Histopathology 2014, 64, 826–839. DOI: 10.1111/his.12331

Epidermal growth factor receptor mutation-specific immunohistochemical antibodies in lung adenocarcinoma Ghassan Allo,1,2 Bizhan Bandarchi,1,3 Naoki Yanagawa,1 Ami Wang,1 Warren Shih,1 Jing Xu,1 Morgan Dalby,4 Hiroaki Nitta,4 Christine To,1 Ni Liu,1 Jenna Sykes5 & Ming S Tsao1,2 1

Department of Pathology, Princess Margaret Cancer Centre, University Health Network, Toronto, ON, Canada, Department of Laboratory Medicine and Pathobiology, University of Toronto, Toronto, ON, Canada, 3Department of Pathology, University of California Los Angeles (UCLA), Los Angeles, CA, USA, 4Ventana Medical Systems, Tucson, AZ, USA, and 5Department of Biostatistics, Princess Margaret Cancer Centre, University Health Network, Toronto, ON, Canada

2

Date of submission 7 July 2013 Accepted for publication 18 November 2013 Published online Article Accepted 20 November 2013

Allo G, Bandarchi B, Yanagawa N, Wang A, Shih W, Xu J, Dalby M, Nitta H, To C, Liu N, Sykes J & Tsao M S (2014) Histopathology 64, 826–839

Epidermal growth factor receptor mutation-specific immunohistochemical antibodies in lung adenocarcinoma Aims: We investigated the sensitivity and specificity of two novel Epidermal growth factor receptor (EGFR) mutation-specific antibodies in the detection of the most common EGFR mutations in lung adenocarcinoma. Methods and results: A total of 241 resected lung adenocarcinoma specimens and six resected post-neoadjuvant gefitinib adenocarcinomas were analysed for EGFR mutation using mass spectrometry, fragment analysis and direct PCR sequencing platforms. Tissue arrays and/or full sections of these cases were evaluated using immunohistochemistry with two novel antibodies (clones SP125 and SP111) and two previously reported antibodies (clones 43B2 and 6B6), specific for L858R or 15-nucleotide exon-19 deletion EGFR mutations. SP125 antibody detected EGFR L858R mutation

with a sensitivity of 76% and positive predictive value of 73%. SP111 antibody stained the 15-nucleotide EGFR exon-19 deletions with a sensitivity of 83% and a positive predictive value of 94%. Pretreatment with gefitinib did not affect antibody performance. Full-section immunohistochemical staining detected heterogeneous mutant EGFR proteins expression in tumours, and revealed L858R mutation in the non-neoplastic bronchial epithelium adjacent to EGFR L858R-carrying carcinomas in three of 16 (19%) cases. Conclusions: Immunohistochemistry using EGFR mutant-specific antibodies may be useful in shortening the diagnostic time of lung adenocarcinoma with most common EGFR mutations, especially in samples with low tumour cellularity.

Keywords: epidermal growth factor receptor, immunohistochemistry, lung adenocarcinoma, mutation

Introduction Lung cancer is the most commonly diagnosed cancer and the leading cause of cancer-related deaths worldwide.1 Eighty per cent of lung cancers are non-small cell lung carcinomas (NSCLCs), with adenocarcinomas Address for correspondence: Dr M S Tsao, Princess Margaret Hospital, 610 University Avenue, 7th Floor, Room 7-611, Toronto, ON M5G 2M9, Canada. e-mail: [email protected] © 2013 John Wiley & Sons Ltd.

being the most common histological subtype.2,3 Epidermal growth factor receptor (EGFR) is a member of the HER/erbB family of receptor tyrosine kinases (TKs), comprising extracellular cysteine-rich ligand-binding, single a-helix transmembrane, cytoplasmic TK and carboxy-terminal signalling domains.4 EGFR dimerization, in response to ligand binding, leads to autophosphorylation of its cytoplasmic domain, interaction with adaptor molecules and activation of downstream signalling pathways, resulting in increased tumour cell

EGFR immunohistochemistry in lung cancer

proliferation, survival and motility.4 Therefore, an inappropriate increase in EGFR TK activity may contribute to tumour progression, and inhibition of this TK activity by small molecular tyrosine kinase inhibitors (TKIs) has been shown to confer survival benefits for advanced NSCLC patients.5–7 A greater response to TKIs has been observed in tumours harbouring activating EGFR TK domain mutations compared to those with wild-type EGFR.6,8–13 These activating mutations have been reported in 10–50% of NSCLCs, associated commonly with adenocarcinoma, female gender, East Asian ethnicity and never smokers.14 More than 85% of these mutations are exon-19 in-frame deletions and exon-21 L858R missense mutations.15,16 There are more than 20 variants of exon-19 deletions, with the most common being delE746-A750, delL747T751insS and delL747-P753insS.17 Testing for the presence of sensitizing mutations in advanced non-squamous NSCLCs has been recommended for selecting patients to receive first-line EGFR TKI therapy.18–20 There are various methods to detect EGFR deletion and point mutations, including direct sequencing of polymerase chain reaction (PCR)-amplified genomic DNA, high-resolution melting analysis, fragment analysis, restriction fragment length polymorphism, the amplification refractory mutation system, and the mass spectrometry-based MassArray platform (Sequenom, San Diego, CA, USA).4 These molecular methods are generally costly. Their sensitivity might vary with the quality of the extracted DNA, tumour cellularity, and the relative contamination with the non-mutated allele. To overcome some of these limitations, mutation-specific antibodies were developed to detect mutant EGFR by immunohistochemistry (IHC).21 In this study we report two novel antibodies targeting EGFR protein carrying L858R substitution (SP125) and 15-nucleotide E746-A750 deletion mutations (SP111). The sensitivity and specificity of these antibodies are compared with two similar antibodies that have been assessed previously.21

Materials and methods PATIENTS AND CLINICAL SAMPLES

Using a protocol approved by the Research Ethics Board at the University Health Network (number 08– 0534-TE), 244 patients with primary lung adenocarcinoma who had undergone surgical resection from 2001 to 2008 were identified in the laboratory information system. Clinical and pathological data were reviewed. Archival formalin-fixed paraffin-embedded (FFPE) tumour tissue blocks were retrieved. Tumour© 2013 John Wiley & Sons Ltd, Histopathology, 64, 826–839.

