Next-Generation Genotyping by Digital PCR to Detect and Quantify the ...

The Journal of Molecular Diagnostics, Vol. 17, No. 4, July 2015

jmd.amjpathol.org

Next-Generation Genotyping by Digital PCR to Detect and Quantify the BRAF V600E Mutation in Melanoma Biopsies Pierre-Jean Lamy,*y Florence Castan,z Nicolas Lozano,* Cécile Montélion,* Patricia Audran,* Frédéric Bibeau,yx Sylvie Roques,* Frédéric Montels,* and Anne-Claire Laberenne* From the Departments of Biology and Oncogenetics,* Statistics,z and Pathology,x and the Biobank,y Institut Régional du Cancer de Montpellier, Montpellier, France Accepted for publication February 18, 2015. Address correspondence to Pierre-Jean Lamy, Pharm.D., Ph.D., Department of Biology and Oncogenetics, Montpellier Cancer Institute, Rue de la Croix Verte, Montpellier, F-34298 France. E-mail: pierre-jean.lamy@icm. unicancer.fr.

The detection of the BRAF V600E mutation in melanoma samples is used to select patients who should respond to BRAF inhibitors. Different techniques are routinely used to determine BRAF status in clinical samples. However, low tumor cellularity and tumor heterogeneity can affect the sensitivity of somatic mutation detection. Digital PCR (dPCR) is a next-generation genotyping method that clonally amplifies nucleic acids and allows the detection and quantification of rare mutations. Our aim was to evaluate the clinical routine performance of a new dPCR-based test to detect and quantify BRAF mutation load in 47 paraffin-embedded cutaneous melanoma biopsies. We compared the results obtained by dPCR with highresolution melting curve analysis and pyrosequencing or with one of the allele-specific PCR methods available on the market. dPCR showed the lowest limit of detection. dPCR and allele-specific amplification detected the highest number of mutated samples. For the BRAF mutation load quantification both dPCR and pyrosequencing gave similar results with strong disparities in allele frequencies in the 47 tumor samples under study (from 0.7% to 79% of BRAF V600E mutations/sample). In conclusion, the four methods showed a high degree of concordance. dPCR was the more-sensitive method to reliably and easily detect mutations. Both pyrosequencing and dPCR could quantify the mutation load in heterogeneous tumor samples. (J Mol Diagn 2015, 17: 366e373; http://dx.doi.org/10.1016/j.jmoldx.2015.02.004)

Mutations and chromosomal translocations underlie key molecular mechanisms that can drive cancer development/ progression. Therefore, treatment strategies that target specific molecules related to gene mutations or chromosome rearrangements were developed. For instance, the BRAF gene encodes the BRAF protein, a cytoplasmic serine/threonine protein kinase that is involved in regulating signaling pathway. BRAF gene mutations are detected in several malignancies, including colorectal, lung, and thyroid cancer and in nearly all cases of classic hairy-cell leukemia.1e3 BRAF somatic missense mutations are also present in most malignant melanomas. Specifically, approximately 75% of melanoma samples harbor the BRAF V600E-activating mutation.4 Importantly, treatment with the BRAF inhibitor vemurafenib shows an 80% response rate in patients with BRAF V600E-positive Copyright ª 2015 American Society for Investigative Pathology and the Association for Molecular Pathology. Published by Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.jmoldx.2015.02.004

metastatic melanoma.5 The identification of diagnostic somatic chromosomal/genetic abnormalities, such as the BRAF V600E-activating mutation, is thus becoming of crucial importance in the field of personalized cancer treatments, because it can guide the therapeutic decisions and can predict therapeutic responses. Many techniques can be used to detect somatic mutations. The gold standard for mutation detection is the Sanger technique that is performed on automated DNA sequencing instruments. However, the low sensitivity of somatic mutation Supported by grants from the French National Institute of Cancer (INCa) and the Montpellier Cancer Institute (to Department of Biology and Oncogenetics as a Génétique Moléculaire des Cancers facility) and Qiagen unconditional grant (P-J.L.). Disclosures: None declared.

