Assessment of BRAF V600E Status in Colorectal Carcinoma ... · conclude that VE1 IHC produces suboptimal results in colo-rectal carcinoma and should not be used to guide patient management

Companion Diagnostics and Cancer Biomarkers

Assessment of BRAF V600E Status in ColorectalCarcinoma: Tissue-Specific Discordancesbetween Immunohistochemistry and SequencingJeannelyn S. Estrella1, Michael T. Tetzlaff1,2, Roland L. Bassett Jr3, Keyur P. Patel4,Michelle D.Williams1, Jonathan L. Curry1, Asif Rashid1, Stanley R. Hamilton1, andRussell R. Broaddus1

Abstract

Although sequencing provides the gold standard for identi-fying colorectal carcinoma with BRAF V600E mutation, immu-nohistochemistry (IHC) with the recently developed mousemonoclonal antibody VE1 for BRAF V600E protein has shownpromise as a more widely available and rapid method. How-ever, we identified anecdotal discordance between VE1 IHC andsequencing results and therefore analyzed VE1 staining by twodifferent IHC methods (Leica Bond and Ventana BenchMark)in whole tissue sections from 480 colorectal carcinomas (323BRAF wild-type, 142 BRAF V600E mutation, and 15 BRAF non-V600E mutation). We also compared the results with melano-mas and papillary thyroid carcinomas (PTC). With the Bondmethod, among 142 BRAF V600E-mutated colorectal carcino-mas, 77 (54%) had diffuse VE1 staining and 48 (33%) hadheterogeneous staining, but 17 (12%) were negative. Among

323 BRAF wild-type colorectal carcinomas, 196 (61%) werenegative, but 127 (39%) had staining, including 7 with diffusestaining. When positivity was defined as staining in �20% oftumor cells, VE1 IHC had sensitivity of 75% and specificity of93% for BRAF V600E mutation. With the Ventana method,among 57 BRAF V600E-mutated colorectal carcinomas, 36(63%) had diffuse VE1 staining, whereas 6 (11%) had no orweak (<20% of tumor cells) staining. Among 33 BRAF wild-type colorectal carcinomas, 16 (48%) had no or weak staining,whereas 15 (45%) had heterogeneous staining. In contrast withcolorectal carcinoma, Bond and Ventana VE1 IHC in melanomaand PTC were highly concordant with sequencing results. Weconclude that VE1 IHC produces suboptimal results in colo-rectal carcinoma and should not be used to guide patientmanagement. Mol Cancer Ther; 14(12); 2887–95. �2015 AACR.

IntroductionThe BRAF proto-oncogene encodes a serine/threonine kinase

belonging to the raf/mil family. The BRAF protein resides at theapex of, and thus provides a critical regulatory function for, theMAPK signaling cascade. The most frequent somatic alteration inthe BRAF gene is a point mutation in codon 600 that replacesvaline with glutamate (NM_004333.4[BRAF]:c.1799T>Ap.V600E, referred to as BRAF V600E hereon) and results inRAS-independent activity of the kinase domain, in turn causingconstitutive activation of the MAPK pathway (1). Using in vivo

mouse models, Rad and colleagues demonstrated that the BRAFV600E-associated pathway of intestinal tumorigenesis occursthrough a hyperplasia/adenoma/carcinoma sequence with sub-sequent acquisition of high levels of microsatellite instability,then activation of the Wnt pathway and intensification of MAPKsignaling, and, finally, late-stage inactivation of p16 and p53 (2).

Colorectal carcinomas that harbor the BRAF V600E mutationare a distinct subset of tumors. They are frequently associatedwithpoor differentiation, mucinous histology, and advanced tumor–node–metastasis stage (3). In patients with microsatellite-stablecolorectal carcinoma (4, 5) and those with advanced colorectalcarcinoma (6), BRAF V600E mutation confers worse survivalcompared with their wild-type counterpart. Furthermore, Lynchsyndrome (hereditary nonpolyposis colorectal cancer syndrome)is virtually excluded when a colorectal carcinoma that exhibitsloss of the MLH1 and PMS2 proteins by immunohistochemistry(IHC) also harbors BRAF V600E mutation (7).

BRAF V600E mutation confers important predictive value inthe treatment of patients with colorectal carcinomas. Some clin-ical studies, supported by in vitro results (8), have reported adetrimental effect of BRAF V600E mutation in patients withcolorectal carcinoma treated with the anti-EGF receptor therapycetuximab or panitumunab (9–12), although data are conflicting.In a subsequent analysis of pooled data from the randomizedphase III trial, Cetuximab Combined with Irinotecan in First-LineTherapy for Metastatic Colorectal Cancer (CRYSTAL), andthe randomized phase II trial, Oxaliplatin and Cetuximab in

1DivisionofPathologyandLaboratoryMedicine,DepartmentofPathol-ogy, The University of Texas MD Anderson Cancer Center, Houston,Texas. 2Department of Translational Molecular Pathology, The Univer-sityofTexasMDAndersonCancerCenter,Houston,Texas. 3Departmentof Biostatistics and Applied Mathematics, The University of Texas MDAnderson Cancer Center, Houston, Texas. 4Division of Pathology andLaboratoryMedicine,DepartmentofHematopathology,TheUniversityof Texas MD Anderson Cancer Center, Houston, Texas.

Note: Supplementary data for this article are available at Molecular CancerTherapeutics Online (http://mct.aacrjournals.org/).

Corresponding Author: Jeannelyn S. Estrella, The University of Texas MDAnderson Cancer Center, 1515 Holcombe Boulevard, Unit 0085, Houston, TX77030. Phone: 713-792-5664; Fax: 713-792-4341; E-mail:[email protected]

doi: 10.1158/1535-7163.MCT-15-0615

�2015 American Association for Cancer Research.

MolecularCancerTherapeutics

www.aacrjournals.org 2887

on August 31, 2020. © 2015 American Association for Cancer Research. mct.aacrjournals.org Downloaded from

Published OnlineFirst October 5, 2015; DOI: 10.1158/1535-7163.MCT-15-0615

http://mct.aacrjournals.org/

First-Line Treatment of Metastatic Colorectal Cancer (OPUS), nosignificant differences in outcome (overall survival, progression-free survival, and best overall response rate) were found betweenpatients with colorectal carcinoma with BRAF mutations andBRAF wild-type (13). In vitro and in vivo xenograft models haveshown that colorectal carcinoma that harbor BRAF V600Emay besensitive to proteasome inhibitors (14). Although colorectalcarcinomas have very limited response to BRAF V600E inhibitormonotherapy (15, 16), combinatorial strategies are currentlyunder investigation to improve response, and outcome for thesepatients (17, 18).

Given that BRAF V600E has these important hereditary, prog-nostic, and therapeutic implications, there is a critical need toensure accurate identification of patients whose colorectal carci-noma has thismutation. PCR-based sequencing assays that detectthe BRAF V600E mutation are considered to be the gold standardfor assessing the mutational status of this gene for patient man-agement decisions. More recently, amousemonoclonal antibodyfor the BRAF V600E protein (clone VE1, Spring Bioscience) hasbecome commercially available for use in IHC. This method hasthe advantages of being relatively fast, inexpensive, and widelyavailable for use in routine formalin-fixed, paraffin-embeddedtissue. In addition, IHC overcomes numerous challenges associ-atedwith sequence analysis that include limited availability of thetechnology, requirement for larger tumor sample, and effects ofdilution with non-neoplastic tissue.

