Context.—A polymerase chain reaction–based companion diagnostic (cobas 4800 BRAF V600 Mutation Test) was recently approved by the US Food and Drug Administration to select patients with BRAF-mutant metastatic melanoma for treatment with the BRAF inhibitor vemurafenib.

Objectives.—(1) To compare the analytic performance of the cobas test to Sanger sequencing by using screening specimens from phase II and phase III trials of vemurafenib, and (2) to assess the reproducibility of the cobas test at different testing sites.

Design.—Specimens from 477 patients were used to determine positive and negative percent agreements between the cobas test and Sanger sequencing for detecting V600E (1799T>A) mutations. Specimens were evaluated with a massively parallel pyrosequencing method (454) to resolve discordances between polymerase chain reaction and Sanger results. Reproducibility of the cobas test was assessed at 3 sites by using 3 reagent lots and an 8-member panel of melanoma samples.

Results.—A valid cobas result was obtained for all eligible patients. Sanger sequencing had a failure rate of 9.2% (44 of 477). For the remaining 433 specimens, positive percent agreement was 96.4% (215 of 223) and negative percent agreement, 80% (168 of 210). Among 42 cobas mutation-positive/Sanger V600E-negative specimens, 17 were V600E positive and 24 were V600K positive by 454. The cobas test detected 70% of V600K mutations. In the reproducibility study, a correct interpretation was made for 100% of wild-type specimens and specimens with greater than 5% mutant alleles; V600E mutations were detected in 90% of specimens with less than 5% mutant alleles.

Conclusions.—The cobas test (1) had a lower assay failure rate than that of Sanger, (2) was more sensitive in detecting V600E mutations, (3) detected most V600K mutations, and (4) was highly reproducible.

The development of novel small-molecule inhibitors of BRAF is changing the standard of care for patients with metastatic melanoma. Until now, this disease has had a grim prognosis, with average survivals of 6 to 9 months, and little benefit from standard chemotherapy.1 However, the selective BRAF inhibitor vemurafenib (RG7204/PLX4032) has demonstrated impressive clinical activity in phase I, II, and III trials in patients with metastatic melanoma harboring BRAF mutations.2,3 By contrast, preclinical and limited clinical data suggest the compound has minimal activity in tumors lacking BRAF mutations.4 Thus, it is critical to have a robust and accurate BRAF mutation assay to select patients for vemurafenib.

There has been a clear unmet medical need for a robust, well-validated, and reproducible BRAF mutation test that can be performed on DNA extracted from formalin-fixed, paraffin-embedded (FFPE) tissue samples. Traditional direct (Sanger) sequencing has been widely used in clinical laboratories for mutation testing, including BRAF mutations,5 but this method suffers from limited sensitivity for detecting mutations that are present in low percentages in a specimen.6 A number of laboratory-developed assays are currently available for detecting BRAF mutations, but the level of validation and characterization of the performance features of these tests is highly variable.

BRAF mutations have been reported to occur in 40% to 70% of melanoma cases, most often affecting codon 600 (reviewed in Ribas and Flaherty7). The most common codon 600 mutation in melanoma and other solid tumors is a single nucleotide mutation (c.1799T>A) that results in a valine-to-glutamic acid (V600E) conversion. These oncogenic V600E mutations result in a gain of function, rendering the RAF pathway constitutively active. In melanoma, the development of BRAF mutations appears to be an early event, occurring in many benign nevi.8 ,BRAF mutations are more prevalent in melanomas occurring in younger patients, and in areas of non–chronically sun-exposed areas of the body (eg, trunk, limbs).9 Although they appear to have no prognostic impact on localized melanoma, one study10 has suggested that BRAF-mutant melanomas may have a worse prognosis in the metastatic setting.