BRAF Mutation Detection by Digital PCR detection by direct sequencing could lead to false-negative results, especially in cancer samples with low tumor cellularity. Another challenge for sequencing is the high intratumor genetic heterogeneity.6 High-resolution, high-throughput next-generation sequencing methods allow detecting mutations even when present at low level or in case of somatic mutational heterogeneity within a single tumor sample.6e8 Allele-specific PCR and pyrosequencing can detect at best 1% to 5% of mutated alleles in wild-type (wt) genomic DNA (gDNA).9 This is a huge improvement compared with the 10% to 20% sensitivity of the Sanger method; however, much effort is currently focused on the development of even more-sensitive and easy-to-use methods. Digital PCR (dPCR) is a technically simple method for the quantification of the total number of initial DNA targets present in a sample.10 The principle of the method is a series of classic PCR end point measurements in microreactors or nanoreactors to provide nucleic acid quantification without the use of standards curves. For that, the sample is randomly partitioned in multiple replicate amplification reactions that contain zero, one, or more template copies and PCR-amplified to the end point. Then, nucleic acids may be quantified by counting the reactions that contain PCR end products. Because the distribution of the targeted nucleic acid in the reactors follows a Poisson distribution, the number of nucleic acids in the sample can be estimated with the following equation: l Z ln (1  P), where l is the average number of target DNA molecules per replicate reaction and P is the fraction of positive end point reactions.11 This method allows the quantification of the mutation load of tumor samples.12 Another advantage of this technique is the independence from the number of amplifications and from the efficiency variability observed in the exponential phase of PCR. This is in relation with the method sensitivity that allows the detection of rare targets. In classical or real-time PCR, during the first cycles of amplification a natural selection leads to amplification of the most frequent nucleic acids.13 Conversely, in dPCR, there is no competition among target nucleic acids, and even rare nucleic acids (as low as 0.01%) are detected at a rate that corresponds to their frequency in the initial sample. Currently, three approaches are used by commercially available dPCR systems. The first approach uses microfluidic chambers or microwells to split the samples,14 and the second, called BEAMing, is based on emulsion PCR to clonally amplify templates in the presence of beads.15 In the third method (used in this study), water-in-oil droplets are generated by loading oil, samples, and PCR reagents in a microfluidic circuit. Then, the PCR amplification is performed on a classic thermocycler, and the fluorescence related to each PCR product is detected in every droplet by an automated droplet flow cytometer.16 Our aim was to evaluate, using 47 paraffin-embedded cutaneous melanoma samples, the clinical routine performance

The Journal of Molecular Diagnostics

-

jmd.amjpathol.org

of the new dPCR-based BRAF load mutation quantification test we developed compared with high-resolution melting (HRM) curve analysis, pyrosequencing, and allele-specific PCR.

Materials and Methods Patients and Samples A total of 47 metastatic melanoma biopsies referred to our oncogenetic laboratory for BRAF mutation testing between January 2012 and June 2013 were analyzed. Biopsies were fixed in ethanol, formalin, and acetic acid for 12 to 24 hours and then embedded in paraffin. Hematoxylin and eosine stained sections were used to identify tumor areas within each biopsy. The regions that contained the greatest amount of tumor cells were located under a microscope, and their contour was highlighted with a marker pen by a pathologist.

DNA Extraction gDNA was extracted from paraffin-embedded melanoma biopsies by using the QIAamp DNA FFPE Kit (Qiagen GmbH, Hilden, Germany) according to the manufacturer’s protocol. DNA yield and purity were assessed by measuring the absorbance at 260 nm and 280 nm with a Nanodrop spectrophotometer (Thermo Scientific, Wilmington, DE). All samples had a 260:280 nm ratio higher than 1.7. DNA was stored at 20 C in buffer of 10 mmol/L Tris and 0.5 mmol/L EDTA (pH 7.6).