Previous studies on the suitability of VE1 IHC for the detectionof BRAF V600E mutation in colorectal carcinoma found sensi-tivity ranging from 59% to 100% and specificity ranging from51% to 100% with use of different IHC techniques, including awide spectrum of antibody conditions, and different sequencingtechniques as the comparator (19–31). In the largest such series todate, investigators usedwhole tissue sections to analyze 113BRAFwild-type and 52 BRAF V600E-mutated colorectal carcinomas inthe validation cohort (21) and reported sensitivity of 96% andspecificity of 99%. Despite these reportedly high concordancesbetween results of VE1 IHC and sequence analysis, in our clinicalpractice we had noted discordances between VE1 IHC andsequencing results when both tests were performed on colorectalcarcinoma samples. In contrast, concordance was high betweenresults of the two tests in melanoma and papillary thyroidcarcinoma samples. Because VE1 IHC has been proposed for useas a surrogate marker for BRAF V600E mutation—even replacingmolecular studies altogether—we believed that more rigorousvalidationofVE1 IHCwith a large cohort of patientswasnecessaryto establish analytical and clinical validity. We therefore analyzedwhole tissue sections from480 colorectal carcinoma cases, includ-ing 323 BRAF wild-type tumors, 142 tumors with BRAF V600Emutation, and 15 tumors with BRAFmutation other than V600E,to compare two different VE1 IHC methods and three differentBRAF sequencing methods. We also compared the results incolorectal carcinoma with those in melanomas and papillarythyroid carcinomas analyzed in our laboratory using the samesequencing and IHC techniques.

Materials and MethodsStudy population

The pathology files of The University of Texas MD AndersonCancer Center were searched for all resection specimens of colo-rectal carcinoma (primary tumors and metastases) from 2008

through 2013 for which BRAF mutation analysis had beenperformed. As previously published studies have reported sen-sitivity and specificity of up to 100% for VE1 IHC, we planned toanalyze a sample size of at least 300 cases each of BRAF wild-type colorectal carcinoma and BRAF V600E colorectal carcino-ma so that 100% concordance between sequencing and IHCwould give us an upper 95% confidence boundary of less than1% discordance. We therefore randomly selected for inclusionin this study a total of 323 colorectal carcinoma cases to includeequal numbers of cases each year from over 1,500 colorectalcarcinoma cases that had no mutation detected in the BRAFgene (wild-type) by sequencing and that had available materialfor IHC testing. Our search of the pathology files revealed only142 colorectal carcinoma cases with BRAF V600E mutation and15 additional cases with BRAF mutation other than V600E onsequencing that had residual material available to perform IHC;all of these cases were included in this study. Thus, a total of 480colorectal carcinomas were analyzed with VE1 IHC, as describedbelow.

For comparison, we also searched our pathology files to iden-tify cases of melanoma and papillary thyroid carcinoma that hadalready been analyzed for BRAF V600E mutation by bothsequence analysis and VE1 IHC in 2013 and 2014. For thesecases, we abstracted the previously recorded findings on sequenceanalysis and VE1 IHC from patient records.

The study was approved by the MD Anderson InstitutionalReview Board.

Sequence analysisDNA had been extracted from microdissected unstained, for-

malin-fixed, paraffin-embedded whole tissue sections, andsequencing had been performed in the College of AmericanPathologists (CAP)-accredited and Clinical Laboratory Improve-ments Amendments (CLIA)-certified Molecular Diagnostics Lab-oratory, as previously described (32–34), as a component ofroutine patient care or eligibility for integral-marker clinical trials.Several different sequencing assays had been employed during theyears covered by our study: (1) DNA pyrosequencing (PSQ96 HSSystem; Biotage AB; n ¼ 202; ref. 34); (2) sequenom matrix-assisted laser desorption/ionization-time of flightmass spectrom-etry (Sequenom MassARRAY; Sequenom; n ¼ 148; ref. 32); and(3) next-generation sequencing (NGS) with the Ion TorrentPersonal Genome Machine (Life Technologies; n ¼ 130; ref. 33).For all three sequencing techniques, a minimum cellularity of20% (i.e., tumor nuclei represent 20% of total nuclei in the testedsample) was required to avoid false negative results.

VE1 IHCIHC was performed for this study on formalin-fixed, paraffin-

embedded whole tissue sections of the 480 colorectal carcinomasin the CAP-accredited and CLIA-certified clinical IHC Laboratorywith theVE1mousemonoclonal antibody toBRAFV600Eprotein(clone VE1, 1:50; Spring Bioscience). For the Leica Bond method(n ¼ 480 colorectal carcinoma), parameters were as follows:antigen retrieval, Tris-EDTA buffer, pH 9.0 (20 minutes);ADV-060 antibody diluent (Spring Bioscience); and Leica Bondautomated systemwith Bond Polymer Refine Detection kit (LeicaBiosystems). For the Ventana BenchMark system subset (n ¼ 92colorectal carcinoma), parameterswere as follows: antigen retriev-al, ultra cell conditioning 1 (64 minutes), and Ventana Bench-Mark ultra-automated system and standard reagents provided by

Estrella et al.

Mol Cancer Ther; 14(12) December 2015 Molecular Cancer Therapeutics2888




Ventana (OptiView DAB IHC Detection Kit; Ventana MedicalSystems).

VE1 cytoplasmic staining was scored microscopically at thetime of the study by semiquantitative analysis by one pathologist(J.S. Estrella) according to the following scoring system: 0, neg-ative; 1, weak staining in <20% of tumor cells (Fig. 1A); 2,moderate to strong staining in <20% of tumor cells (Fig. 1B);3, weak staining in 20% to 70% of tumor cells; 4, moderate tostrong staining in 20% to 70% of tumor cells; 5, weak staining in>70% of tumor cells (Fig. 1C); and 6, moderate to strong stainingin >70% of tumor cells (Fig. 1D). Cases that had weakly positivestaining (scores 1, 3, and 5), cases that exhibited moderately tostrongly positive staining in <20% of tumor cells (score 2), andcases discordant with the sequencing results were reviewed by twoother pathologists (M.T. Tetzlaff and R.R. Broaddus) to evaluatethe initial interpretation. Difference in classification was resolvedby a consensus agreement. None of the pathologists (J.S. Estrella,M.T. Tetzlaff, and R.R. Broaddus) had access to the sequencingresults at the time of IHC scoring.

Statistical analysisSensitivity was defined as the ratio of true positives on VE1 IHC

to positives identified by BRAF sequencing, and specificity was

defined as the ratio of true negatives on VE1 IHC to negatives onsequencing, both expressed as percentage. Spearman rho (r) wascalculated to assess the nonparametric correlation between VE1IHC results and results of the different sequencing methods.

ResultsStudy population

The characteristics of patients included in the study are sum-marized in Table 1. The 15 BRAF mutations other than V600Eidentified by sequencing were as follows: D594G (6), K601E (2),G464V (1),G466V (1),G466E (1),G469R (1),D594N(1), L597R(1), and V600K (1).