A number of less common variant codon 600 mutations have been reported to occur in malignant melanoma.11 The most frequent of these lesions are dinucleotide mutations that result in different amino acid substitutions for the valine at codon 600, for example, V600K (c.1798_1799GT>AA:p.Val600Lys), V600R (c.1798_1799GT>AG:p.Val600Arg), V600D (c.1799_1800TG>AC:p.Val600Asp), and V600M (c.1798G>A:Val600Met). However, one of the mutations, referred to as “V600E2” (c.1799_1800TG>AA:p. Val600Glu), results in the same amino acid substitution as the predominant V600E (c.1799T>A) mutation. The clinical implications of these rare genetic variants are uncertain. In the COSMIC (Catalogue of Somatic Mutations in Cancer) database, these non-V600E mutations comprise approximately 10% of all codon 600 mutations reported in melanoma.11 

As part of the clinical development plan for vemurafenib, we undertook in parallel the development of a companion diagnostic test, the cobas 4800 BRAF V600 Mutation Test, which was used to select patients for the pivotal phase II and phase III trials of the drug. The cobas test is designed for use with FFPE tissue samples and uses real-time polymerase chain reaction (PCR) amplification of target DNA with a complementary primer pair and 2 oligonucleotide probes labeled with different fluorescent dyes: 1 probe to detect the wild-type (WT) BRAF V600 sequence and 1 to detect the V600E mutation sequence.

The results of our internal analytic performance studies of the cobas BRAF test, including studies of analytic sensitivity, minimal tumor content, interfering substances (including melanin), and cross-reactivity with non-V600E mutations, have been described in a separate article.12 Here we present data on the multisite external validation studies of the assay performed in the context of the clinical trials of vemurafenib in metastatic melanoma. Using specimens from the 2 pivotal clinical trials, we compared the analytic performance of the cobas test with conventional Sanger sequencing and used quantitative massively parallel pyrosequencing (454 GS-Titanium, 454 Life Sciences, Branford, Connecticut) to resolve discordances observed between the cobas test and Sanger sequencing. We also conducted a separate study to assess the reproducibility of the cobas test when used to test a panel of melanoma samples at 3 clinical laboratories with different operators and reagent lots.

Clinical Comparator Study Assessing the Agreement Between the cobas Test and Direct Sanger Sequencing for the Detection of V600E Mutations

Study Population.—Five hundred forty-two (542) sequential patients who underwent screening for the phase II and phase III clinical trials of vemurafenib (RG7204/PLX4032) in metastatic melanoma were enrolled in this comparison study of the cobas test versus Sanger sequencing (Table 1). The prospective cobas testing for these 542 patients was performed at 3 laboratories—2 in the United States and 1 in Australia—according to a protocol approved by an institutional review board. A positive cobas test result was an inclusion criterion for both the phase II and phase III trials of vemurafenib. Specimens were retrospectively subjected to 2× bidirectional Sanger sequencing and, if indicated, quantitative massively parallel pyrosequencing (454 GS Titanium, hereafter referred to as 454 sequencing) at a single reference laboratory using validated methods.

Table 1. 

Study Population

Study Population
Study Population

Inclusion criteria for this comparison study of cobas testing versus Sanger sequencing included documentation of signed informed consent form, age 18 years or older, and histologically confirmed metastatic melanoma. Sixty-five patients failed to meet eligibility criteria, most often (60 of 65) because they lacked documentation of metastatic disease, leaving 477 eligible patients.

The demographic profiles of phase II and phase III evaluable patients were similar with respect to age, sex, and ethnicity (Table 2). Greater than 70% of patients tested were from North America, with a higher percentage in phase II than in phase III.

Table 2. 

Demographics, Collection Site Regions, and Test Site Regions of Evaluable Patients

Demographics, Collection Site Regions, and Test Site Regions of Evaluable Patients
Demographics, Collection Site Regions, and Test Site Regions of Evaluable Patients

Tumor Specimens.—The melanoma tumor specimens submitted for BRAF mutation screening for the phase II and phase III trials were either FFPE tissue blocks or consecutive 5-µM unstained sections mounted on slides from a single FFPE tissue block. If an archival specimen did not exist for a patient, a new biopsy was performed and the tumor tissue was subjected to standard formalin fixation and paraffin embedding. Both primary and metastatic lesions were permitted for mutation analysis.