BRAF V600E Mutation Analysis gDNA from the IGR37 cell line (BRAF V600E mutation) was used as positive control, and DNA from the LNCaP and SW620 cell lines (wt BRAF) was used as negative control. Water was used as negative control for PCR contamination. Normal human DNA without mutations in the BRAF gene (Roche Diagnostics GmbH, Mannheim, Germany) was used to prepare the serial dilutions of gDNA from IGR37 cells to determine the sensitivity of dPCR. HRM Analysis PCR amplification and HRM analysis were performed on a Rotor-Gene 6000 (Corbett Research, Mortlake, NSW, Australia) by using the Light Cycler 480 HRM Master MIX kit (Roche Diagnostics, Meylan, France). Primers were designed to amplify BRAF fragments that span the region of exon 15 in which the BRAF V600E mutation is located: BRAF exon 15 forward Z 50 -bio CCACAAAATGGATCCAGACA-30 and BRAF exon 15 reverse Z 50 -TTCATGAAGACCTCACAGTAAAAA-30 . PCR amplifications were performed in a final volume of 20 mL that included 10 ng of purified gDNA, 10 mL of 2 PCR mix, 2.4 mL of 25 mmol/L MgCl2, and 0.4 mL of each 10 mmol/L forward and reverse primer.

367

Lamy et al Table 1

Quantification of the Limit of BRAF V600E Mutation Detection by dPCR Therascreen BRAF Pyro

Therascreen BRAF RQG PCR

Results

MT, %

Result

DCt

Result

MT MT MT MT MT MT MT MT MT MT MT WT

0 e 34 26 5 3 0 0 0 0 0 0 0 0

WT e MT MT MT MT WT WT WT WT WT WT WT WT

0 e 1.72 2.18 6.63 6.46 8.33 9.51 9.6 10 9.89 10.1 10.6 10.4

WT e MT MT MT MT WT WT WT WT WT WT WT WT

dPCR DNA gDNA NTC IGR37 IGR37 IGR37 IGR37 IGR37 IGR37 IGR37 IGR37 IGR37 IGR37 IGR37 IGR37

Dil e

1

1 2 10 20 40 80 160 320 640 1280 2560 5120

MT, %

GE

Droplets FAM, % (n)

Droplets VIC, % (n)

0 0 50 25 5 2.50 1.25 0.625 0.312 0.156 0.0780 0.0390 0.0195 0.0098

60,606 0 30,303 15,152 3030 1515 758 379 189 95 47 24 12 6

0 0 63.3 23.6 4.39 2.31 0.99 0.46 0.24 0.17 0.077 0.038 0.023 0

2990 0 36.7 76.4 95.61 97.69 99.01 99.54 99.76 99.83 99.92 99.96 99.97 100

(0) (7741) (4475) (1092) (555) (246) (123) (62) (46) (19) (10) (6) (0)

(16,734) (0.1) (4775) (13,312) (16,952) (16,400) (16,628) (18,072) (17,209) (18,560) (16,554) (17,858) (17,425) (17,597)

Ct, cycle threshold; Dil, dilution of mutated DNA in wild-type DNA (gDNA); dPCR, digital PCR; GE, genome equivalent; MT, mutated; NTC, no template control; Pyro, pyrosequencing; WT, wild-type.

The following cycling conditions were used: one cycle at 95 C for 5 minutes, 50 cycles at 95 C for 15 seconds and 63 C for 15 seconds with a touchdown program of 0.5 C/ cycle during the first 11 cycles, 72 C for 20 seconds. The melting conditions included one cycle at 95 C for 1 minute, one cycle at 40 C for 1 minute, and one cycle at 65 C for 2 seconds, followed by a melt from 65 C to 95 C, rising at a rate of 0.1 C per second. HRM data were analyzed with the Rotor-Gene 6000 version 1.7 (Corbett Research). For each sample, the normalized melting curves were evaluated, and samples were compared with the negative controls (no mutation: gDNA from LNCaP and SW620 cells) in a deduced difference plot. Significant deviations from the horizontal line relative to the spread of the wt controls were indicative of sequence changes within the analyzed amplicon. Results were given in percentage of homology. According to our experience, a percentage of homology 7 were classified as negative or beyond the limits of detection of the kit. Samples with a DCt value 7 were considered as positive.

