because most were detected with minimal mutant signal intensity. To further evaluate the specificity and sensitivity of the UltraSEEK melanoma panel, 49 plasma samples obtained from 20 healthy donors and 29 melanoma patients were tested next. Of the 29 melanoma patients, 22 were indicated to be negative for BRAF mutations in the archival pathology reports, and 7 were BRAF V600E or V600K positive in their tumor, but found negative in the blood sample by ddPCR. None of the seven plasma samples from the plasma-negative BRAF mutant patients were scored as positive by UltraSEEK, indicating the absence of detectable ctDNA in these samples by both UltraSEEK as well as ddPCR. Six of the 22 BRAF wild-type samples were found to carry mutations using the UltraSEEK panel (Supplemental Table S3). One had a BRAF V600E mutation and one a DPH3 promoter mutation that were not found by ddPCR. Two had NRAS Q61K/DPH3 C>T and NRAS Q61K/CDKN2A R80X mutations that were confirmed by ddPCR. The YAE1D1_c39605965G>A mutation found in sample E40, but was not tested by ddPCR. To perform unbiased assessment of concordance, both methods were compared for the detection of BRAF V600E in 77 samples that were tested for this mutation in both platforms. A substantial agreement was observed with a Cohen's κ coefficient 0.826 (bias-corrected and accelerated 95% CI, 0.669–0.946) (Table 2).Table 2BRAF Mutation Detection Concordance between PlatformsddPCRTotalUndetectedDetectedUltraSEEK UndetectedCount48351Expected count33.817.251.0 DetectedCount32326Expected count17.28.826.0Total Count512677 Expected count51.026.077.0 Open table in a new tab Longitudinally collected plasma from three melanoma patients treated with PD-1 inhibitors (pembrolizumab or nivolumab) were analyzed next. Figure 3 shows that the normalized intensities for BRAF V600E detected by UltraSEEK (right y axis) correlated with copies by ddPCR (left y axis), and with changes in disease status, declining with ongoing response to treatment and rising upon disease progression. These results demonstrate that the panel may be used for noninvasive disease monitoring. Advancements in ultrasensitive genotyping methods have created great interest in the application of somatic mutation detection from plasma DNA as a “liquid biopsy” for individualized patient management.23Diaz Jr., L.A. Bardelli A. Liquid biopsies: genotyping circulating tumor DNA.J Clin Oncol. 2014; 32: 579-586Crossref PubMed Scopus (1527) Google Scholar In the absence of a patient's tumor genotype, there is a need to accurately screen for multiple mutations from a blood sample at low mutant abundance and with small amounts of DNA input. Here, a comprehensive UltraSEEK panel, specifically designed for the detection of melanoma-associated mutations, was assessed. Samples analyzed in the study were known to carry mutations in plasma by ddPCR (K.H., B.C., and D.I. were blinded to ddPCR results) but were blinded in the UltraSEEK analysis (E.S.G, T.W., M.P., L.C., M.A.K., J.R., A.H., J.C., S.S., G.A.M., M.M., M.Z., A.D., and S.Q.W. were blinded to UltraSEEK results), including a set of 20 healthy control samples used for assessment of specificity. The UltraSEEK Oncogene Panel assay uses a mass spectrometer for detection and does not require the accessory equipment and support often needed with next-generation sequencing–derived data. However, it is still able to interrogate multiple informative variants within a single reaction. The UltraSEEK chemistry is amenable to a manual workflow, but is also compatible with high-throughput processes using various automated liquid dispensing platforms. UltraSEEK differs from similar biochemistries in that it enriches the minor alleles by probing them specifically in a post-PCR primer extension step that omits the wild-type allele.24Mosko M.J. Nakorchevsky A.A. Flores E. Metzler H. Ehrich M. van den Boom D.J. Sherwood J.L. Nygren A.O. Ultrasensitive detection of multiplexed somatic mutations using MALDI-TOF mass spectrometry.J Mol Diagn. 2016; 18: 23-31Abstract Full Text Full Text PDF PubMed Scopus (53) Google Scholar Therefore, UltraSEEK can only provide semiquantitative measurement of the mutant allele. By comparison, methods like ddPCR provide absolute quantification of copies per volume of plasma of the mutant allele. Cohen's κ coefficient was used to compare both methodologies. The Cohen κ coefficient represents a considerable improvement over percent agreement calculations because the κ statistic provides a quantitative measure of agreement that has been adjusted for the degree of agreement expected solely on the basis of chance.25Viera A.J. Garrett J.M. Understanding interobserver agreement: the kappa statistic.Fam Med. 2005; 37: 360-363PubMed Google Scholar In addition, the κ coefficients were used with bias-corrected and accelerated 95% CIs, which automatically adjusts for bias and skewness in the bootstrap distribution.26Kang C. Qaqish B. Monaco J. Sheridan S.L. Cai J. Kappa statistic for clustered dichotomous responses from physicians and patients.Stat Med. 2013; 32: 3700-3719Crossref PubMed Scopus (22) Google Scholar BRAF mutations were only analyzed for this comparison for two reasons: all samples were analyzed for BRAF mutations in both assays and, secondly, many of the discordant results between the two platforms were observed in BRAF V600E mutations. Although no healthy donor samples were found to contain melanoma-associated mutations, multiple melanoma samples were found to have mutations that were not confirmed by ddPCR. Given this potential for false positives in the current assay design, we would recommend any putative mutations detected by this assay to be confirmed using an orthogonal method, for example, ddPCR or samples could be run in duplicate reactions. Moreover, this study was performed at the time when the panel was still being optimized by Agena Bioscience. A new version of the UltraSEEK panel is now available aiming to provide better specificity and sensitivity across all mutations. In addition to the UltraSEEK panel performance, the observed discordant results may be attributed to the handling of the samples during shipment between laboratories and to differences between sample processing, because in some cases, a separate plasma aliquot from the same blood collection time point was extracted to be analyzed by ddPCR. However, no specific event was identified to explain the discordant results (Supplemental Table S2), because all three samples where UltraSEEK failed to detect the BRAF V600E mutations were processed within 2 hours. Pre-analytical variables may have potentially confounded the concordance of the results in this study, for example, freeze-thawing of the same sample tested leading to potential degradation. These results therefore highlight that proper quality control assessment of a plasma DNA sample is warranted to ensure that the optimal amount of template and integrity of a sample are acceptable before testing. Although the screening of mutations in the TERT promoter were not possible in this current assay due to the GC richness of this loci, other highly recurrent promoter mutations were included in the panel, including DPH3 and RPS27, allowing serial mutation tracking in patients whose tumors are BRAF/NRAS wild type. The panel can also be useful in the detection of resistance to mitogen-activated protein kinase inhibitors, for example, NRAS mutations as highlighted in Patient P16. In conclusion, our results indicate that the UltraSEEK melanoma panel is as sensitive as ddPCR for the detection of ctDNA. This highly multiplexed assay allows for rapid and sensitive screening for the detection of multiple melanoma–associated mutations in plasma. Download .docx (.01 MB) Help with docx files Supplemental Table S1 Download .docx (.02 MB) Help with docx files Supplemental Table S2 Download .docx (.02 MB) Help with docx files Supplemental Table S3