cobas 4800 BRAF V600 Mutation Test.—The results of our internal studies of the analytic performance of this assay have been presented elsewhere.12 This assay uses complementary primers that define a 116–base pair sequence of human genomic DNA containing the BRAF codon 600 site in exon 15. The entire BRAF gene is not amplified. The test is designed to detect the nucleotide 1799 T>A change in the BRAF gene, which results in a valine-to-glutamic acid substitution at codon 600 (V600E). One probe is designed to detect the WT BRAF V600 sequence and one is designed to detect the V600E mutation sequence. Two external run controls are provided and the WT allele serves as an internal, full process control. The analytic time required to perform the test is approximately 8 hours for 24 samples run on 1 instrument.

cobas 4800 BRAF V600 Mutation Test Procedure.—A slide from the middle of the serial sections was stained with hematoxylin-eosin and examined by the laboratory pathologist to determine the need for macrodissection. If the sample contained at least 50% tumor cells, the tissue used for DNA isolation was processed without macrodissection. If the estimated percentage of tumor content was less than 50%, the tissue was macrodissected (if possible). The DNA from 1 section was isolated by using the cobas DNA Sample Preparation Kit (Roche Molecular Systems, Pleasanton, California), diluted to 5 ng/µL, and tested on a cobas 4800 System v2.0 (Roche Molecular Systems). Further details of the cobas DNA Sample Preparation Kit and the cobas 4800 BRAF V600 Mutation Test are provided in the package insert. The analytic time for the assay, including DNA isolation, amplification, and mutation detection, is less than 8 hours.

Sanger and 454 Sequencing.—Residual DNA extracts from all patients were subjected to retrospective 2× bidirectional Sanger sequencing at a single Clinical Laboratory Improvement Amendments of 1988 (CLIA)–certified central laboratory (SeqWright, Houston, Texas). This Sanger sequencing was based on automated dye-terminator chemistry and used software (ABI Sequencing Analysis Software version 5.1) and equipment supplied by Applied Biosystems (Foster City, California), following a validated sequencing protocol. Pass criteria required a contiguous Phred20 score of greater than 400. The estimated limit of detection of this sequencing method is 20% mutant alleles.

The same laboratory also performed 454 sequencing, using a validated method, on (1) all specimens discordant for codon 600 status by the cobas test and Sanger sequencing, (2) a randomly selected subset of concordant V600E-positive and concordant wild-type samples, and (3) samples for which Sanger sequencing gave invalid test results. Four hundred fifty-four sequencing is a quantitative, massively parallel pyrosequencing method run on the Genome Sequencer FLX system (454 Life Sciences). This technique relies on emulsion PCR on a fiber optic chip (PicoTiterPlate [454 Life Sciences]), and is based on sequencing by synthesis.13 For a valid mutation call, at least 3000 nucleotide reads of the target region were required per specimen, with a minimum of 500 reads in each direction. At least 10 reads had to contain the same nucleotide variation in order for the sample to be called mutant-positive. The estimated limit of detection of this sequencing method is 1% mutant alleles.

Agreement Analysis.—In the agreement analyses, the test performance was characterized by comparing the evaluable paired results between the cobas BRAF V600 Mutation Assay and Sanger sequencing (reference method). “Mutation detected” is defined as either having a mutation-detected result by cobas test or observing the presence of the predominant BRAF mutation type, V600E, as identified by bidirectional Sanger sequencing. Thus, positive percent agreement (PPA) is calculated as the percentage of mutation-detected cobas BRAF test results among all patients carrying the BRAF V600E mutation according to Sanger sequencing. Also, negative percent agreement (NPA) is calculated as the percentage of patients with mutation-not-detected test results among all those without V600E mutations by Sanger sequencing. Overall percent agreement (OPA) is the percentage of total patients for whom cobas BRAF test and the Sanger result agree. The 2-sided 95% Wilson score confidence intervals were calculated for these measures of agreement.14 

The reproducibility of the cobas test in the detection of the V600E (1799 T>A) mutation was evaluated by using an 8-member panel (see Table 3 and Figure, A through C) of malignant melanoma specimens, which included challenging sample types (eg, low tumor content, pigmentation). Testing was performed at 3 sites by using multiple (3) reagent lots and 2 operators per site. On 5 nonconsecutive days, operators performed blinded runs (2 replicates of each panel member/run), using cobas 4800 System v2.0.