Statistical Analysis Data were summarized by using frequencies and percentages for categorical variables or means and SDs for continuous variables. The percentages of BRAF V600E mutations measured by dPCR and by pyrosequencing were compared with the paired Student’s t-test. The association between these two methods was described with linear regression and Spearman correlations coefficient. Statistical analyses were performed with Stata 13 (Stata Corporation, College Station, TX).

The Journal of Molecular Diagnostics

-

jmd.amjpathol.org

Results Melanoma Samples Because of the careful selection of the tumor areas, all surgical specimens displayed >20% of tumor cells. Nine samples (19.2%) showed low tumor cellularity (20% to 50%), and 38 samples (80.8%) had a tumor cellularity >50%. The yield of gDNA extracted from the paraffinembedded melanoma samples ranged between 158 and 994 ng/mL (median, 501 ng/mL; effective concentration, 214 ng/ mL). All tumor DNA samples were amplified with acceptable Ct values (mean, 28.86; range, 22.78 to 32.64).

Sensitivity of the Methods To assess the limit of detection of the different methods, serial dilutions (from 1/2 to 1/5120) of gDNA extracted from IGR37 cells (BRAF V600E mutation) in gDNA were prepared. dPCR could detect the BRAF V600E mutation in samples, including 0.0195% mutated allele (Table 1 and Figure 1). At this dilution the CV was 61%. The limit of detection for pyrosequencing and allele-specific amplification was 2.5%. To assess dPCR limit of quantification (the lowest concentration that results in a CV Z 20%), analyses were performed in quadruplicate, and the CV was calculated for each dilution. CV ranged from 3% to 19% for dilutions from 1/2 to 1/160 and was >20% for dilutions >1/160. Consequently, the limit of quantification for dPCR was 0.3%.

Rate of BRAF V600E Detection according to the Different Methods The detection rate of the BRAF V600E mutation in the 47 melanoma samples according to the four methods is shown

369

Lamy et al Table 2

BRAF Genotyping in Melanoma Samples HRM*

Therascreen BRAF Pyroy

Therascreen BRAF RGQ PCRz

dPCRx

Samples

Cellularity, %

Ct, mean

Homology, %

Status

% MT

Result

DCt

Result

% MT

% WT

FAM/VIC, n/n

Result

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47

>50 >50 >50 >50 >50 >50 20e50 >50 >50 >50 >50 >50 >50 >50 >50 >50 >50 >50 20e50 >50 >50 >50 >50 >50 >50 20e50 >50 >50 >50 >50 >50 >50 >50 >50 >50 >50 >50 >50 >50 >50 >50 20e50 20e50 20e50 20e50 20e50 20e50

23.02 27.61 27.40 30.43 28.47 30.04 28.74 29.99 31.92 32.55 30.99 32.12 29.65 27.74 29.00 28.41 31.03 29.48 29.10 29.16 27.94 28.04 30.22 27.63 27.60 28.83 26.47 27.48 28.85 28.93 28.52 28.68 27.03 29.79 29.35 30.82 29.00 28.98 27.98 28.07 27.85 37.19 35.37 28.75 29.08 29.11 29.89

28.26 92.95 98.2 97.17 21.5 41.6 98.3 97.36 8.79 95.85 7.4 10.35 95.85 95.8 96.35 96.49 96.9 11.06 81.16 74.45 96.03 97.82 8.1 96.17 97.78 4.08 98.37 95.49 4.04 90.2 23.47 92.14 22.34 15.12 93.38 95.96 34.65 63.36 94.6 96.48 43.65 94.52 49.05 98.31 96.7 81.16 94.69

MT WT WT WT MT MT WT WT MT WT MT MT WT WT WT WT WT MT MT MT WT WT MT WT WT MT WT WT MT WT MT WT MT MT WT WT MT MT WT WT MT WT MT WT WT WT WT

25 0 0 0 39 12 2 1 55 0 49 61 1 0 0 0 1 68 3 0 0 0 50 0 0 56 0 1 59 1 19 2 22 56 1 1 68 6 0 1 72 0 19 2 0 4 0