827

rich areas were annotated for tissue microarray (TMA) construction and for DNA isolation. Quadruple-core (three tumour and one non-tumour control cores) TMA was constructed using a 0.6-mm needle microarray corer (Beecher Instruments, Inc., Sun Prairie, WI, USA). DNA was isolated from tumour-rich cores of FFPE tissue blocks using a 1-mm core needle. In addition, IHC-positive areas from mutation-negative tumours and IHC-positive non-neoplastic bronchial epithelium were recored and reanalysed by macrodissecting IHC-positive areas from 5-l-thick sections stained with toluidine blue (Fisher Scientific Canada, Ottowa, ON, Canada). For DNA extraction, FFPE tumour tissue was deparaffinized using graded concentrations of xylene and alcohol. This was followed by overnight digestion with proteinase K. DNA was extracted using the phenol chloroform method and quantified using spectrophotometry. EGFR MUTATIONS ANALYSIS

All cases were screened initially for mutations using the OncoCarta Panel version 1.0 MassArray System (Sequenom), which analyses for 238 somatic mutations in 19 oncogenes.22 Negative cases (n = 72) were screened further for exon-19 deletions using the fragment length analysis method, as described previously.23 All mutations were confirmed using dideoxy nucleotide sequencing, as described previously.11 IMMUNOHISTOCHEMISTRY

TMA sections and control slides were stained using two anti-EGFR L858R rabbit monoclonal antibodies, SP125 (Ventana Medical Systems, Tucson, AZ, USA) and 43B2 (Cell Signalling, Danvers, MA, USA), and two anti-EGFR E746-A750 deletion rabbit monoclonal antibodies, SP111 (Ventana) and 6B6 (Cell Signalling) (Supporting information, Table S1). In addition, tumour full-section staining was performed on: (i) cases with inadequate tumour on the TMA; (ii) cases with discrepant IHC results on TMA (see below); (iii) cases that were IHC-positive on the TMA but lacked the specific EGFR mutation (possible false-positive); and (iv) in all cases with a proven EGFR mutation, regardless of their IHC status. Fourl-thick sections were stained using each of the four antibodies, according to the suppliers’ protocol. Scoring of immunohistochemical expression All slides were reviewed and scored by at least two observers, blinded to the EGFR mutation status. Interobserver concordance was assessed; discordant cases were

828

G Allo et al.

Table 1. The immunohistochemical scoring system for assessing the cytoplasmic and membranous expression of all four antibodies used in this study, and the criteria used to determine positivity (adapted from Brevet et al.24) Cytoplasmic staining

Membranous staining

Score 0: No staining OR faint staining in 10%

Score 1+: Faint partial membrane staining

Score 2+: Moderate staining, any %

Score 2+: Weak, complete staining in >10%

Score 3+: Strong staining, any %

Score 3+: Intense, complete staining in >10%

ies, we analysed tumours from patients who received preresection gefitinib therapy in a Phase II neoadjuvant study.7,25 Two and four cases were stained with SP125 and SP111, respectively, including cases with known L858R mutation, 15-nucleotide exon-19 deletion and other point mutations. STATISTICS

Criteria A (‘low’ threshold) Positive: At least one TMA core is 1+

Positive: At least one TMA core is 1+

Negative: All TMA cores are 0

Negative: All TMA cores are 0

Criteria B (‘high’ threshold)

The concordance between the two scores was compared using the prevalence-adjusted, bias-adjusted kappa to minimise the prevalence bias that might occur, as the majority of the observations were classified as negative.26 The performance of each antibody was assessed by evaluating sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV) and the area under the receiver operating characteristic curve (AUC). A non-parametric test27 was used to compare the AUCs for the antibodies. All statistical analyses were performed using R version 2.12.1. The package ‘concord’ was used to generate kappa values and the package ‘clinfun’ was used to compare AUCs. A P-value of G) exon-21 missense mutation, which occurred in 20 of 49 (40.8%) tumours. This was followed by the 15-nucleotide exon-19 deletions, which included 10 K745-A750 deletions and seven E746-A750 deletions. One adenocarcinoma harboured both L858R and K745-A750 deletion mutations. Non-15-nucleotide deletions constituted six of 24 (25%) exon-19 deletions. EXON-21 L858R-SPECIFIC ANTIBODIES

SP125 antibody immunohistochemistry For SP125 staining, the OptiView DAB IHC Detection Kit was used and 239 tumours on the TMAs were evaluable. Using low- and high-scoring thresholds, one observer scored 28 (11.7%) and 21 (8.7%) among the 239 tumours as positive, respectively © 2013 John Wiley & Sons Ltd, Histopathology, 64, 826–839.

EGFR immunohistochemistry in lung cancer

A

B

C

D

829

Figure 1. EGFR mutation-specific immunohistochemistry of lung adenocarcinoma TMA cores. Examples of cases scored as: (A) negative, (B) 1+, (C) 2+ and (D) 3+ positivity are given (SP125).

(Supporting information, Table S2). The corresponding scores for the second observer were 26 (9.9%) and 16 (6.6%), respectively. The prevalence-adjusted, bias-adjusted kappa coefficient between the two observers reached 0.90 and 0.96 for the low and high threshold criteria, respectively. Full-section tumour slides from 42 cases were stained using the SP125 antibody and the OptiView © 2013 John Wiley & Sons Ltd, Histopathology, 64, 826–839.

method (Supporting information, Table S3). Twentyone tumours were scored as positive for displaying clear homogeneous cytoplasmic staining and three also showed complete strong membranous staining (Figure 2A). Three of the remaining cases demonstrated apical granular positivity focally within the tumour cells (Figure 2B); these were scored as negative for the lack of homogeneous cytoplasmic staining,

830

G Allo et al.