Representative photomicrographs of malignant melanoma specimens used in the clinical reproducibility study, showing the range of pigmentation represented in the panel. A, Nonpigmented specimen (0%). B, Moderately pigmented specimen (20%). C, Heavily pigmented specimen (70%) (hematoxylin-eosin, original magnifications ×20).

Representative photomicrographs of malignant melanoma specimens used in the clinical reproducibility study, showing the range of pigmentation represented in the panel. A, Nonpigmented specimen (0%). B, Moderately pigmented specimen (20%). C, Heavily pigmented specimen (70%) (hematoxylin-eosin, original magnifications ×20).

Table 3. 

Clinical Reproducibility Panel Members

Clinical Reproducibility Panel Members
Clinical Reproducibility Panel Members

To prepare the panel members, FFPE melanoma specimens were obtained from external vendors and evaluated to determine BRAF mutation status, tumor content, degree of pigmentation, and percentage of mutant alleles as determined by 454 sequencing. The specimens were divided into 5-µm sections that were mounted individually on slides. DNA was isolated from the specimens according to the standard cobas test procedure described above. Seven of the 8 panel members were DNA extracts from individual melanoma tumors; the remaining member was a blend of WT and mutant DNA from a mutant and a WT tumor to achieve a percentage of mutant alleles near the limit of detection (5% mutation). The percentage of mutant alleles was determined by 454 sequencing.

Clinical Comparator Study Comparison of cobas 4800 BRAF V600 Mutation Test and Sanger Sequencing Results

The primary objective of this study was to compare the analytic performance of the cobas test with 2× bidirectional direct sequencing (Sanger) for the detection of the predominant V600E mutation. As indicated in “Materials and Methods,” specimens from 477 sequential eligible patients from the phase II and phase III trials were subjected to retrospective Sanger sequencing. Of the 477 eligible patients, 473 cases (99.2%) had a valid initial cobas BRAF test result; all 4 cases (0.8%) with an initially invalid cobas result yielded a valid result upon retesting. When DNA samples from the 477 eligible patients were subjected to Sanger sequencing, 44 samples (9.2%) gave invalid test results despite retesting, leaving 433 evaluable specimens for the comparison of the cobas assay and Sanger test results. The proportion of phase II enrollees with evaluable specimens (80%) was similar to that of phase III enrollees (79%). All 44 specimens with invalid Sanger sequencing results were subjected to 454 sequencing, which identified 28 specimens with V600E mutations, 4 with V600K mutations, 3 with V600R mutations, 1 with V600M, and 8 with no codon 600 mutation.

For the 433 evaluable specimens, the PPA between the methods was 96.4%; NPA was 80.0%; and OPA was 88.5% (Table 4). For this analysis, mutation positive was defined as the presence of the predominant BRAF mutation type V600E (c.1799 T>A), as identified by the specified sequencing method; all other valid non-V600E results were considered mutation negative. Test site-specific and macrodissection-specific PPAs and NPAs were comparable to the overall results (not shown).

Table 4. 

Comparison of the BRAF V600 Mutation Test With Sanger Sequencing for Detection of V600E Mutations

Comparison of the BRAF V600 Mutation Test With Sanger Sequencing for Detection of V600E Mutations
Comparison of the BRAF V600 Mutation Test With Sanger Sequencing for Detection of V600E Mutations

Assessment of Discrepant Samples by 454 Sequence Analysis.—In total, 50 specimens had discrepant cobas and Sanger sequencing results. All 50 specimens with discordant results (42 cobas mutation detected/Sanger sequence V600E mutation not detected, and 8 cobas mutation not detected/Sanger sequence mutation not detected) were subjected to 454 sequencing, as well as a random sampling of specimens that gave concordant results.

Among the 42 specimens that were cobas mutation detected and Sanger V600E mutation not detected, Sanger sequence detected a V600K mutation in 26 specimens, a V600D mutation in 1 case, and WT results in 15. Among these 42 discordant specimens, 454 sequencing detected a V600E mutation in 17 cases, including all 15 specimens that were WT by Sanger and 2 specimens that were V600K by Sanger. It is noteworthy that the percentage of mutant alleles, as determined by 454 sequencing in the 15 specimens that were WT by Sanger, ranged from 2.8% to 38.8%, with a median of 8.0%; 14 of 15 samples had a percentage of mutant alleles of less than 25%. Twenty-four of the 25 remaining discordant samples were positive for the V600K mutation by both Sanger and 454 sequencing. The 1 sample that had a V600D mutation identified by Sanger sequencing was WT by 454 sequencing. Thus, all of the 257 samples that were mutation detected by cobas testing had a codon 600 mutation confirmed by either Sanger or 454 or both.