MT WT WT WT MT MT WT WT MT WT MT MT WT WT WT WT WT MT MT WT WT WT MT WT WT MT WT WT MT WT MT WT MT MT WT WT MT MT WT WT MT WT MT WT WT MT WT

1.27 11.6 12.6 >7 0.72 1.37 5.44 >7 0.16 >7 0.51 0.53 12 12.2 >7 >7 >7 0.83 3.54 12.3 12.3 >7 0.78 12.8 11.7 0.48 8.78 8.64 0.3 8.58 0.32 7.3 1.44 1.51 8.58 8.57 0.78 3.02 8.92 8.58 0.38 >7 0.45 6.35 >7 4.29 >7

MT WT WT WT MT MT MT WT MT WT MT MT WT WT WT WT WT MT MT WT WT WT MT WT WT MT WT WT MT WT MT WT MT MT WT WT MT MT WT WT MT WT MT MT WT MT WT

22 0 0 0 37 14 2 0 57 0 48 67 0 0 0 0 0 69 3 0 0 0 55 0 0 23 0 0 63 0 22 0 22 66 0 0 78 5 0 0 79 0.7 20 1.5 0 2.8 0

78 100 100 100 63 86 98 100 43 100 52 33 100 100 100 100 100 31 97 100 100 100 45 100 100 77 100 100 37 100 78 100 78 34.3 100 100 22 95 100 100 21 99.3 80 98.5 100 97.2 100

221/794 0.1/1060 0.1/2710 0/719 627/1090 56.4/356 26.8/1680 0.09/660 643/494 0/1030 926/993 283/142 0/2320 0/2270 0/1790 0/1740 0/919 1310/579 91.6/3490 0.1/1830 0.1/1630 0.1/1250 186/153 0/2200 0/1930 120/411 0/1125 0/691 622/373 0/1630 102/362.5 0/283.5 386/1400 118.5/62 0/312 0/695 318/91 25.4/481 0/708 0/862 691/187 1/142 64/256 286/14,325 1/8066 778/18,318 0/1792

MT WT WT WT MT MT MT WT MT WT MT MT WT WT WT WT WT MT MT WT WT WT MT WT WT MT WT WT MT WT MT WT MT MT WT WT MT MT WT WT MT MT MT MT WT MT WT

*Homology < 90% Z mutated, homology > 90% Z wild-type. y Therascreen BRAF pyro: 2% Z wild-type, >2% Z mutated. z Therascreen BRAF RGQ PCR: DCt  7 Z mutated, DCT > 7 Z wild-type. x dPCR: % MT > 0.3% Z mutated; % MT < 0.3% Z wild-type. dPCR, digital PCR; HRM, high-resolution melting; MT, mutated; PYRO, pyrosequencing; WT, wild-type.

in Table 2. With HRM, samples with a rate of homology 50%) presented a mutated HRM profile but was considered as wt by the three other methods. We confirmed by Sanger sequencing that no sequence variation was present in this sample.

Determination of the Mutation Load by dPCR and Pyrosequencing The BRAF V600E mutation load of a tumor sample can be estimated by using quantitative methods, such as dPCR or pyrosequencing. Globally, the mutation load in the melanoma samples under study was not statistically different between these methods (P Z 0.97). The mutation load obtained by dPCR and pyrosequencing were highly correlated (Spearman’s correlation coefficient Z 0.9143). The regression line for the mutation load obtained with the two methods was close to the theoretical line with a slope equal to 1 and an intercept of 0.47 (Figure 2).