Table 2. Clinical characteristics of adenocarcinoma patients evaluated in this study Frequency (%) Patient demographics (n = 235)* Sex Female Male Age (years) Median (range)

132 (56.2) 103 (43.8) 66.8 (40.7–88.0)

Tumour characteristics (n = 241) Differentiation grade Well Moderate Poor Stage (pT) 1a

54 (22.4) 118 (49) 69 (28.6) 35 (14.5)

1b

38 (15.8)

2a

107 (44.4)

2b

19 (7.9)

3

26 (10.8)

4

16 (6.6)

*Six patients had two synchronous tumours in the resected specimens.

and were L858R-negative by molecular testing. The presence of focal apical granules in these three cases has led at least one of the two observers to score them as positive on the TMA. Based on the scoring results of the TMA and the full-section slides for both observers, the final number of positive cases was 22 of 241 cases (9.1%) (Table 4). 43B2 antibody immunohistochemistry UltraView was used as the detection system for 43B2 antibody. Nineteen (8.0%) and seven (3.0%) of 237 cases were scored as positive using the low and high thresholds, respectively (Supporting information, Table S2). Full-section tumour slides from 18 cases that were discrepant between the TMA IHC and mutation status or were missing in the TMAs were also stained with this antibody (Supporting information, Table S3); only 17 contained sufficient evaluable tumour cells. One case was found to be positive with 43B2 IHC but had an EGFR wild-type genotype. The remaining 16 negative cases included eight false-negative cases

(seven with L858R mutation and one with both K745-A750 deletion and L858R mutation) and eight true negative cases (six wild-type EGFR and two with non-L858R substitution mutations). None of the cases demonstrated positivity in the non-neoplastic tissue. After review of discordant cases and full-section slides, the final number of positive cases was 15 of 241 tumours (6.2%) (Table 4). Correlation of L858R immunohistochemistry with mutation status Table 5 summarizes the correlation between the presence of EGFR L858R mutation and the performance of the two mutation-specific IHC methods. While the specificity of both the SP125 and 43B2 antibodies was comparable, with a low scoring threshold at 97% and 99%, respectively, the sensitivity of SP125 was higher than that of 43B2 antibody (76% versus 62%). When the accuracy of each of the IHC method was compared by the AUC, SP125 had a slightly higher AUC than 43B2, but this was not statistically significant (P = 0.233) (Figure 3A). Using the ‘high threshold’, the sensitivity for these antibodies dropped significantly (SP125, 62%; 43B2, 24%) (Supporting information, Table S4). Initially, seven adenocarcinomas showed false-positive SP125 staining (SP125-positive on TMA; L858R mutation-negative on initial genotype screening). Full sections of these seven cases were stained with SP125, revealing heterogeneous positivity; tumour areas, from which tissue cores were obtained for DNA extraction, were SP125-negative (Figure 2C–D). Therefore, we macrodissected SP125-positive tumour areas, extracted DNA and resequenced for L858R mutation. Resequencing showed L858R mutation in one of the seven cases, with the remaining six cases being false-positive for L858R (Tables 3–5). This result suggests the presence of mutation heterogeneity in this tumour. For the remaining six false-positive cases, we compared the pattern of full-section staining to that of the 15 true-positive cases. False-positive staining areas tend to be less extensive than for true-positive tumours (mean percentage positivity 35% versus 62%, respectively, P = 0.047), with four of the six (67%) having ≤20% positively stained areas compared to only one of the 15 (6.7%) true-positive cases showing ≤20% positively stained areas. All false-positive cases showed a staining intensity of 1+, while eight of the 15 (53%) true-positive cases stained 1+. These results suggest that with low staining intensity, SP125 may potentially give false-positive results in six of 220 (2.7%) of cases. © 2013 John Wiley & Sons Ltd, Histopathology, 64, 826–839.

EGFR immunohistochemistry in lung cancer

831

Table 3. Sensitivity and specificity of four EGFR-mutant specific antibodies for tumours with various types of kinase domain mutations Mutation status

Number of positively stained tumours

Number Exon-19 deletions

L858R mutation

Ex19-deletion

SP125

SP111

43B2

6B6

24

2/24

0/24

15/24

14/24

10nt del. L747-TTAAGAGAAG-A750, ins. C

2

1/2

0/2

0/2

0/2

13nt del. L747-TTAAGAGAAGCAA-T751, ins. C

1

0/1

0/1

0/1

0/1

15nt del. K745-GGAATTAAGAGAAGC-A750

10

1/10

0/10

9/10

10/10

15 nt del. E746-GAATTAAGAGAAGCA-A750

6

0/6

0/6

5/6

3/6

15nt del. E746-GAATTAAGAGAAGCA-A750, V742I*

1

0/1

0/1

1/1

1/1

15nt del. K745-GGAATTAAGAGAAGC-A750, L858R*

1

0/1

0/1

0/1

0/1

18nt del. L747-TAAGAGAAGCAACATCTC-P753

1

0/1

0/1

0/1

0/1

19nt del. E746-AATTAAGAGAAGCAACATC-S752, ins.T

2

0/2

0/2

0/2

0/2

Point mutations

25

L858R

20

16/20

13/20

1/20

1/20

G719C, S768I

1

0/1

0/1

0/1

0/1

P733S

1

0/1

0/1

0/1

0/1

E758K

1

0/1

0/1

0/1

0/1

D830N

1

0/1

0/1

0/1

0/1

L861Q

1

0/1

0/1

0/1

0/1

EGFR wild-type

192

4/192

2/192

0/192

3/192

Total

241

22/241

15/241

16/241

18/241

*Combined point mutation and exon-19 deletion.