Approximately equal numbers of BRAF wild-type cases hadbeen analyzed with one of the three sequencing techniques,because BRAF wild-type cases were randomly selected to includean equivalent number of cases from each year of the study. Incontrast, most (63%) of the tumors with a BRAF mutation,including both V600E and mutations other than V600E, hadbeen identified using pyrosequencing. The percentage of BRAFmutation by pyrosequencing was 49% (99/202); by Sequenom,28% (41/148); and by NGS, 13% (17/130). These data highlightan important limitation in that the relative frequency ofmutation

Figure 1.Examples of VE1 IHC in colorectalcarcinomas. A, weak cytoplasmicstaining in <20% of tumor cells(score ¼ 1). B, moderate to strongcytoplasmic staining in <20% of tumorcells (score ¼ 2). C, weak cytoplasmicstaining in >70% of tumor cells(score ¼ 5). D, moderate to strongcytoplasmic staining in >70% of tumorcells (score ¼ 6). A–D, magnification,�100.

Table 1. Colorectal carcinoma patient characteristics

BRAF wild-type BRAF V600E mutation BRAF non-V600E mutationCharacteristic (n ¼ 323) (n ¼ 142) (n ¼ 15)

Median age, years (range) 59 (14–91) 65 (27–89) 69 (30–81)Sex, female:male 150:173 77:65 9:6Sequencing method, no. (%)Pyrosequencing 103 (32) 91 (64) 8 (53)Sequenom 107 (33) 37 (26) 4 (27)Next-generation sequencing 113 (35) 14 (10) 3 (20)

BRAF Immunohistochemistry in Colorectal Carcinoma

www.aacrjournals.org Mol Cancer Ther; 14(12) December 2015 2889




may be dependent upon the methodology employed. Additionalstudies are warranted to determine the most appropriate goldstandard for comparison to IHC, although, these additionalstudies are beyond the scope of the current study. As a result ofthese findings, however, the Bond IHC method was evaluated inrelation to the individual sequencing methods, as well as to allthree sequencing methods combined.

VE1 IHC using the Bond method in all colorectal carcinomaResults of VE1 IHC using the Bond method are tabulated in

Supplementary Table S1. Among the 142 colorectal carcinomacases that had a BRAF V600E mutation by sequence analysis, 77cases (54%) had cytoplasmic staining in >70% of tumor cells(scores 5–6)byVE1 IHC, 29 cases (20%)had cytoplasmic stainingin 20% to 70% of tumor cells (scores 3–4), and 19 cases (13%)had cytoplasmic staining in <20% of tumor cells (score 1).However, 17 cases (12%) known to have a BRAF V600Emutationby sequencing did not stain with the VE1 antibody. None of thecolorectal carcinoma cases with BRAF V600Emutation had mod-erate to strong cytoplasmic staining in <20% of tumor cells (score2). Thus, the sensitivity of VE IHC was low, as only 54% of thecolorectal carcinoma cases with BRAF V600E mutation bysequence analysis had unambiguously positive VE1 IHC.

Among the 323 colorectal carcinoma caseswithwild-typeBRAFby sequence analysis, 196 cases (61%) did not stain with the VE1antibody. However, 104 cases (32%) exhibited cytoplasmic stain-ing in <20% of tumor cells (scores 1–2), and 23 cases (7%) hadcytoplasmic staining in�20%of tumor cells (scores 3–6), includ-ing 7 cases with diffuse staining. Thus, the specificity of VE1 IHCwas low, as 39% of BRAF wild-type colorectal carcinomas exhib-ited VE1 IHC staining.

Among the 15 colorectal carcinoma cases with BRAFmutationother than V600E by sequence analysis, 12 cases (80%) did notstain with the VE1 antibody, although 3 cases (20%) had cyto-plasmic staining in <20% of tumor cells.

The sensitivity and specificity using different definitions forpositive VE1 staining are summarized in Table 2. The definitionyielding the highest sensitivity and specificity was staining in�20% of tumor cells regardless of intensity, which we appliedin the subanalysis below.

Nonspecific staining was frequently seen in both BRAF wild-type and BRAF-mutated cases (Supplementary Fig. S1) andincluded the following patterns: weak to strong nuclear andcytoplasmic staining of histopathologically normal colonicmucosa and inflammatory cells; weak to strong granular cyto-plasmic staining of smooth muscle in muscularis mucosae, mus-cularis propria, and thick-walled blood vessels; strong staining ofluminal mucin; and strong staining of brush border of bronchialepithelium accompanying lung metastasis.

VE1 IHC using the Bond method in comparison withsequencing method

The sensitivity and specificity of VE1 IHC using the Bondmethod and applying the definition of positive staining ascytoplasmic staining in �20% of tumor cells are summarizedin Table 3 according to the BRAF sequencing technique that wasused. For all three sequencing techniques, the majority ofcolorectal carcinoma cases with BRAF V600E identified bysequencing was positive by VE1 IHC in �20% of tumor cells;specificities ranged from 89% to 95%. However, for eachsequencing technique, some cases identified as positive forBRAF V600E were negative by VE1 IHC; sensitivities rangedfrom 70% to 84%. The relatively low sensitivity of VE1 IHC todetect BRAF V600E has great potential clinical impact, as thisfinding indicates that some colorectal carcinoma cases withBRAF V600E mutation would be missed if only IHC were usedas the method of detection.

Among 36 cases with BRAF V600E mutation by sequenceanalysis but negative cytoplasmic staining (scores 0 and 1) onVE1 IHC (Supplementary Table S1), BRAFmutation status in the

Table 2. Sensitivity and specificity of VE1 IHC in colorectal carcinoma using the Bond method

Definition of positive VE1 IHC resulta

BRAF mutation status by sequence analysis VE1 IHC result Scores 1–6 Scores 3–6 Scores 2–6 Scores 2, 4, and 6

WT or non-V600E mutation Negative 208 315 299 309Positive 130 23 39 29

V600E mutation Negative 17 36 36 97Positive 125 106 106 45

Sensitivity 88% 75% 75% 32%Specificity 62% 93% 88% 91%

Abbreviation: WT, wild-type.a0, negative; 1, weak staining in <20% of tumor cells; 2, moderate to strong staining in <20% of tumor cells; 3, weak staining in 20% to 70% of tumor cells; 4, moderateto strong staining in 20% to 70% of tumor cells; 5, weak staining in >70% of tumor cells; and 6, moderate to strong staining in >70% of tumor cells.

Table 3. Sensitivity and specificity of VE1 IHC in colorectal carcinoma using the Bond method according to sequencing method used as the criterion standard

Pyrosequencing result Sequenom result Next-generation sequencing result

VE1 IHC resulta

BRAF WT ornon-V600Emutation

BRAF V600Emutation


BRAF V600Emutation


BRAF V600Emutation

Negative 99 27 106 6 110 3Positive 12 64 5 31 6 11

Sensitivity 70% 84% 79%Specificity 89% 95% 95%Spearman's rho 0.487 0.707 0.576

Abbreviation: WT, wild-type.aA positive VE1 IHC result was defined as staining in �20% of tumor cells.