Discussion Previous studies showed that the detection rate of the BRAF V600E mutation in formalin-fixed, paraffin-embedded melanoma biopsies ranges from 22% to 72%.17 This variation can be mainly explained by patients’ selection in the different cohorts but also by the different sensitivity of the methods used for the detection and/or the percentage of tumor cells

The Journal of Molecular Diagnostics

-

jmd.amjpathol.org

present in the samples. To improve the detection of potential BRAF mt subclones in melanoma samples, we developed a highly sensitive dPCR method that also allows the quantification of the mutational load in melanoma samples, and we compared the results with the strategy we currently use in clinical routine (HRM þ pyrosequencing) and also with an allele-specific amplification technique we routinely use to analyze samples with discrepancies or with noninterpretable results after HRM and pyrosequencing analyses. The method we developed is specific for the BRAF V600E mutation and cannot be used to detect other BRAF V600 mutations. dPCR presented the highest sensitivity with a limit of detection determined by serial dilution of mutated DNA in wt gDNA equal to 0.0195%, whereas the limit of quantification was 0.3%.18 In our series, the detection rate was 46.8% for dPCR and allele-specific amplification and 40.4% for HRM and pyrosequencing. Discrepancies among the four methods were mostly observed for samples with low tumor cellularity, with the exception of sample 20 (tumor cellularity >50%) that gave a false-positive profile by HRM. Altogether, our findings indicate that the frequency of BRAF V600E mutation detection is influenced, particularly for low tumor cellularity samples, by the analytic sensitivity of the used method when sensitive techniques are used. Despite the same limit of detection, allelespecific amplification seems to be more sensitive than pyrosequencing. This good concordance among the different methods could be due to the systematic macrodissection of all biopsies, leading to the selection of tumor areas with a high percentage of tumor cells in samples and also to the use of highly sensitive methods. It is generally admitted that the Sanger method presents a 10% to 20% sensitivity, and studies that compared more-sensitive techniques with the Sanger method reported higher numbers of discordant results.18 We confirm that PCR amplification-based tests are currently the more-sensitive methods available on the market.19 We also show that dPCR is even more sensitive and could also be more efficient than pyrosequencing, a method that was previously qualified as the most-efficient technique for the detection of BRAF mutations in melanoma samples,20 particularly in biopsies with low tumor cellularity. Moreover, highly sensitive methods, such as dPCR, could avoid false-negative results related to intratumor heterogeneity,11 when assessing samples enriched in tumor cells. The intrapatient tumor heterogeneity in melanomas was recently questioned; however, it seems that this question is inherent to the method sensitivity.21,22 Until now, only qualitative methods were used to evaluate BRAF mutation status in clinical trials. However, it could be interesting to quantify also the mutation load. Indeed, the tumor cellularity (percentage of tumor cells relative to all of the cells of each tissue section) and the rate of mutations detected in the DNA extracted from the corresponding biopsies are not comparable.23 As intratumor heterogeneity was described in melanomas, different clones harboring different molecular profiles (wt or mt BRAF, for instance) could be

371

Lamy et al present in the same tumor. Moreover, one of the mechanisms underlying anti-BRAF V600E treatment could be the selection of tumor clones that express wt BRAF. Therefore, the question of whether the rate of mutated tumor cells can influence the tumor response to anti-BRAF therapies could be relevant. In addition, BRAF inhibitors can activate the mitogen-activated protein kinase pathway in wt BRAF cells, and this, in turn, induces cell proliferation and may promote the development of cutaneous squamous cell carcinoma or keratoacanthoma.24 It would be interesting to know whether a correlation exists between the tumor mutational load and the occurrence of this serious adverse effect. Both pyrosequencing and dPCR are quantitative methods. Because the fluorescence data after amplification can be fitted with the Poisson distribution, dPCR enables the absolute quantitation of nucleic acids in a sample. Pyrosequencing, a technology based on the principle of sequencing by synthesis detects by bioluminescence which base is added at each step of the synthesis of the new strand and generates a pyrogram with a peak for each nucleotide. Differences in peak heights are in relation with the presence of a mutation, and the peak rate allows the relative quantification of alleles. Surprisingly, both absolute and relative quantitative methods showed a good correlation in our study. We observed a great disparity in the mutation load in our samples (0.7% to 79%). The clinical impact of this disparity remains to be investigated, including whether the response to BRAF inhibitors in patients with a low mutation rate is altered.