The three tumours with coarse apical granules did not harbour L858R mutation, but each of them harboured a form of exon-19 deletion: 15-nt E746-A750 del, 15-nt K745-A750 del and 13nt del. L747T751ins C, respectively. EXON-19 DELETION-SPECIFIC ANTIBODIES

SP111 antibody immunohistochemistry The number of assessable adenocarcinoma cases on the TMA for SP111 staining was 235 (Supporting information, Table S2). The number of cases scored positive was 27 (11.5%) and 19 (8.1%) using low and high positivity thresholds, and with the prevalence-adjusted, bias-adjusted kappa of 0.86 and 0.9 (Supporting information, Table S2), respectively. © 2013 John Wiley & Sons Ltd, Histopathology, 64, 826–839.

Review of discordant cases revealed the presence of a heterogeneous coarse granular cytoplasmic positivity that was considered positive by one observer. After a consensus discussion among the group, this granular pattern was discounted, and only a diffuse homogeneous cytoplasmic positivity was considered positive. Up to nine (3.7%) demonstrated apparent membranous positivity. Full sections from 51 cases were stained using the SP111 antibody (Supporting information, Table S3) and 16 (31.4%) revealed homogeneous cytoplasmic positivity in the tumour cells. In the remaining cases, tumour cell cytoplasm was devoid of any staining. Including discordant cases and full-section slide staining, the final overall number of cases positive for SP111 was 16 (6.6%) (Table 4).

832

G Allo et al.

A

B

† ‡ *

C

D

Figure 2. Staining patterns observed with EGFR mutant-specific antibodies. (A), SP111 immunohistochemistry of full-section adenocarcinoma showing homogeneous cytoplasmic and membranous positivity. SP125 staining showed a similar staining pattern. (B), SP125 immunohistochemistry of full-section adenocarcinoma showing apical coarse granules. This pattern was not considered as truly positive. (C, D) Two fields from a case with tumour heterogeneity for SP125 expression. Incidentally, coring for the TMA (marked as †) was obtained from an SP125-positive tumour area (C), while that for DNA extraction (marked as ‡) was obtained from an SP125-negative tumour area (marked as *) (D). Testing DNA extracted from the SP125-positive tumour area confirmed the presence of L858R mutation in this case that was missed on the initial assessment.

Table 4. Results of immunohistochemical staining for the four EGFR mutation-specific antibodies

Antibody

Number positive (%)

Number negative (%)

Total evaluated

SP125

22 (9.1)

219 (90.9)

241

43B2

15 (6.2)

226 (93.8)

241

SP111

16 (6.6)

225 (93.4)

241

6B6

21 (8.7)

220 (91.3)

241

6B6 antibody Thirty-four (14.4%) of the 236 assessable TMA cases stained positively using the exon-19 deletion-specific

antibody 6B6. For various reasons, full-section staining was performed on 31 cases (Supporting information, Table S3). Five cases showed positive staining. After taking into account the results of full-section staining of the discordant cases, the total number of true 6B6-positive cases was 21 of 241 (8.7%) (Table 4). Correlation of exon-19 deletion immunohistochemistry with mutation status Lower cytoplasmic intensity and scoring thresholds were more sensitive than stronger intensity in recognizing 15-nt exon-19 deletions, as 13 of the 15 (86.7%) SP111-positive carcinomas with such mutations showed 1+ IHC intensity on full-section staining. Table 6 and © 2013 John Wiley & Sons Ltd, Histopathology, 64, 826–839.

EGFR immunohistochemistry in lung cancer

833

Table 5. Correlations between immunohistochemistry using the EGFR exon-21 L858R mutant-specific antibodies with lowthreshold scoring, and the presence of L858R mutation L858R mutation Antibody SP125

43B2

Present

Absent

Total

Positive

16

6

22

Negative

5

214

219

Total

21

220

241

Positive

13

2

15

Negative

8

218

226

21

220

241

Total

Sensitivity (%)

Specificity (%)

PPV (%)

NPV (%)

76.2

97.3

72.7

97.7

61.9

99.1

86.7

96.4

PPV, positive predictive value; NPV, negative predictive value.

Supporting information, Table S5, summarize the correlation between the presence of 15-nucleotide exon-19 deletion and the performance of the two deletion-specific antibodies, and Figure 3B–D illustrates the accuracy of each of these antibodies in detecting the specific deletion. For predicting the 15-nt E746-A750 deletion, SP111 had a borderline significantly higher AUC than 6B6 (0.907 versus 0.897, respectively, P = 0.057). However, there was no difference between these two antibodies in their ability to detect the 15-nt K745-A750 deletion or to detect both 15-nt exon-19 deletions (P = 0.42 and 0.51, respectively). Six of the seven adenocarcinomas (85.7%) with 15-nt E746-A750 were positive using SP111 (Table 3 and Supporting information, Table S5). The performance of this antibody was comparable when the 15-nucleotide deletion starting at one codon in the 5′ direction (15-nt K745-A750) was analysed (nine of 11, 81.8%). However, it was negative in all four tumours carrying an exon-19 deletion starting one codon in the 3′ direction at L747 (Table 3). Moreover, while the case of combined 15-nt E746A750 deletion and V742I mutation was positive for SP111, the tumour with combined K745-A750 deletion and L858R mutations and those with 19nt del. E746 -S752 insT were negative using SP111. Effect of TKIs on the performance of EGFR mutationspecific antibodies Immunohistochemical staining with SP125 revealed moderate cytoplasmic and membranous positivity in more than 95% of tumour cells in the tumour harbouring the L858R EGFR mutation and negativity of the control non-L858R case. Furthermore, staining with SP111 was positive in the adenocarcinoma with © 2013 John Wiley & Sons Ltd, Histopathology, 64, 826–839.