Among the 8 cobas mutation-not-detected/Sanger mutation-detected discordant samples, the results of 454 sequencing revealed WT BRAF sequences in 2 specimens, the V600K variant in 2 specimens, “V600E2” in 1 specimen, and V600E in 3 specimens. Two of the 3 latter specimens had percentage of mutant alleles by 454 sequencing at the 5% limit of detection of the cobas test.

The 454 assay result agreement for specimens that were mutation-detected concordant with cobas and Sanger sequence analysis was 100% (39 of 39) and for mutation-not-detected concordant samples, 94.5% (52 of 55). For 2 of 3 cases in which a V600E mutation was detected by 454 sequencing, the percentage of mutant alleles was at or below the limit of detection of the cobas test.

Distribution of Codon 600 Mutation Variants in the BRAF Mutation Population.—By the Sanger and 454 sequence analyses conducted on our cohort of 477 eligible cases, V600E accounted for 81.7% of all codon 600 mutations, with V600K the second most frequent mutation at 12.4%. The remaining codon 600 mutation variants were uncommon at 3.1% or less (Table 5). Of the 40 specimens that were identified as having V600K mutations by Sanger sequencing, the cobas test result was positive in 28 (70%).

Table 5. 

Distribution of Codon 600 Variants in the BRAF-Mutant Population as Determined by Sanger and 454 Sequencinga

Distribution of Codon 600 Variants in the BRAF-Mutant Population as Determined by Sanger and 454 Sequencinga
Distribution of Codon 600 Variants in the BRAF-Mutant Population as Determined by Sanger and 454 Sequencinga

Overall, 92 of 94 runs (97.9%) were considered valid on the basis of the assay quality control criteria (1 run was invalid owing to an invalid control; 1 run was invalid owing to operator error). Of 1442 samples tested, 2 specimens (0.14%) gave invalid results (1 V600 E [5% mutation], 1 pigmented WT [75% tumor content]). Of a total of 1440 valid specimen results, all WT specimens, regardless of pigmentation, gave “mutation not detected” results (360 specimen results). For the valid mutant-positive specimens, “mutation detected” results were obtained in 1062 and mutation-not-detected results were obtained in 18. The only V600E-mutant samples that gave mutation-not-detected results were specimens below the assay limit of detection of 5%. In 180 samples with 4% mutation, mutation-detected results were obtained in 162 samples (90%). Data are shown in Table 6.

Table 6. 

Reproducibility Across Reagent Lots, Operators, and Testing Dates

Reproducibility Across Reagent Lots, Operators, and Testing Dates
Reproducibility Across Reagent Lots, Operators, and Testing Dates

In the clinical utility study, the cobas 4800 BRAF V600 Mutation Test showed a high PPA with bidirectional Sanger sequencing (96.4%) for the detection of V600E (1799 T>A) mutations. The lower NPA (80.0%) was primarily due to the greater sensitivity of the cobas test for the detection of V600E mutations and cross-reactivity of the test with non-V600E mutations. Indeed, in every case for which the cobas test detected a mutation, a codon 600 mutation was confirmed by Sanger and/or 454 sequencing. The cobas test, while not specifically designed for the detection of non-V600E mutations or being able to discriminate between codon 600 mutations, demonstrated substantial cross-reactivity with some non-V600E mutations, and thus detected 70% of the V600K mutations and the only V600D mutation in this cohort. It should be noted that preclinical studies demonstrated that V600K- and V600D-mutant melanoma cell lines are also sensitive to vemurafenib,4 and clinical responses were observed in some patients with V600K-mutant melanomas in both the phase II and phase III trials of vemurafenib.3,15 However, because of the small number of such cases in these 2 trials, further clinical studies are needed to establish the efficacy of vemurafenib in melanomas with non-V600E mutations.