Estrella et al.





majority of cases was determined by pyrosequencing (27 cases,75%). One possible explanation for the discordance was thatpyrosequencing may have erroneously detected the BRAF V600Emutation. Indeed, comparison of VE1 IHC results against pyr-osequencing results yielded the lowest correlation coefficient(r ¼ 0.487, Table 3) and the lowest sensitivity (70%) andspecificity (89%).We attempted to repeatBRAFmutation analysisby NGS to address this issue, but insufficient DNA remained inour laboratory. Among the 27 cases with BRAF V600E mutationby pyrosequencing, results of additional molecular analyses per-formed on the same tumor block were as follows: 21 of 22 wereKRAS wild-type in codons 12, 13, and 61, as expected becauseKRAS and BRAFmutations are usually mutually exclusive; 5 casesexhibited high levels of microsatellite instability by PCR andmethylation of the MLH-1 promoter region, again expectedbecause MSI-H secondary to somatic MLH-1 hypermethylationormutation is commonly associated with BRAFmutation; 2 casesexhibited methylation of theMLH-1 promoter region but did nothave microsatellite instability analyzed by PCR; 2 cases exhibitedhigh levels of microsatellite instability by PCR but did not haveMLH-1 promoter methylation analyzed. Only two cases did nothave additional molecular analysis.

We reviewed the sequencing result of 18 caseswithBRAFV600Emutation but negative (score 0) VE1 IHC by the Bond method todetermine the mutation signal of these cases. Among 18 cases,17 (94%)were definitive for theBRAFV600Emutationwith alleleburden ranging from 10% to 60% (median ¼ 30%). In one case,there is a low level mutation (allele burden ¼ 6%); however,repeat testing yielded the same result.

Among the 13 cases that were BRAF wild-type by sequenceanalysis but had moderate to strong cytoplasmic staining in�20% of tumor cells (scores 4 and 6) on VE1 IHC (Supple-mentary Table S1), BRAF mutation status had been determinedby pyrosequencing in five cases (38%, 5/13) and by Sequenomand NGS in four cases each (31%, 4/13). One possible expla-nation for the discordance was that the tissue was amenable toIHC studies, whereas DNA isolated from the tumor was sub-optimal for molecular testing. However, in 9 of the 13 cases,molecular testing performed on the same tumor block gener-ated informative results for other genes that were tested; in 4cases other sequence analyses yielded a negative result. Addi-tional molecular analyses for all the 13 cases are tabulated inSupplementary Table S3.

VE1 IHC using the Ventana method in colorectal carcinomaBecause the Ventana platform was the method recom-

mended by the VE1 antibody vendor and the majorityof previously published reports used the Ventana method(22–24, 26, 27, 29, 30), we also performed VE1 IHC on selectedcolorectal carcinoma cases, which by the Bond method wereconcordant or discordant with the sequencing result. A total of92 cases was re-evaluated in this manner: 57 cases harboringthe BRAF V600E mutation, 33 BRAFwild-type cases, and 2 casesharboring BRAF mutations other than V600E. Results are sum-marized in Supplementary Table S2. With the Ventana method,among the 57 colorectal carcinoma cases that harbored BRAFV600E mutation by sequencing, 36 cases (63%) exhibiteddiffuse cytoplasmic staining, whereas 6 cases (11%) had nostaining or weak staining in <20% of tumor cells. Among theBRAF wild-type colorectal carcinomas, 16 (48%) had no stain-ing or weak staining in <20% of tumor cells, whereas 15 (45%)

had positive cytoplasmic staining, including 5 cases with mod-erate to strong cytoplasmic staining in >20% of tumor cells.

To compare our Bond and Ventana VE1 IHC findings in these92 cases, we plotted the Bond score (x-axis) versus the Ventanascore (y-axis) for each individual case (Fig. 2). As depicted inthe figure, there is a random distribution of individual cases,both BRAF V600E and BRAF wild-type/non-V600E, across VE1IHC scores using the two different IHC staining platforms.Neither BRAF V600E nor wild-type/non-V600E cases demon-strated concordance with respect to their BRAF status by VE1IHC.

VE1 IHC with Bond method in melanoma sequenced by NGSVE1 IHC with the Bond method had been performed at our

laboratory on 144melanomas with known BRAFmutation statusby NGS (49 with BRAF V600E, 74 BRAF wild-type, and 21 withBRAF non-V600Emutations). Among the 49melanomas harbor-ing BRAF V600E mutation, 48 stained with the VE1 antibody,representing 98% sensitivity. Among the 95 BRAF wild-type andnon-V600Emelanomas, 91 were negative with the VE1 antibody,representing 96% specificity [data presented in part in Tetzlaff andcolleagues (35)].

VE1 IHCwithVentanamethod inmelanoma sequenced byNGSVE1 IHC had been performed with the Ventana method in our

laboratory on 90 melanomas with known BRAF mutation statusby NGS (25 with BRAF V600E, 48 BRAF wild-type, and 17 withBRAF non-V600E mutations). Among the 25 melanomas withBRAF V600E mutation, 23 were positive for the VE1 antibody,representing sensitivity of 92%. Among the 65 BRAF wild-typeand non-V600E melanomas, 63 were negative for the VE1 anti-body, representing specificity of 97%.

Figure 2.Scatter plot of VE1 IHC scoring results with the Bond method (x-axis) andVentana method (y-axis) for 92 colorectal carcinoma cases. Blue representsBRAF V600E, and red represents BRAF wild-type or BRAF mutations otherthan V600Emutation. There is a randomdistribution of individual cases, bothBRAF V600E and BRAF wild-type/non-V600E, across VE1 IHC scores usingthe two different IHC staining platforms.






VE1 IHC in papillary thyroid carcinomaOur sample size of papillary thyroid carcinoma is limited. Only

36 caseswere previously tested by SequenomorNGS.Of these, 16harbored BRAF V600E mutation, and 20 were BRAF wild-type.VE1 IHC by the Bondmethod was performed in 17 cases (9 BRAFV600E and 8 wild-type). All BRAF V600E cases were positive byVE1 IHC, and all BRAFwild-type cases were negative by VE1 IHC,yielding sensitivity and specificity of 100%. VE1 IHC by theVentana method was performed in 19 cases (7 BRAF V600E and12 wild-type). All BRAF V600E cases were positive by VE1 IHC,and 11 of 12BRAFwild-type cases were negative by VE1 IHC, for asensitivity of 100% and a specificity of 92%. The discordant casewas a metastatic papillary thyroid carcinoma involving a lymphnode that showed strong VE1 expression by IHC but nomutationby NGS. Given the small amount of tumor present in the tissuetested, it is likely that tumor DNA was diluted with lymphocyteDNA during DNA extraction, resulting in a false-negative findingon sequencing.

DiscussionIn our study, we examined VE1 IHC findings as a surrogate for

BRAF V600E mutation status in colorectal carcinoma and com-pared the results with melanoma and papillary thyroid carcino-ma. We believe this is the largest study of its type to date incolorectal carcinoma. In colorectal carcinoma, VE1 IHC using theBond method and a definition of positivity of cytoplasmic stain-ing in �20% of tumor cells yielded a sensitivity of 75% and aspecificity of 93%. When selected cases were re-evaluated usingthe Ventana method, a disturbing proportion of cases againshowed IHC-sequencing discordance: 11% of colorectal carcino-mas with BRAF V600E were negative with the VE1 antibody, and18% of colorectal carcinomas with BRAF wild-type exhibiteddiffuse cytoplasmic staining. Our findings show that in colorectalcarcinoma, VE1 IHC is not a reliable method for determining theBRAF V600E mutational status. In contrast, our results in mela-noma and papillary thyroid carcinoma are in keeping with pre-viously published data showing excellent concordance betweenmolecular analysis and VE1 IHC (36–41).