the 15-nucleotide E746-A750 deletion, but negative in the remaining three cases with other forms of the exon-19 deletion genotype (Table 7). Staining in the non-tumour component Full-section staining using SP125 displayed cytoplasmic positivity in some of the histologically normal non-neoplastic bronchial epithelium in 36 of 42 cases studied (85.7%), either as homogeneous cytoplasmic positivity (eight cases, 22.2%), granular staining (four cases, 11.1%) or a combination of both (24 cases, 66.7%) (Figure 4A). Macrophages also demonstrated positive granular cytoplasmic staining. Eighteen of these 36 cases (50%) showed SP125 positivity in corresponding adenocarcinomas and 16 (44.4%) contained L858R mutation (Supporting information, Table S6). Five of the eight cases with homogeneous cytoplasmic staining of the non-neoplastic bronchi showed homogeneous staining of the adenocarcinoma. Moreover, the three L858R-negative adenocarcinomas with apical granular positivity also exhibited granular positivity within the non-neoplastic bronchial epithelium. The SP125+ bronchial epithelium in the 16 cases with L858R+ tumour was macrodissected and sequenced for L858R mutation (Figure 4A,B). Three of the 16 (18.8%) demonstrated L858R mutation in the non-neoplastic bronchial tissue. The pattern of SP125 staining in the non-neoplastic bronchial epithelium was not helpful in differentiating between L858R+/SP125+ versus L858R-/SP125+ non-neoplastic bronchial epithelium, as both showed a mixed homogeneous and granular pattern. A finely to coarsely granular cytoplasmic SP111 positivity was noted in non-neoplastic endothelium, smooth muscle, bronchial glands and chondrocytes in

834

G Allo et al.

L858R antibodies

B 1.0

0.8

0.8

0.6

0.4

0.6

0.4

0.2

0.2

SP121, AUC = 0.907

SP125, AUC = 0.861 42B3, AUC = 0.795

6B6, AUC = 0.897 0.0

0.0 0.0

0.2

0.4 0.6 False positive rate

0.8

1.0

E745-A750 EGFR deletion

C

0.0

D 1.0

0.8

0.8

0.6

0.4

0.2

0.4 0.6 False positive rate

0.8

1.0

Either 15-nt EGFR deletion

1.0

True positive rate

True positive rate

E746-A750 EGFR deletion

1.0

True positive rate

True positive rate

A

0.6

0.4

0.2

0.2

SP111, AUC = 0.914

SP111, AUC = 0.894

6B6, AUC = 0.933

6B6, AUC = 0.931

0.0

0.0 0.0

0.2

0.6 0.4 False positive rate

0.8

1.0

0.0

0.2

0.4 0.6 False positive rate

0.8

1.0

Figure 3. Receiver operating characteristic (ROC) curves, illustrating the performance of the immunohistochemical methods to detect EGFR protein with L858R substitution. (A) Staining with SP125 antibody using the OptiView detection method yielded the highest accuracy, represented by the area under the curve (AUC). (B–D) ROC curves, illustrating the performance of two immunohistochemical methods to detect EGFR proteins with exon-19, 15-nucleotide deletion. The two antibodies performed equivalently with regards to 15-nt E746-A750 deletion (B), the 15-nt K745-A750 deletion (C) and both deletions (D).

17 of 51 full sections (33.3%), three of which had SP111-positive adenocarcinoma (Figure 4C). Of these 17 cases, three (17.6%) carried SP111-positive carcinoma with the 15-nt exon-19 deletion (Supporting information, Table S6). The carcinomas in the remaining cases neither had this particular deletion nor were positive for SP111 IHC. There was no sig-

nificant correlation between the presence of this specific deletion and positivity in the non-neoplastic component (P = 0.071). Because the positivity in non-neoplastic tissue was mainly in pulmonary soft tissue and not in bronchial epithelium, the presence of exon-19 deletion in the non-neoplastic tissue was not investigated further. © 2013 John Wiley & Sons Ltd, Histopathology, 64, 826–839.

EGFR immunohistochemistry in lung cancer

835

Table 6. Correlations between immunohistochemistry using the exon-19 deletion mutation-specific antibodies using low threshold scoring, and the presence of 15-nucleotide exon-19 deletion variants E746-A750/K745-A750 del Present SP111

6B6

Absent

Total

Positive

15

1

16

Negative

3

222

225

Total

18

223

241

Positive

16

5

21

Negative

2

218

220

18

223

241

Total

Sensitivity

Specificity

PPV

NPV

83.3

99.6

93.8

98.7

88.9

97.8

76.2

99.1

PPV, positive predictive value; NPV, negative predictive value.

Table 7. Post-gefitinib adenocarcinoma cases with EGFR mutation and immunohistochemical staining results

ID

EGFR mutation

Antibody

Immunostaining result

1

L833V, H835L

SP125

Negative

2

L858R

SP125

Positive

3

del746-A750

SP111

Positive

4

delE746-S752, insV

SP111

Negative

5

delL747-P752

SP111

Negative

6

delL747-P753, insS

SP111

Negative

Discussion Testing for EGFR mutations in lung adenocarcinomas has been introduced into clinical practise due to the therapeutic implications that these mutations carry. Routine testing necessitates the development of an efficient, cost-effective and practical test that can be universally standardized and performed by pathology laboratories, either to triage specimens, to substitute for more expensive molecular tests, or to fast-track treatment decisions by a shorter diagnosis time. Immunohistochemistry has been used in routine clinical pathology practise based on its relatively high throughput and fast turnaround time, pathologists’ experience in rapidly interpreting the findings, and its affordability and universal availability. These advantages make IHC a useful test that can potentially be a cost-effective method to screen for or even diagnose EGFR mutations. In this report, we have detailed the © 2013 John Wiley & Sons Ltd, Histopathology, 64, 826–839.

performance of two novel antibodies targeting the most common EGFR mutations, L858R substitution (SP125) and E746-A750 deletion (SP111), and compared them to two previously reported antibodies.21 This study found that SP125 and SP111 EGFR antibodies are reasonably useful tools for rapid screening to detect the two most common forms of targetable EGFR mutations, SP125 detecting L858R mutations with a sensitivity of 76% and PPV of 73% and, SP111 detecting the 15 base-pair exon-19 deletions with a sensitivity of 83% and PPV of 94%. While their suboptimal sensitivity precludes use as a screening tool, these antibodies could potentially be used to overcome some of the hurdles that conventional mutation analysis methods face. These include minimal sample materials (e.g. core biopsies and fineneedle aspirates) and low tumour cellularity resulting from tumour contamination with non-neoplastic tissue (such as dense inflammation which can dilute the tumour DNA pool and result in a false-negative result). Immunohistochemistry can be used on small amounts of tissue and on both tissue biopsies and cytology-derived tissue blocks, permitting direct visualization of tumour cells and circumventing potential dilution by host non-tumour cell contamination. The high specificity and positive predictive value may also provide a rationale for their use in detecting the presence of the most common types of EGFR mutations – L858R and 15-nt E745/6-A750 deletions – while awaiting results from the conventional mutation studies that may take a longer time to complete. This is especially applicable to SP111, which carries a high specificity of 99.6% and a positive predictive value of 94%. If validated further in larger sample-size studies, SP111-positive tumours would very probably harbour

836

G Allo et al.