We noted a test failure/invalid rate of 9.2% for Sanger sequencing in this study. There are few published data on the test failure rate for Sanger sequencing in the routine analysis of archival tissue samples for mutation analysis, but 2 other studies16,17 have reported similar invalid rates for Sanger sequencing, suggesting that this methodology has limitations when applied to DNA from FFPE materials. In addition, in a separate external study comparing the cobas BRAF test to Sanger sequencing, performed at a different clinical laboratory, an invalid rate of 6.9% was observed for Sanger sequencing.18 Lastly, it should be noted that, in this study, every specimen that gave invalid Sanger sequencing results gave valid results by both cobas and 454 testing.

Sanger sequencing also failed to detect a number of V600E mutations that were detected by the cobas test and confirmed by 454 sequencing; such cases would have been excluded from the vemurafenib clinical trial if Sanger sequencing had been used to screen patients for enrollment. Not surprisingly, a large percentage of these cases had low percentage of mutant alleles, confirming the results of a number of prior studies showing poor sensitivity of Sanger in specimens with less than 20% to 25% mutant alleles.6,19 Efforts to secure specimens with high tumor content and to enrich for tumor by macrodissection or microdissection may enhance the accuracy of Sanger sequencing. However, in clinical practice, specimens with high tumor content may not always be available and macrodissection may not be possible in every case.

Furthermore, the pathological estimation of tumor content is an inherently imprecise exercise, as shown in the recent study by Bellon et al,20 in which the authors observed widely divergent estimates of tumor content by a panel of expert pathologists. In the recent study by Halait et al,12 we found no correlation between the estimated tumor content of a specimen and the percentage of mutant alleles as determined by 454 sequencing. Thus, current methods of estimating tumor content and enriching for tumor by various macrodissection and microdissection techniques do not ensure that a given specimen will have a sufficient level of mutant alleles to be reliably detected by Sanger sequencing. It is important to stress that our test procedures mandated a macrodissection step for all specimens with estimated tumor content less than 50% before DNA isolation, for both cobas testing and Sanger sequencing. Yet, despite that enrichment step, Sanger sequencing still failed to detect BRAF V600E mutations in several specimens.

Conversely, we noted 2 cases in which a V600E mutation identified by Sanger sequencing was not confirmed by either the cobas test or 454 sequencing, suggesting that Sanger sequencing may sometimes yield a falsely positive mutation result. Such false-positive results could have significant clinical impact, since preclinical data indicate that BRAF inhibitors such as vemurafenib may paradoxically stimulate the growth of BRAFWT melanomas bearing upstream pathway activation.21 

In the clinical reproducibility study, the assay detected the mutation in all of the mutant samples except for the samples below the limit of detection. For the specimens with 4% mutant alleles, the V600E mutation was detected in 90% (162 of 180) of cases. The assay also showed an invalid rate of 0.14%, and no false-positive calls in this study. Only 1 invalid test result was observed for the 2 pigmented samples, which is noteworthy, since melanin has been previously reported to interfere with DNA polymerases,22 thus potentially leading to invalid test results.

Reproducibility and robustness of a mutation assay have significant clinical ramifications in terms of assuring physicians and their patients that the test result is not dependent on where the laboratory testing was done. Recent quality assurance studies performed to assess the ability of different clinical laboratories to ascertain the mutation status of a standard panel of tumors have shown that a significant proportion of laboratories do not correctly identify the mutation status of 100% of the panel members, even when they are using the same or similar testing methodologies.20,23 For assays that involve mutation analysis from tumor samples, important factors contributing to the assay performance include the reagents and methodology, the experience of the laboratory, and the appropriate involvement of the pathologist. Furthermore, invalid test results can cause significant delays in determining patients' BRAF mutation status and in implementing effective therapy for their disease.