Other published studies examining the concordance betweenIHC and sequencing for assessing BRAF V600E in colorectal

carcinoma are summarized in Table 4. Our results are similar tothose of three prior studies (19, 25, 30). Adackapara and collea-gues evaluated 52 colorectal carcinoma whole tissue sections andreported a sensitivity of 71% and a specificity of 74% using amanual staining technique (19). Lasota and colleagues used theLeica Bond platform to evaluate 113 colorectal carcinoma wholetissue sections and reported a sensitivity of 89% and specificity of51% when weak staining was considered positive, and a sensi-tivity of 85%and specificity of 68%whenonlymoderate to strongstaining was considered positive. Finally, Loes and colleaguesevaluated 99 colorectal carcinoma tissue microarray cores usingthe Ventana BenchMark XT platform and reported a sensitivity of59% and a specificity of 84%. In contrast, multiple other studieshave reported sensitivities of 96% to 100% and specificities of94% to 100% in colorectal carcinoma using the Leica Bond andVentana platforms (20–24, 29).

Asmultiple studies have highlighted, a proper scoring system isnecessary to reduce false-positive and false-negative cases; how-ever, VE1 IHC scoring is controversial. The studies by Lasota andcolleagues (25) andKuanand colleagues (29) suggested thatweakstaining (even diffuse) was an unreliable result, although thestudy by Bledsoe and colleagues (21) concluded that diffuse (ornear-uniform) weak positive staining should be regarded as true-positive staining. In our study, we found that defining positivestaining as cytoplasmic staining in �20% of tumor cells, regard-less of intensity, yielded the best combination of sensitivity andspecificity.

The discrepancies between these different studies have beenattributed to methodology and equipment (21, 23, 24, 26). Mul-tiple factors have been shown to affect the performance of the VE1antibody. Staining was reported to differ significantly betweendifferent antibody lots from the vendor and different batches ofreagents used (29). In addition, certain pre-analytical variables,including fixation for <6 or >72 hours and a delay in fixation of >6hours, negatively impact the staining pattern and signal intensity(24). Testing of these variables was beyond the scope of this studyusing retrospective specimens, and thus we cannot determinewhether these issues contributed to our observed low sensitivityand specificity. In addition, heat-induced antigen retrieval usingan acidic solution was reported to be suboptimal compared withthe use of a basic solution (29, 31). In our laboratory, we used a

Table 4. Studies of concordance between BRAF V600E IHC (VE1 IHC) and sequencing results in colorectal carcinoma

First author Sequencing methodTissuesource

Detection methodfor IHC

þVE1 IHC/BRAFV600E bysequencing

þVE1 IHC/BRAF WT bysequencing

Sensitivity(%)

Specificity(%)

Adackapara et al. (11) Pyrosequencing WS Manual 12/17 9/35 71 74Affolter et al. (12) Pyrosequencing WS Ventana Ultraview 14/14 0/17 100 100Bledsoe et al. (13) Multiplex PCR TMAþWS Leica Bond-III 57/59 2/145 96 99Capper et al. (14) Sanger and pyrosequencing WS Ventana OptiView 11/11 1/80 100 99Day et al. (15) Sanger and SNaPShot TMA Ventana OptiView 59/59 0/416 100 100Dvorak et al.a (16) Sanger, SNapShot, and NGS TMA Ventana OptiView 86/86 2/193 100 99Kuan et al.b (21) PCR WS Ventana BenchMark Ultra 74/74 3/54 100 94Lasota et al. (17) Multiple analyses WS Leica Bond-Max 24/27 22/86 89 51Loes et al. (22) Sanger and LightMix TMA Ventana BenchMark XT 13/22 12/77 59 84Rossle et al. (18) Sanger and ultra-deep sequencing WS Ventana OptiView 37/37 1/21 100 95Roth et al. (23) Multiplex PCR WS Leica Bond 28/32 0/23 88 100Sinicrope et al. (19) Multiplex allele-specific PCR WS Ventana OptiView 49/49 0/25 100 100Toon et al.c (20) Multiplex PCR and RT-PCR WS Leica Bond-III 43/44 0/157 98 100

Abbreviations: RT, reverse transcription; TMA, tissue microarray; WS, whole tissue sections; WT, wild-type.aOnly colorectal cases were included.bCombined retrospective and prospective.cResults using whole tissue sections only were included.

Estrella et al.





basic solution for antigen retrieval (Tris-EDTA, pH 9.0) with theBond method and used ultra cell conditioning 1 (recommendedby the vendor) with the Ventana method. Studies also showedthat the use of VE1 antibody with the Ventana detection kit andVentana automated stainer yielded optimal results (24). In ourstudy, VE1 IHC using the Ventanamethod on selected cases againprovided results with substantial IHC-sequencing discordance.

Kuan and colleagues performed rigorous optimization of VE1IHC staining conditions and concluded that this step is necessaryto achieve reliable results (29). Our laboratory similarly testednumerous protocols and employed two different automatedstainers to optimize staining conditions for the VE1 antibody toattempt to improve sensitivity and specificity, albeit the VE1antibody optimization relied heavily upon BRAF V600E mutantmelanomas with fewer colorectal carcinoma and papillary thy-roid carcinomas as positive controls. Despite these steps, in ourexperience, VE1 IHC was a suboptimal technique for detectingBRAF V600E mutation in colorectal carcinoma. In contrast, wefound that VE1 IHC reliably detected BRAF V600E mutation inboth melanoma and papillary thyroid carcinoma. Similarly, Loesand colleagues demonstrated high concordance between molec-ular and VE1 IHC results in melanoma (sensitivity, 89%; speci-ficity, 100%) but suboptimal VE1 IHC results in colorectal car-cinoma (sensitivity, 59%; specificity, 84%; ref. 30). Together,these findings suggest that the IHC-sequencing discordanceobserved in colorectal carcinoma may not be secondary to theIHC protocols and platforms used.

Multiple studies have suggested that cases in which VE1 IHCresults are positive but sequencing results are negative might beattributable to the lower sensitivity of molecular analysis(24, 26, 27). Specifically, the contention is that low tumor volumesand dilution with nontumoral tissue can produce a false-negativemolecular result. Although this may be the case in studies usingSanger sequencing in which the limit of detection is approximately20% (i.e., at least 20% of tumor cells must be present in order toreliably detect themutationof interest; ref. 42), ourCAP-accredited,CLIA-certified molecular laboratory has shown the analytical sen-sitivity of pyrosequencing to be 5% (34) and the analytical sensi-tivity of both Sequenom (32) and NGS (33) to be approximately10%. Moreover, in our study, all the 13 cases in which sequenceanalysis indicated wild-type BRAF and VE1 IHC had either mod-erate to strong staining in �20% of tumor cells had adequatetumor formultiple additionalmolecular analyses (SupplementaryTable S3), with informative molecular results reported for 9 of the13 cases, including 4 with KRAS mutation, 2 with high levels ofmicrosatellite instability by PCR, and 2 with CTNNB1 mutation.