A

B

C Figure 4. EGFR mutation in normal bronchial epithelium revealed by the mutant-specific antibodies. (A) Non-neoplastic bronchial epithelium adjacent to EGFR L858R-carrying adenocarcinoma staining positively for SP125 immunohistochemistry (inset: high-power magnification). (B) The same area after coring for DNA extraction and mutation evaluation. This case showed EGFR L858R mutation in the shown non-neoplastic bronchial epithelium. (C) Non-neoplastic lung tissue, showing finely granular immunohistochemical staining in smooth muscle cells and chondrocytes using SP111 antibody.

the corresponding 15-nt exon-19 deletion. If both antibodies are used in conjunction, they would detect the most common EGFR mutations with a sensitivity of 82% and a positive predictive value of 82% (Supporting information, Table S7). However, this high positive predictive value comes with the caveat of a small, yet noteworthy, false-positive rate (six of 220, 2.7% for SP125 and one of 223, 0.5% for SP111; Tables 5 and 6), which may affect the utility of these antibodies, especially in small material such as biopsies. On further analysis and full-section examination, we found that false-positive cases are associated with 1+ staining, while all cases with 2+ staining were true-positive.

Further study may be needed to establish the extent of false positivity associated with these antibodies. In one tumour case showing discrepancy between IHC and molecular analysis, full-section staining and reanalysis revealed mutation heterogeneity. The question of whether or not heterogeneity of EGFR mutation status can occur in a tumour is controversial. While some reports suggest that EGFR mutation, being a driver mutation, occurs homogeneously within a primary tumour,28 others have reported the coexistence of EGFR wild-type and mutant tumour cells on analysing different areas of a single tumour after macrodissection or laser capture microdissec© 2013 John Wiley & Sons Ltd, Histopathology, 64, 826–839.

EGFR immunohistochemistry in lung cancer

tions.29,30 Furthermore, genetic heterogeneity in primary tumours has been inferred from reports of metastases showing discordant mutation status compared to the primary tumour.31,32 Further studies of this subject may be warranted. Using these mutant-specific antibodies, we have also inadvertently identified EGFR mutations in histologically normal bronchial epithelium. Sixteen cases with L858R mutations demonstrated SP125 positivity in adjacent non-neoplastic bronchial epithelium, with three (19%) being confirmed to have L858R mutations. These findings are consistent with a previous report of EGFR mutations in histologically normal epithelium in bronchi and bronchioles within the tumour (nine of 21 cases, 43%) and within adjacent lung tissue (seven of 29 cases, 24%) in EGFR mutation-carrying adenocarcinomas, suggesting that EGFR alterations are early events in lung adenocarcinoma tumorigenesis with a field effect.33 While the previous study reported that exon-19 deletions were more likely to be present in normal lung than L858R mutations (14 of 26, 54%; versus 2 of 28, 7%; P = 0.02),33 we found L858R mutations only in the normal lung tissue, as only SP125 stained the bronchioles. Conversely, SP111, the exon-19 deletion-specific antibody, did not stain bronchial epithelium and, therefore, deletions were not investigated further. The sensitivity and specificity of the previously reported antibodies (clones 43B2 and 6B6) has varied in different studies with varying study populations (Supporting information, Table S8).21,24,34–41 Nevertheless, the performance of these antibodies in this study resembled that in the previous reports (Supporting information, Figure S1). More importantly, the two new antibodies we studied have comparable performance to the previously reported EGFR mutation antibodies;21 SP125 achieved an equivalent performance to 43B2, with a higher sensitivity, comparable specificity and negative predictive value, but a lower positive predictive value. SP111 detected the 15nucleotide deletion with comparable sensitivity, specificity and negative predictive value to 6B6 and with a higher positive predictive value. One major caveat for the use of immunohistochemistry to detect EGFR mutations is that the currently available mutant-specific antibodies are specific only to their target mutations. Despite their reasonably high sensitivity and negative predictive value in detecting L858R and 15-nt exon-19 deletion, SP125 would miss six of 26 (23%) of any form of point mutations, including one case with combined L858R and exon-19 deletion. The SP111 would miss nine of 24 (38%) of the exon-19 deletions, belonging mainly © 2013 John Wiley & Sons Ltd, Histopathology, 64, 826–839.

837

to non-15 nucleotide deletions. Furthermore, the rigorous assessment with full-section staining of the false-positive and -negative cases may potentially overestimate the sensitivity and specificity of the studied antibodies. Our study demonstrates the importance of careful interpretation of antibody staining, as a substantial proportion of mutation-positive carcinomas showed weak (1+) staining intensity. While the ‘low-threshold’ scoring criterion, which defines a positive tumour as that showing any positivity, better predicts tumour positivity, the potential false positivity resulting from use of this threshold should be noted. The pattern of IHC positivity was also important; only homogeneous cytoplasmic staining correlated with the presence of the mutation of interest. Moreover, it is important to note the distribution of staining, as non-tumour tissue may stain positively with either antibody regardless of EGFR mutation status. This is particularly relevant in small biopsies. In summary, we have reported the performance of two new EGFR mutant-specific antibodies that could potentially be used clinically to overcome the challenges of mutation testing in small and low tumour cellularity specimens. More studies are needed to validate these results further in larger cohorts of patients and to evaluate their utility in treatment decisions.