Conclusion The four assessed methods show a good concordance, especially when samples with high tumor cellularity are analyzed. Nevertheless, dPCR allowed the detection of the highest number of mutated samples. dPCR could be used for BRAF V600E mutation quantification in melanoma biopsies. The higher sensitivity of this method could allow investigating tumor cell mutation heterogeneity and its changes during tumor progression in a more thorough way than with the currently used sequencing methods. dPCR is easy to perform and economically competitive. Moreover, because of its high sensitivity and the use of short amplicons, it could be also used for the analysis of circulating DNA, opening the way to the development of new strategies for the patients’ follow-up during treatment.25

Acknowledgments We thank Hélène Frugier for technical assistance and Elisabetta Andermarcher for editing the manuscript.

References 1. Davies H, Bignell GR, Cox C, Stephens P, Edkins S, Clegg S, et al: Mutations of the BRAF gene in human cancer. Nature 2002, 417: 949e954

372

2. Tiacci E, Trifonov V, Schiavoni G, Holmes A, Kern W, Martelli MP, Pucciarini A, Bigerna B, Pacini R, Wells VA, Sportoletti P, Pettirossi V, Mannucci R, Elliott O, Liso A, Ambrosetti A, Pulsoni A, Forconi F, Trentin L, Semenzato G, Inghirami G, Capponi M, De Raimondo F, Patti C, Arcaini L, Musto P, Pileri S, Haferlach C, Schnittger S, Pizzolo G, Foa R, Farinelli L, Haferlach T, Pasqualucci L, Rabadan R, Falini B: BRAF mutations in hairy-cell leukemia. N Engl J Med 2011, 364:2305e2315 3. Sobrinho-Simoes M, Preto A, Rocha AS, Castro P, Maximo V, Fonseca E, Soares P: Molecular pathology of well-differentiated thyroid carcinomas. Virchows Arch 2005, 447:787e793 4. Greaves WO, Verma S, Patel KP, Davies MA, Barkoh BA, Galbincea JM, Yao H, Lazar AJ, Aldape KD, Medeiros LJ, Luthra R: Frequency and spectrum of BRAF mutations in a retrospective, single-institution study of 1112 cases of melanoma. J Mol Diagn 2013, 15:220e226 5. Ravnan MC, Matalka MS: Vemurafenib in patients with BRAF V600E mutation-positive advanced melanoma. Clin Ther 2012, 34: 1474e1486 6. Swanton C: Intratumor heterogeneity: evolution through space and time. Cancer Res 2012, 72:4875e4882 7. Gerlinger M, Rowan AJ, Horswell S, Larkin J, Endesfelder D, Gronroos E, Martinez P, Matthews N, Stewart A, Tarpey P, Varela I, Phillimore B, Begum S, McDonald NQ, Butler A, Jones D, Raine K, Latimer C, Santos CR, Nohadani M, Eklund AC, Spencer-Dene B, Clark G, Pickering L, Stamp G, Gore M, Szallasi Z, Downward J, Futreal PA, Swanton C: Intratumor heterogeneity and branched evolution revealed by multiregion sequencing. N Engl J Med 2012, 366:883e892 8. Shah SP, Roth A, Goya R, Oloumi A, Ha G, Zhao Y, et al: The clonal and mutational evolution spectrum of primary triple-negative breast cancers. Nature 2012, 486:395e399 9. Gonzalez de Castro D, Angulo B, Gomez B, Mair D, Martinez R, Suarez-Gauthier A, Shieh F, Velez M, Brophy VH, Lawrence HJ, Lopez-Rios F: A comparison of three methods for detecting KRAS mutations in formalin-fixed colorectal cancer specimens. Br J Cancer 2012, 107:345e351 10. Sykes PJ, Neoh SH, Brisco MJ, Hughes E, Condon J, Morley AA: Quantitation of targets for PCR by use of limiting dilution. Biotechniques 1992, 13:444e449 11. Hindson BJ, Ness KD, Masquelier DA, Belgrader P, Heredia NJ, Makarewicz AJ, et al: High-throughput droplet digital PCR system for absolute quantitation of DNA copy number. Anal Chem 2011, 83: 8604e8610 12. Yeh I, von Deimling A, Bastian BC: Clonal BRAF mutations in melanocytic nevi and initiating role of BRAF in melanocytic neoplasia. J Natl Cancer Inst 2013, 105:917e919 13. Bizouarn F: Clinical applications using digital PCR. Methods Mol Biol 2014, 1160:189e214 14. Warren L, Bryder D, Weissman IL, Quake SR: Transcription factor profiling in individual hematopoietic progenitors by digital RT-PCR. Proc Natl Acad Sci U S A 2006, 103:17807e17812 15. Dressman D, Yan H, Traverso G, Kinzler KW, Vogelstein B: Transforming single DNA molecules into fluorescent magnetic particles for detection and enumeration of genetic variations. Proc Natl Acad Sci U S A 2003, 100:8817e8822 16. Beer NR, Hindson BJ, Wheeler EK, Hall SB, Rose KA, Kennedy IM, Colston BW: On-chip, real-time, single-copy polymerase chain reaction in picoliter droplets. Anal Chem 2007, 79:8471e8475 17. Lee JH, Choi JW, Kim YS: Frequencies of BRAF and NRAS mutations are different in histological types and sites of origin of cutaneous melanoma: a meta-analysis. Br J Dermatol 2011, 164:776e784 18. Lade-Keller J, Romer KM, Guldberg P, Riber-Hansen R, Hansen LL, Steiniche T, Hager H, Kristensen LS: Evaluation of BRAF mutation testing methodologies in formalin-fixed, paraffin-embedded cutaneous melanomas. J Mol Diagn 2013, 15:70e80 19. Lopez-Rios F, Angulo B, Gomez B, Mair D, Martinez R, Conde E, Shieh F, Vaks J, Langland R, Lawrence HJ, de Castro DG: Comparison of testing methods for the detection of BRAF V600E mutations in