Although other studies reported excellent concordancebetweenmolecular analysis and VE1 IHC in some central nervoussystem tumors, specifically epithelioid glioblastomas (43), pleo-morphic xanthoastrocytoma (44), and dysembryoplastic neuroe-pithelial tumor (45), other authors found that the utility of VE1IHC is limited in pituitary adenomas (46), similar to what wefound in colorectal carcinoma. Sperveslage and colleagues eval-uated 78 pituitary adenomas using two different platforms (LeicaBond-Max and Ventana OptiView). They found 10 cases thatexhibited VE1 cytoplasmic staining, including 4 with moderateto strong staining, but did not harbor BRAF V600E mutation bySanger sequencing. Similarly, we and others have reported non-specific VE1 staining on normal cells adjacent to colorectal car-cinoma (20, 24, 25, 28) and in normal anterior pituitary andadrenal cortex (47). Recently, Jones and colleagues showed

sequence homology between the VE1 synthetic peptide andregions of multiple axonemal dynein heavy chain proteins(DNAH2, DNAH7, and DNAH12). Similar proteins may bepresent in some colorectal carcinomas and pituitary adenomasthat cross-react with the VE1 antibody, significantly impacting thespecificity of VE1 IHC in these tumors. It is also possible thattissue-specific factors unique to colonic mucosa (i.e., not presentin melanoma or papillary thyroid carcinoma) influence expres-sion or modification of BRAF proteins independent of mutationstatus, resulting in altered antigenicity (and nonmutation-relatedVE1 IHC reactivity). Moreover, the possible roles of pathwayregulation, especially potential influences on the relationshipbetween gene expression and mutation, in cancers remain to bedetermined. The Cancer Genome Atlas studies report that muta-tion at the DNA level without expression at the mRNA level isrelatively common (48–50). The possibility remains that proteinproduct expression is also influenced by these factors, such asregulatory noncoding RNAs and posttranslational modifications,thereby providing a biologic, rather than a purely methodologic,explanation of some of the discordances between IHC andsequencing results for BRAF in colorectal carcinoma.

Multiple studies have suggested that one of themost importantadvantages of IHC is that tissue-limited samples that could not beanalyzed by molecular techniques would be amenable to IHC(20, 21, 23, 24). In our study, we found significant heterogeneityin VE1 IHC staining in colorectal carcinoma. Among 480 colo-rectal carcinomas evaluated by IHC, 126 cases (26%) had focal(<20%) cytoplasmic staining in tumor cells. Similarly, Affolterand colleagues evaluated 14 colorectal carcinomas with BRAFV600E mutation and found heterogeneous staining in four cases(29%; ref. 20). Thus, significant sampling error may be encoun-tered when VE1 IHC is performed on a biopsy specimen. Incontrast, reliable molecular results may be achieved from limitedtissue. Our laboratory has shown that 10 ng of formalin-fixed,paraffin-embedded DNA is sufficient to amplify mutation hot-spot regions in 46 genes (including BRAF) using Ion TorrentPersonal Genome Machine (NGS; ref. 33). In fact, we routinelyperform DNA sequencing on biopsy specimens, including fine-needle aspiration specimens, in our laboratory.

In summary, we performed the largest study to date to evaluatethe reliability of VE1 IHC as a surrogate marker for BRAF V600Emutation in colorectal carcinoma.We found that, in contrast withwhat has been reported in melanoma and papillary thyroidcarcinoma, the sensitivity and specificity of VE1 IHC in colorectalcarcinoma are suboptimal, indicating that this technique shouldnot be used to guide therapy and clinical management of patientswith colorectal carcinoma. In the era of personalized medicine,more rigorous validation of tests with prognostic and predictiveimportance is necessary to optimize care for cancer patients.

Disclosure of Potential Conflicts of InterestNo potential conflicts of interest were disclosed.

Authors' ContributionsConception and design: J.S. Estrella, A. Rashid, R.R. BroaddusDevelopment ofmethodology: J.S. Estrella, K.P. Patel, A. Rashid, R.R. BroaddusAcquisition of data (provided animals, acquired and managed patients,provided facilities, etc.): J.S. Estrella, M.T. Tetzlaff, K.P. Patel, M.D. Williams,A. Rashid, R.R. BroaddusAnalysis and interpretation of data (e.g., statistical analysis, biostatistics,computational analysis): J.S. Estrella, M.T. Tetzlaff, R.L. Bassett Jr, M.D.Williams,S.R. Hamilton, R.R. Broaddus






Writing, review, and/or revision of the manuscript: J.S. Estrella, M.T. Tetzlaff,R.L. Bassett Jr, M.D. Williams, J.L. Curry, R.R. BroaddusAdministrative, technical, or material support (i.e., reporting or organizingdata, constructing databases): J.S. Estrella, J.L. Curry, S.R. HamiltonStudy supervision: R.R. Broaddus

AcknowledgmentsThe authors thank Kim-Anh Vu from the Department of Pathology Digital

Imaging for her work in figures layout and Stephanie Deming from theDepartment of Scientific Publications for editing the article.

Grant SupportThis work was supported by Frederick F. Becker Distinguished University

Chair in Cancer Research, University of Texas MD Anderson Cancer Center(to S.R. Hamilton).

The costs of publication of this articlewere defrayed inpart by the payment ofpage charges. This article must therefore be hereby marked advertisement inaccordance with 18 U.S.C. Section 1734 solely to indicate this fact.

Received July 23, 2015; revised September 15, 2015; accepted September 18,2015; published OnlineFirst October 5, 2015.

References1. Davies H, Bignell GR, Cox C, Stephens P, Edkins S, Clegg S, et al. Mutations

of the BRAF gene in human cancer. Nature 2002;417:949–54.2. Rad R, Cadinanos J, Rad L, Varela I, Strong A, Kriegl L, et al. A genetic

progression model of Braf(V600E)-induced intestinal tumorigenesisreveals targets for therapeutic intervention. Cancer Cell 2013;24:15–29.

3. Chen D, Huang JF, Liu K, Zhang LQ, Yang Z, Chuai ZR, et al. BRAFV600Emutation and its association with clinicopathological features of colo-rectal cancer: a systematic review and meta-analysis. PLoS One 2014;9:e90607.

4. Samowitz WS, Sweeney C, Herrick J, Albertsen H, Levin TR, Murtaugh MA,et al. Poor survival associated with the BRAF V600E mutation in micro-satellite-stable colon cancers. Cancer Res 2005;65:6063–9.

5. Pai RK, Jayachandran P, Koong AC, Chang DT, Kwok S, Ma L, et al. BRAF-mutated, microsatellite-stable adenocarcinoma of the proximal colon: anaggressive adenocarcinoma with poor survival, mucinous differentiation,and adverse morphologic features. Am J Surg Pathol 2012;36:744–52.

6. Richman SD, SeymourMT,Chambers P, Elliott F,DalyCL,MeadeAM, et al.KRAS and BRAF mutations in advanced colorectal cancer are associatedwith poor prognosis but do not preclude benefit from oxaliplatin oririnotecan: results from the MRC FOCUS trial. J Clin Oncol 2009;27:5931–7.