In a recent draft guidance issued by the US Food and Drug Administration,24 the agency has stated that in vitro diagnostic companion assays should be developed in parallel with novel therapeutic products, such as targeted anticancer drugs, to ensure the safe and effective use of those therapeutic agents. The codevelopment of this assay in parallel with the clinical drug development of vemurafenib is the first prospective example of this process. A major strength of this approach, wherein the analytic performance of the assay is established by using specimens and data from the clinical development program of the drug, is that the clinical efficacy and safety of the drug have been demonstrated in patients selected for treatment by this assay. Based on the study presented here, the cobas assay has been shown to be a robust, accurate companion diagnostic assay for selecting patients for treatment with the targeted anticancer therapy vemurafenib.

We acknowledge Felice Shieh, PhD, and Suzanne Cheng, PhD, for their contributions to manuscript preparation, and Matt Lear, BA, for his assistance with specimen acquisition for the clinical reproducibility study.

1.
Garbe
C
,
Eigentler
TK
,
Keilholz
U
,
Hauschild
A
,
Kirkwood
JM
.
Systematic review of medical treatment in melanoma: current status and future prospects
.
Oncologist
.
2010
;
16
(
1
):
5
24
.
2.
Flaherty
KT
,
Puzanov
I
,
Kim
KB
, et al.
Inhibition of mutated, activated BRAF in metastatic melanoma
.
N Engl J Med
.
2010
;
363
(
9
):
809
819
.
3.
Chapman
PB
,
Hauschild
A
,
Robert
C
, et al.
Improved survival with vemurafenib in melanoma with BRAF V600E mutation
.
N Engl J Med
.
2011
;
364
(
26
):
2507
2516
.
4.
Yang
H
,
Higgins
B
,
Kolinsky
K
, et al.
RG7204 (PLX4032), a selective BRAF V600E inhibitor, displays potent antitumor activity in preclinical melanoma models
.
Cancer Res
.
2010
;
70
(
13
):
5518
5527
.
5.
Sanger
F
,
Nicklen
S
,
Coulson
AR
.
DNA sequencing with chain-terminating inhibitors
.
Proc Natl Acad Sci U S A
.
1977
;
74
(
12
):
5463
5467
.
6.
Tan
YH
,
Liu
Y
,
Eu
KW
, et al.
Detection of BRAF V600E mutation by pyrosequencing
.
Pathology
.
2008
;
40
(
3
):
295
298
.
7.
Ribas
A
,
Flaherty
KT
.
BRAF targeted therapy changes the treatment paradigm in melanoma
.
Nat Rev Clin Oncol
.
2011
;
8
(
7
):
426
433
.
8.
Pollock
PM
,
Harper
UL
,
Hansen
KS
, et al.
High frequency of BRAF mutations in nevi
.
Nat Genet
.
2003
;
33
(
1
):
19
20
.
9.
Viros
A
,
Fridlyand
J
,
Bauer
J
, et al.
Improving melanoma classification by integrating genetic and morphologic features
.
PLoS Med
.
2008
;
5
(
6
):
0941
0952
.
10.
Long
GV
,
Menzies
AM
,
Nagrial
AM
, et al.
Prognostic and clinicopathologic associations of oncogenic BRAF in metastatic melanoma
.
J Clin Oncol
.
2011
;
29
(
10
):
1239
1246
.
11.
Wellcome Trust Sanger Institute
.
Catalogue of somatic mutations in cancer (COSMIC) database
.
12.
Halait
H
,
De Martin
K
,
Shah
S
, et al.
Analytical performance of a real-time 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
.
In press.
13.
Thomas
RK
,
Nickerson
E
,
Simons
JF
, et al.
Sensitive mutation detection in heterogeneous cancer specimens by massively parallel picoliter reactor sequencing
.
Nat Med
.
2006
;
12
(
7
):
852
855
.
14.
Wilson
EB
.
Probable inference, the law of succession and statistical inference
.
J Am Stat Assoc
.
1927
;
22
:
209
212
.
15.
Sosman
JA
,
Kim
KB
,
Schuchter
L
, et al.
Long-term survival in vemurafenib-treated BRAF V600-mutant advanced melanoma
.
New Eng J Med
.
2012
.
In press.
16.
Gallegos Ruiz
MI
,
Floor
K
,
Rijmen
F
,
Grunberg
K
,
Rodriguez
JA
,
Giaccone
G
.
EGFR and K-ras mutation analysis in non-small cell lung cancer: comparison of paraffin embedded versus frozen specimens
.
Cell Oncol
.
2007
;
29
(
3
):
257
264
.
17.
Ogino
S
,
Kawasaki
T
,
Brahmandam
M
, et al.
Sensitive sequencing method for KRAS mutation detection by pyrosequencing
.
J Mol Diagn
.
2005
;
7
(
3
):
413
421
.
18.
Lopez-Rios
F
,
Angulo
B
,
Gomez
B
, et al.
Comparison of molecular testing methods for the detection of BRAF V600E mutations in formalin-fixed paraffin-embedded tissue (FFPET) specimens of malignant melanoma
.
Virchows Arch
.
2011
;
459
(
suppl 1
):
S323
.
19.
Kotoula
V
,
Charalambous
E
,
Biesmans
B
, et al.
Targeted KRAS mutation assessment on patient tumor histologic material in real time diagnostics
.
PLoS One
.
2009
;
4
(
11
):
1
13
.
20.
Bellon
E
,
Ligtenberg
MJ
,
Tejpar
S
, et al.
External quality assessment for KRAS testing is needed: setup of a European program and report of the first joined regional quality assessment rounds
.
Oncologist
.
2011
;
16
(
4
):
467
478
.
21.
Hatzivassiliou
G
,
Song
K
,
Yen
I
, et al.
RAF inhibitors prime wild-type RAF to activate the MAPK pathway and enhance growth
.
Nature
.
2010
;
464
(
7287
):
431
435
.
22.
Eckhart
L
,
Bach
J
,
Ban
J
,
Tschachler
E
.
Melanin binds reversibly to thermostable DNA polymerase and inhibits its activity
.
Biochem Biophys Res Commun
.
2000
;
271
(
3
):
726
730
.
23.
Beau-Faller
M
,
Degeorges
A
,
Rolland
E
, et al.
Cross-validation study for epidermal growth factor receptor and KRAS mutation detection in 74 blinded non-small cell lung carcinoma samples: a total of 5550 exons sequenced by 15 molecular French laboratories (Evaluation of the EGFR Mutation Status for the Administration of EGFR-TKIs in Non-Small Lung Carcinoma [ERMETIC] Project-part 1)
.
J Thorac Oncol
.
2011
;
6
:
1006
1015
.
24.
US Food and Drug Adminstration
.
Draft guidance for industry and Food and Drug Administration staff: in vitro companion diagnostic devices.
July 14, 2011
.