jmd.amjpathol.org

-

The Journal of Molecular Diagnostics

BRAF Mutation Detection by Digital PCR malignant melanoma: pre-approval validation study of the companion diagnostic test for vemurafenib. PLoS One 2013, 8:e53733 20. Colomba E, Helias-Rodzewicz Z, Von Deimling A, Marin C, Terrones N, Pechaud D, Surel S, Cote JF, Peschaud F, Capper D, Blons H, Zimmermann U, Clerici T, Saiag P, Emile JF: Detection of BRAF p.V600E mutations in melanomas: comparison of four methods argues for sequential use of immunohistochemistry and pyrosequencing. J Mol Diagn 2013, 15:94e100 21. Heinzerling L, Baiter M, Kuhnapfel S, Schuler G, Keikavoussi P, Agaimy A, Kiesewetter F, Hartmann A, Schneider-Stock R: Mutation landscape in melanoma patients clinical implications of heterogeneity of BRAF mutations. Br J Cancer 2013, 109:2833e2841 22. Menzies AM, Wilmott JS, Long GV, Scolyer RA: Intra-patient heterogeneity of BRAF mutation status: fact or fiction? Br J Cancer 2014, 111: 1678e1679

The Journal of Molecular Diagnostics

-

jmd.amjpathol.org

23. Halait H, Demartin K, Shah S, Soviero S, Langland R, Cheng S, Hillman G, Wu L, Lawrence HJ: Analytical performance of a realtime PCR-based assay for V600 mutations in the BRAF gene, used as the companion diagnostic test for the novel BRAF inhibitor vemurafenib in metastatic melanoma. Diagn Mol Pathol 2012, 21: 1e8 24. Su F, Viros A, Milagre C, Trunzer K, Bollag G, Spleiss O, et al: RAS mutations in cutaneous squamous-cell carcinomas in patients treated with BRAF inhibitors. N Engl J Med 2012, 366:207e215 25. Thierry AR, Mouliere F, El Messaoudi S, Mollevi C, Lopez-Crapez E, Rolet F, Gillet B, Gongora C, Dechelotte P, Robert B, Del Rio M, Lamy PJ, Bibeau F, Nouaille M, Loriot V, Jarrousse AS, Molina F, Mathonnet M, Pezet D, Ychou M: Clinical validation of the detection of KRAS and BRAF mutations from circulating tumor DNA. Nat Med 2014, 20:430e435

373