7. Palomaki GE, McClain MR, Melillo S, Hampel HL, Thibodeau SN. EGAPPsupplementary evidence review: DNA testing strategies aimed at reducingmorbidity and mortality from Lynch syndrome. Genet Med 2009;11:42–65.

8. DiNicolantonio F,MartiniM,Molinari F, Sartore-Bianchi A, Arena S, SalettiP, et al. Wild-type BRAF is required for response to panitumumab orcetuximab in metastatic colorectal cancer. J Clin Oncol 2008;26:5705–12.

9. Di Fiore F, Sesboue R, Michel P, Sabourin JC, Frebourg T. Moleculardeterminants of anti-EGFR sensitivity and resistance in metastatic colorec-tal cancer. Br J Cancer 2010;103:1765–72.

10. Laurent-Puig P, Cayre A, Manceau G, Buc E, Bachet JB, Lecomte T, et al.Analysis of PTEN, BRAF, and EGFR status in determining benefit fromcetuximab therapy inwild-type KRASmetastatic colon cancer. J ClinOncol2009;27:5924–30.

11. De Roock W, De Vriendt V, Normanno N, Ciardiello F, Tejpar S. KRAS,BRAF, PIK3CA, and PTENmutations: implications for targeted therapies inmetastatic colorectal cancer. Lancet Oncol 2011;12:594–603.

12. Pietrantonio F, Petrelli F, Coinu A, Di BartolomeoM, Borgonovo K, MaggiC, et al. Predictive role of BRAF mutations in patients with advancedcolorectal cancer receiving cetuximab and panitumumab: a meta-analysis.Eur J Cancer 2015;51:587–94.

13. Bokemeyer C, Van Cutsem E, Rougier P, Ciardiello F, Heeger S, SchlichtingM, et al. Addition of cetuximab to chemotherapy as first-line treatment forKRAS wild-type metastatic colorectal cancer: pooled analysis of the CRYS-TAL and OPUS randomised clinical trials. Eur J Cancer 2012;48:1466–75.

14. Zecchin D, Boscaro V, Medico E, Barault L, Martini M, Arena S, et al. BRAFV600E is a determinant of sensitivity to proteasome inhibitors.Mol CancerTher 2013;12:2950–61.

15. Falchook GS, Long GV, Kurzrock R, Kim KB, Arkenau TH, BrownMP, et al.Dabrafenib in patients with melanoma, untreated brain metastases, andother solid tumours: a phase 1 dose-escalation trial. Lancet 2012;379:1893–901.

16. Kopetz S, Desai J, Chan E, Hecht JR, O'Dwyer PJ, Lee RJ, et al. PLX4032 inmetastatic colorectal cancer patients with mutant BRAF tumors. J ClinOncol 2010;28:15s (suppl; abstr 3534).

17. Mao M, Tian F, Mariadason JM, Tsao CC, Lemos R Jr., Dayyani F, et al.Resistance to BRAF inhibition in BRAF-mutant colon cancer can be over-come with PI3K inhibition or demethylating agents. Clin Cancer Res2013;19:657–67.

18. Yaeger R, Cercek A, O'Reilly EM, Reidy DL, Kemeny N, Wolinsky T, et al.Pilot trial of combined BRAF and EGFR inhibition in BRAF-mutantmetastatic colorectal cancer patients. Clin Cancer Res 2015;21:1313–20.

19. Adackapara CA, Sholl LM, Barletta JA, Hornick JL. Immunohistochemistryusing the BRAFV600Emutation-specificmonoclonal antibodyVE1 is not auseful surrogate for genotyping in colorectal adenocarcinoma. Histopa-thology 2013;63:187–93.

20. Affolter K, Samowitz W, Tripp S, Bronner MP. BRAF V600E mutationdetection by immunohistochemistry in colorectal carcinoma. Genes Chro-mosomes Cancer 2013;52:748–52.

21. Bledsoe JR, Kamionek M, Mino-Kenudson M. BRAF V600E immunohis-tochemistry is reliable in primary and metastatic colorectal carcinomaregardless of treatment status and shows high intratumoral homogeneity.Am J Surg Pathol 2014;38:1418–28.

22. CapperD, Voigt A, BozukovaG, Ahadova A, Kickingereder P, vonDeimlingA, et al. BRAF V600E-specific immunohistochemistry for the exclusion ofLynch syndrome in MSI-H colorectal cancer. Int J Cancer 2013;133:1624–30.

23. Day F, Muranyi A, Singh S, Shanmugam K, Williams D, Byrne D, et al. Amutant BRAF V600E-specific immunohistochemical assay: correlationwith molecular mutation status and clinical outcome in colorectal cancer.Targeted Oncol 2015;10:99–109.

24. Dvorak K, Aggeler B, Palting J, McKelvie P, Ruszkiewicz A, Waring P.Immunohistochemistrywith the anti-BRAF V600E (VE1) antibody: impactof pre-analytical conditions and concordance with DNA sequencing incolorectal and papillary thyroid carcinoma. Pathology 2014;46:509–17.

25. Lasota J, Kowalik A, Wasag B, Wang ZF, Felisiak-Golabek A, Coates T, et al.Detection of the BRAF V600E mutation in colon carcinoma: criticalevaluation of the imunohistochemical approach. Am J Surg Pathol 2014;38:1235–41.

26. Rossle M, Sigg M, Ruschoff JH, Wild PJ, Moch H,Weber A, et al. Ultra-deepsequencing confirms immunohistochemistry as a highly sensitive andspecific method for detecting BRAF V600E mutations in colorectal carci-noma. Virchows Archiv 2013;463:623–31.

27. Sinicrope FA, Smyrk TC, Tougeron D, Thibodeau SN, Singh S, Muranyi A,et al. Mutation-specific antibody detects mutant BRAFV600E proteinexpression in human colon carcinomas. Cancer 2013;119:2765–70.

28. Toon CW, Walsh MD, Chou A, Capper D, Clarkson A, Sioson L, et al.BRAFV600E immunohistochemistry facilitates universal screening of colo-rectal cancers for Lynch syndrome. Am J Surg Pathol 2013;37:1592–602.

29. Kuan SF, Navina S, Cressman KL, Pai RK. Immunohistochemical detectionof BRAF V600E mutant protein using the VE1 antibody in colorectalcarcinoma is highly concordant with molecular testing but requires rigor-ous antibody optimization. Hum Pathol 2014;45:464–72.

30. Loes IM, Immervoll H, Angelsen JH, Horn A, Geisler J, Busch C, et al.Performance comparison of three BRAF V600E detection methods inmalignant melanoma and colorectal cancer specimens. Tumour Biol2015;36:1003–13.

31. Roth RM, Hampel H, Arnold CA, Yearsley MM, Marsh WL, Frankel WL. Amodified Lynch syndrome screening algorithm in colon cancer: BRAFimmunohistochemistry is efficacious and cost beneficial. Am J Clin Pathol2015;143:336–43.

Estrella et al.





32. Greaves WO, Verma S, Patel KP, Davies MA, Barkoh BA, Galbincea JM,et al. Frequency and spectrum of BRAF mutations in a retrospective,single-institution study of 1112 cases of melanoma. J Mol Diagn 2013;15:220–6.