Author notes

From Oncology and Genetics, Laboratory Corporation of America, Research Triangle Park, North Carolina (Dr Anderson); GE Healthcare/Clarient, Inc, Aliso Viejo, California (Dr Bloom); TMD - A Quintiles Central Laboratory, Westmont, Illinois (Dr Vallera); Targos Molecular Pathology, Kassel, Germany (Dr Rueschoff); Molecular Pathology Level 1, Peter MacCallum Cancer Centre, East Melbourne, Australia (Dr Meldrum); Clinical Operations (Mr Schilling and Ms Kovach), Biometrics Department (Dr Lee and Ms Ochoa), Research Department (Ms Langland), Development Department (Mr Halait), and Medical & Scientific Affairs (Drs Lawrence and Dugan), Roche Molecular Systems, Inc, Pleasanton, California.

Clarient, Inc, was contracted by Roche to complete BRAF testing.

This study was sponsored by Roche Molecular Systems, Pleasanton, California. Dr Meldrum received a fee-for-service provisions of testing associated with development of companion diagnostics and received honorarium for a presentation at a conference and a position on Roche's advisory boards.

Dr Ju Ruey-Jiuan Lee is a contract employee at Roche. Drs Dugan and Lawrence, Messrs Schilling and Halait, and Misses Kovach, Ochoa, and Langland are employees of Roche. All other authors have no relevant financial interest in the products or companies described in this article. Dr Dugan is now located at MCDXI Medical Diagnostics International, Santa Rosa Valley, California. Dr Lawrence is now located at Genomic Health, Inc, Redwood City, California.

Results of this study were presented in part at the European Multidisciplinary Cancer Congress, European Society of Medical Oncology, Stockholm, Sweden, September 26, 2011.