Conflicts of interests Dr Tsao receives a research grant from Ventana ME, Inc.

Acknowledgements This work is supported by grants from the Canadian Cancer Society Research Institute (#020527), Ontario Institute of Cancer Research, Terry Fox Foundation STIHR at CIHR grant TGT-53912 (G.A., B.B., N.Y., M.S.T.) and Ontario Ministry of Health and Long Term Care. Dr Tsao is the M. Qasim Choksi Chair in Lung Cancer Translational Research. We would like to thank Ms Michele Astry for assay development and validation of Ventana clones, Ms Krista Costa for TMA staining for CST clones and Dr Leigh Ann Henricksen for bridging study over lots and fullsection staining.

References 1. Ferlay J, Shin HR, Bray F, Forman D, Mathers C, Parkin DM. GLOBOCAN 2008 v1.2: cancer incidence and mortality worldwide:

838

2.

3. 4. 5.

6.

7.

8.

9.

10.

11.

12.

13.

14.

15.

16.

17.

18.

G Allo et al.

IARC CancerBase no. 10 [internet]. Lyon, France: International Agency for Research on Cancer, 2010. Available at: http:// globocan.iarc.fr. (accessed 10 April 2012). Travis WD, World Health Organization, International Agency for Research on Cancer et al. Pathology and genetics of tumours of the lung, pleura, thymus and heart. Lyon, France: IARC Press, 2004. Subramanian J, Govindan R. Lung cancer in never smokers: a review. J. Clin. Oncol. 2007; 25; 561–570. da Cunha Santos G, Shepherd FA, Tsao MS. EGFR mutations and lung cancer. Ann. Rev. Pathol. 2011; 15; 49–69. Shepherd FA, Rodrigues Pereira J, Ciuleanu T et al. Erlotinib in previously treated non-small-cell lung cancer. N. Engl. J. Med. 2005; 353; 123–132. Jackman DM, Miller VA, Cioffredi LA et al. Impact of epidermal growth factor receptor and KRAS mutations on clinical outcomes in previously untreated non-small cell lung cancer patients: results of an online tumor registry of clinical trials. Clin. Cancer Res. 2009; 15; 5267–5273. Lara-Guerra H, Waddell TK, Salvarrey MA et al. Phase II study of preoperative gefitinib in clinical stage I non-small-cell lung cancer. J. Clin. Oncol. 2009; 27; 6229–6236. Lynch TJ, Bell DW, Sordella R et al. Activating mutations in the epidermal growth factor receptor underlying responsiveness of non-small-cell lung cancer to gefitinib. N. Engl. J. Med. 2004; 350; 2129–2139. Paez JG, Janne PA, Lee JC et al. EGFR mutations in lung cancer: correlation with clinical response to gefitinib therapy. Science 2004; 304; 1497–1500. Pao W, Miller V, Zakowski M et al. EGF receptor gene mutations are common in lung cancers from ‘never smokers’ and are associated with sensitivity of tumors to gefitinib and erlotinib. Proc. Natl Acad. Sci. U.S.A. 2004; 101; 13306–13311. Tsao MS, Sakurada A, Cutz JC et al. Erlotinib in lung cancer – molecular and clinical predictors of outcome. N. Engl. J. Med. 2005; 353; 133–144. Zhu CQ, da Cunha Santos G, Ding K et al. Role of KRAS and EGFR as biomarkers of response to erlotinib in National Cancer Institute of Canada Clinical Trials Group Study BR.21. J. Clin. Oncol. 2008; 26; 4268–4275. Mok TS, Wu YL, Thongprasert S et al. Gefitinib or carboplatinpaclitaxel in pulmonary adenocarcinoma. N. Engl. J. Med. 2009; 361; 947–957. Sharma SV, Bell DW, Settleman J, Haber DA. Epidermal growth factor receptor mutations in lung cancer. Nat. Rev. Cancer 2007; 7; 169–181. Sakurada A, Shepherd FA, Tsao MS. Epidermal growth factor receptor tyrosine kinase inhibitors in lung cancer: impact of primary or secondary mutations. Clin. Lung Cancer 2006; 7 (Suppl. 4); S138–S144. Yamamoto H, Toyooka S, Mitsudomi T. Impact of EGFR mutation analysis in non-small cell lung cancer. Lung Cancer 2009; 63; 315–321. Cheng L, Alexander RE, Maclennan GT et al. Molecular pathology of lung cancer: key to personalized medicine. Mod. Pathol. 2012; 25; 347–369. Canadian Agency for Drugs and Technologies in Health (CADTH). Epidermal growth factor receptor mutation analysis in advanced non-small cell lung cancer: a review of the clinical effectiveness and guidelines. Ottawa, ON: Canadian Agency for Drugs and Technologies in Health, 2010. Available at: http:// www.cadth.ca/media/pdf/M0017_EGFR_Testing_for_NSCLC_e. pdf (accessed 10 April 2012).