33. Singh RR, Patel KP, Routbort MJ, Reddy NG, Barkoh BA, Handal B, et al.Clinical validation of a next-generation sequencing screen for muta-tional hotspots in 46 cancer-related genes. J Mol Diagn 2013;15:607–22.

34. Verma S, Greaves WO, Ravandi F, Reddy N, Bueso-Ramos CE, O'Brien S,et al. Rapid detection and quantitation of BRAF mutations in hairy cellleukemia using a sensitive pyrosequencing assay. Am J Clin Pathol2012;138:153–6.

35. Tetzlaff MT, Pattanaprichakul P, Wargo J, Fox PP, Patel KP, Estrella JS, et al.Utility of BRAF V600E immunohistochemical expression pattern as asurrogate of BRAF mutation status in 154 patients with advanced mela-noma. Hum Pathol 2015;46:1101–10.

36. Capper D, Berghoff AS, Magerle M, Ilhan A, Wohrer A, Hackl M, et al.Immunohistochemical testing of BRAF V600E status in 1,120 tumor tissuesamples of patients with brain metastases. Acta Neuropathol 2012;123:223–33.

37. Long GV, Wilmott JS, Capper D, Preusser M, Zhang YE, Thompson JF,et al. Immunohistochemistry is highly sensitive and specific for thedetection of V600E BRAF mutation in melanoma. Am J Surg Pathol2013;37:61–5.

38. Busam KJ, Hedvat C, Pulitzer M, von Deimling A, Jungbluth AA. Immu-nohistochemical analysis of BRAF(V600E) expression of primary andmetastatic melanoma and comparison with mutation status and melano-cyte differentiation antigens of metastatic lesions. Am J Surg Pathol2013;37:413–20.

39. Feller JK, Yang S,MahalingamM. Immunohistochemistry with amutation-specific monoclonal antibody as a screening tool for the BRAFV600Emutational status in primary cutaneousmalignantmelanoma.ModPathol2013;26:414–20.

40. Ilie MI, Lassalle S, Long-Mira E, Bonnetaud C, Bordone O, Lespinet V, et al.Diagnostic value of immunohistochemistry for the detection of the BRAF

(V600E) mutation in papillary thyroid carcinoma: comparative analysiswith three DNA-based assays. Thyroid 2014;24:858–66.

41. McKelvie PA, Chan F, Yu Y, Waring P, Gresshoff I, Farrell S, et al. Theprognostic significance of the BRAF V600E mutation in papillary thyroidcarcinoma detected by mutation-specific immunohistochemistry. Pathol-ogy 2013;45:637–44.

42. Harrington CT, Lin EI, Olson MT, Eshleman JR. Fundamentals of pyrose-quencing. Arch Pathol Lab Med 2013;137:1296–303.

43. Kleinschmidt-DeMasters BK, Aisner DL, Foreman NK. BRAF VE1 immu-noreactivity patterns in epithelioid glioblastomas positive for BRAF V600Emutation. Am J Surg Pathol 2015;39:528–40.

44. Ida CM, Vrana JA, Rodriguez FJ, Jentoft ME, Caron AA, Jenkins SM, et al.Immunohistochemistry is highly sensitive and specific for detection ofBRAF V600E mutation in pleomorphic xanthoastrocytoma. Acta Neuro-pathol Commun 2013;1:20.

45. Chappe C, Padovani L, Scavarda D, Forest F, Nanni-Metellus I, Loundou A,et al. Dysembryoplastic neuroepithelial tumors share with pleomorphicxanthoastrocytomas and gangliogliomas BRAF(V600E) mutation andexpression. Brain Pathol 2013;23:574–83.

46. Sperveslage J, GierkeM, Capper D, Honegger J, Sipos B, Beschorner R, et al.VE1 immunohistochemistry in pituitary adenomas is not associated withBRAF V600E mutation. Acta Neuropathol 2013;125:911–2.

47. Mordes DA, Lynch K, Campbell S, Dias-Santagata D, Nose V, Louis DN,et al. VE1 antibody immunoreactivity in normal anterior pituitary andadrenal cortexwithout detectable BRAFV600Emutations. Am JClin Pathol2014;141:811–5.

48. Ciriello G, Miller ML, Aksoy BA, Senbabaoglu Y, Schultz N, Sander C.Emerging landscape of oncogenic signatures across human cancers. NatGenet 2013;45:1127–33.

49. Cancer Genome Atlas Research N Weinstein JN, Collisson EA, Mills GB,Shaw KR, Ozenberger BA, et al. The Cancer Genome Atlas Pan-Canceranalysis project. Nat Genet 2013;45:1113–20.

50. Akbani R,NgPK,WernerHM, ShahmoradgoliM, Zhang F, Ju Z, et al. Apan-cancer proteomic perspective on The Cancer Genome Atlas. Nat Commun2014;5:3887.






2015;14:2887-2895. Published OnlineFirst October 5, 2015.Mol Cancer Ther Jeannelyn S. Estrella, Michael T. Tetzlaff, Roland L. Bassett, Jr, et al. SequencingTissue-Specific Discordances between Immunohistochemistry and

V600E Status in Colorectal Carcinoma:BRAFAssessment of

Updated version

10.1158/1535-7163.MCT-15-0615doi:

Access the most recent version of this article at:

Material

Supplementary

http://mct.aacrjournals.org/content/suppl/2015/10/02/1535-7163.MCT-15-0615.DC1

http://mct.aacrjournals.org/content/suppl/2015/10/20/1535-7163.MCT-15-0615.DC2Access the most recent supplemental material at:

Cited articles

http://mct.aacrjournals.org/content/14/12/2887.full#ref-list-1

This article cites 49 articles, 7 of which you can access for free at:

Citing articles

http://mct.aacrjournals.org/content/14/12/2887.full#related-urls

This article has been cited by 3 HighWire-hosted articles. Access the articles at:

E-mail alerts related to this article or journal.Sign up to receive free email-alerts

Subscriptions

Reprints and

[email protected]

To order reprints of this article or to subscribe to the journal, contact the AACR Publications Department at

Permissions

Rightslink site. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC)

.http://mct.aacrjournals.org/content/14/12/2887To request permission to re-use all or part of this article, use this link



http://mct.aacrjournals.org/lookup/doi/10.1158/1535-7163.MCT-15-0615

http://mct.aacrjournals.org/content/suppl/2015/10/20/1535-7163.MCT-15-0615.DC2 http://mct.aacrjournals.org/content/suppl/2015/10/02/1535-7163.MCT-15-0615.DC1

http://mct.aacrjournals.org/content/suppl/2015/10/20/1535-7163.MCT-15-0615.DC2 http://mct.aacrjournals.org/content/suppl/2015/10/02/1535-7163.MCT-15-0615.DC1

http://mct.aacrjournals.org/content/14/12/2887.full#ref-list-1

http://mct.aacrjournals.org/content/14/12/2887.full#related-urls

http://mct.aacrjournals.org/cgi/alerts

mailto:[email protected]

http://mct.aacrjournals.org/content/14/12/2887


Documents

Assessment of BRAF V600E Status in Colorectal Carcinoma ... · conclude that VE1 IHC produces suboptimal results in colo-rectal carcinoma and should not be used to guide patient management