19. Keedy VL, Temin S, Somerfield MR et al. American Society of Clinical Oncology provisional clinical opinion: epidermal growth factor receptor (EGFR) mutation testing for patients with advanced non-small-cell lung cancer considering first-line EGFR tyrosine kinase inhibitor therapy. J. Clin. Oncol. 2011; 29; 2121–2127. 20. Ellis PM, Blais N, Soulieres D et al. A systematic review and Canadian consensus recommendations on the use of biomarkers in the treatment of non-small cell lung cancer. J. Thorac. Oncol. 2011; 6; 1379–1391. 21. Yu J, Kane S, Wu J et al. Mutation-specific antibodies for the detection of EGFR mutations in non-small-cell lung cancer. Clin. Cancer Res. 2009; 15; 3023–3028. 22. Thomas RK, Baker AC, Debiasi RM et al. High-throughput oncogene mutation profiling in human cancer. Nat. Genet. 2007; 39; 347–351. 23. Pan Q, Pao W, Ladanyi M. Rapid polymerase chain reactionbased detection of epidermal growth factor receptor gene mutations in lung adenocarcinomas. J. Mol. Diagn. 2005; 7; 396–403. 24. Brevet M, Arcila M, Ladanyi M. Assessment of EGFR mutation status in lung adenocarcinoma by immunohistochemistry using antibodies specific to the two major forms of mutant EGFR. J. Mol. Diagn. 2010; 12; 169–176. 25. Lara-Guerra H, Chung CT, Schwock J et al. Histopathological and immunohistochemical features associated with clinical response to neoadjuvant gefitinib therapy in early stage nonsmall cell lung cancer. Lung Cancer 2012; 76; 235–241. 26. Byrt T, Bishop J, Carlin JB. Bias, prevalence and kappa. J. Clin. Epidemiol. 1993; 46; 423–429. 27. DeLong ER, DeLong DM, Clarke-Pearson DL. Comparing the areas under two or more correlated receiver operating characteristic curves: a nonparametric approach. Biometrics 1988; 44; 837–845. 28. Yatabe Y, Matsuo K, Mitsudomi T. Heterogeneous distribution of EGFR mutations is extremely rare in lung adenocarcinoma. J. Clin. Oncol. 2011; 29; 2972–2977. 29. Sakurada A, Lara-Guerra H, Liu N, Shepherd FA, Tsao MS. Tissue heterogeneity of EGFR mutation in lung adenocarcinoma. J. Thorac. Oncol. 2008; 3; 527–529. 30. Nakano H, Soda H, Takasu M et al. Heterogeneity of epidermal growth factor receptor mutations within a mixed adenocarcinoma lung nodule. Lung Cancer 2008; 60; 136–140. 31. Gow CH, Chang YL, Hsu YC et al. Comparison of epidermal growth factor receptor mutations between primary and corresponding metastatic tumors in tyrosine kinase inhibitor-naive non-small-cell lung cancer. Ann. Oncol. 2009; 20; 696–702. 32. Park S, Holmes-Tisch AJ, Cho EY et al. Discordance of molecular biomarkers associated with epidermal growth factor receptor pathway between primary tumors and lymph node metastasis in non-small cell lung cancer. J. Thorac. Oncol. 2009; 4; 809–815. 33. Tang X, Shigematsu H, Bekele BN et al. EGFR tyrosine kinase domain mutations are detected in histologically normal respiratory epithelium in lung cancer patients. Cancer Res. 2005; 65; 7568–7572. 34. Ho HL, Chang FP, Ma HH et al. Molecular diagnostic algorithm for epidermal growth factor receptor mutation detection in Asian lung adenocarcinomas: comprehensive analyses of 445 Taiwanese cases with immunohistochemistry, PCR-direct sequencing and Scorpion/ARMS methods. Respirology 2013; 18; 1261–1270. 35. Cooper WA, Yu B, Yip PY et al. EGFR mutant-specific immunohistochemistry has high specificity and sensitivity for detecting © 2013 John Wiley & Sons Ltd, Histopathology, 64, 826–839.

EGFR immunohistochemistry in lung cancer

36.

37.

38.

39.

40.

41.

targeted activating EGFR mutations in lung adenocarcinoma. J. Clin. Pathol. 2013; 66; 744–748. Fan X, Liu B, Xu H et al. Immunostaining with EGFR mutation-specific antibodies: a reliable screening method for lung adenocarcinomas harboring EGFR mutation in biopsy and resection samples. Hum. Pathol. 2013; 44; 1499–1507. Xiong Y, Bai Y, Leong N et al. Immunohistochemical detection of mutations in the epidermal growth factor receptor gene in lung adenocarcinomas using mutation-specific antibodies. Diagn. Pathol. 2013; 8; 27. Kozu Y, Tsuta K, Kohno T et al. The usefulness of mutationspecific antibodies in detecting epidermal growth factor receptor mutations and in predicting response to tyrosine kinase inhibitor therapy in lung adenocarcinoma. Lung Cancer 2011; 73; 45–50. Simonetti S, Molina MA, Queralt C et al. Detection of EGFR mutations with mutation-specific antibodies in stage IV nonsmall-cell lung cancer. J. Transl. Med. 2010; 8; 135. Kato Y, Peled N, Wynes MW et al. Novel epidermal growth factor receptor mutation-specific antibodies for non-small cell lung cancer: immunohistochemistry as a possible screening method for epidermal growth factor receptor mutations. J. Thorac. Oncol. 2010; 5; 1551–1558. Kitamura A, Hosoda W, Sasaki E, Mitsudomi T, Yatabe Y. Immunohistochemical detection of EGFR mutation using mutation-specific antibodies in lung cancer. Clin. Cancer Res. 2010; 16; 3349–3355.

Supporting Information Additional supporting information may be found in the online version of this article: Table S1. Characteristics of antibodies used for immunohistochemistry

© 2013 John Wiley & Sons Ltd, Histopathology, 64, 826–839.

839

Table S2. Results of immunohistochemistry using the Sp125 and SP111 EGFR mutation-specific antibodies among the two observers Table S3. Full section staining with the four EGFR mutation specific antibodies Table S4. Correlations between presence of EGFR L858R mutation and mutation-specific antibody IHC using high-threshold scoring Table S5. Correlation between the presence of 15-nucleotide exon-19 deletion variants and immunohistochemistry of the exon-19 deletion mutationspecific antibodies Table S6. Cases with positive immunohistochemistry in non-neoplastic component: Mutation status and pattern of staining in associated tumour Table S7. The predictive value of combining EGFR mutation-specific antibodies in detecting the common EGFR mutation (L858R mutation and the 15-nucleotide exon 19 deletion) Table S8. Reports of EGFR-specific immunohistochemistry using 43B2 and 6B6 clones targeting EGFR L858R point mutation, and EGFR exon 19 deletion, respectively. Figure S1. Forest plots summarizing previously reported antibodies targeting EGFR L858R mutation (clone 43B2) and 15-nucleotide exon-19 deletin (clone 6B6) in comparison to the current study.