Assessment Run B23 2017

HER-2 IHC

Material
The slide to be stained for HER-2 comprised the following 9 materials:
IHC: HER-2
FISH: HER-2/chr17
Score*
ratio**
(0, 1+, 2+, 3+)
1. Cell line 1, Horizon Discovery*** 3+
2. Cell line 2, Horizon Discovery*** 2+
3. Cell line 3, Horizon Discovery*** 1+
4. Cell line 4, Horizon Discovery*** 0
5. Breast carcinoma, no. 1 3+ > 6.0 (clusters) (amplified)
6. Breast carcinoma, no. 2 2+ 2.3 – 2.9 (amplified)
7. Breast carcinoma, no. 3 0-1+ 1.1 – 1.4 (unamplified)
8. Breast carcinoma, no. 4 1-2+ 1.3 – 1.7 (unamplified)
9. Breast carcinoma, no. 5 0-1+ 1.2 – 1.4 (unamplified)
* HER-2 immunohistochemical score (see table below) as achieved by using the three FDA approved kits and
antibodies, HercepTest™ (Dako), Oracle™ (Leica) and PATHWAY® (Ventana), in NordiQC reference laboratories.
** HER-2/chr17 ratios achieved using ZytoLight ® SPEC HER2/CEN 17 Dual Color FISH (Zytovision)
*** The cell lines were not included in the assessment. Data will be analyzed subsequently by digital image analysis.

All carcinomas were fixed for 24 - 48 h in 10% neutral buffered formalin.

IHC scoring system according to the 2013 ASCO/CAP guidelines

Score 0 No staining is observed or incomplete membrane staining is observed in ≤ 10% of the tumour cells.

Score 1+ A faint perceptible and incomplete membrane staining is observed in more than 10% of the tumour
cells.

A weak to moderate circumferential incomplete membrane staining is observed in more than 10% of
Score 2+ the tumour cells or an intense circumferential complete membranous staining in ≤ 10% of the tumour
cells.

An intense circumferential complete membrane staining is observed in more than 10% of the tumour
Score 3+
cells.

Criteria for assessing a HER-2 staining as optimal were:

 Staining corresponding to score 0 or 1+ in carcinomas no. 3 and 5.

 Staining corresponding to score 0, 1+ or 2+ in carcinoma no. 4.
 Staining corresponding to score 2+ or 3+ in carcinoma no. 2.
 Staining corresponding to score 3+ in carcinoma no. 1.
 No or only weak cytoplasmic reaction that did not interfere with the interpretation.

Staining was assessed as good, if (1) the HER-2 gene amplified tumour no. 1 showed a 2+ reaction and
the other breast carcinomas showed reaction pattern as described above (equivocal 2+ IHC staining
should always be analyzed by ISH according to the ASCO/CAP guidelines) or (2) the HER-2 gene non-
amplified tumour no. 3 and/or 5 showed a 2+ reaction and the other breast carcinomas showed the
expected reaction pattern.

Staining was assessed as borderline, if the signal-to-noise ratio was low, e.g., because of moderate
cytoplasmic reaction, excessive counterstaining or excessive retrieval hampering the interpretation.

Staining was assessed as poor in case of a false negative staining (e.g., the 3+ tumour or the 2+ tumour
with gene amplification showed a 0 or 1+ reaction) or a false positive staining (e.g., the 0/1+ tumors and
the 2+ tumour without gene amplification showing a 3+ reaction).

Participation
Number of laboratories registered for HER-2, run B23 380
Number of laboratories returning slides 372 (98%)
Nordic Immunohistochemical Quality Control, HER-2 IHC run B23 2017 Page 1 of 7
Results: 372 laboratories participated in this assessment and 95% achieved a sufficient mark.
Assessment marks for IHC HER-2 assays and HER-2 antibodies are summarised in table 1.

Table 1. Assessment marks for IHC assays and antibodies run B23, HER-2 IHC
FDA approved HER-2 Suff.
n Vendor Optimal Good Borderline Poor Suff.1
assays OPS2
PATHWAY® rmAb clone
204 Ventana/Roche 186 13 2 3 98% 98%
4B5, 790-2991
CONFIRM™, rmAb clone
15 Ventana/Roche 12 3 0 0 100% 100%
4B5, 790-4493
HercepTest™ SK001 33 Dako/Agilent 32 0 0 1 97% 97%
HercepTest™ SK001 4
8 Dako/Agilent 8 0 0 0 100% -
HercepTest™ K5207 2 Dako/Agilent 2 0 0 0 - -
HercepTest™ K5204 2 Dako/Agilent 1 1 0 0 - -
Oracle™ mAb clone
9 Leica 4 4 0 1 89% 88%
CB11, TA9145
Antibodies3 for
laboratory developed Suff.
n Vendor Optimal Good Borderline Poor Suff.1
HER-2 assays, OPS2
conc. antibody
mAb clone BS24 1 Nordic Biosite 0 1 0 0 - -
7 Leica/Novocastra
mAb clone CB11 4 4 0 0 100% 100%
1 Biogenex
1 Biocare
1 Cell Marque
rmAb clone EP3 1 Celnovte 3 2 0 0 100% 100%
1 Thermo/NeoMarkers
1 PathnSitu
17 Thermo/NeoMarkers
3 Zytomed
rmAb clone SP3 1 Cell Marque 12 8 0 3 87% 100%
1 Spring Bioscience
1 Thermo/Pierce
pAb clone A0485 50 Dako/Agilent 35 11 1 3 92% 93%
Antibodies for
Suff.
laboratory developed n Vendor Optimal Good Borderline Poor Suff.1
OPS2
HER-2 assays, RTU
mAb clone CB11,
4 Leica/Novocastra 0 0 3 1 - -
NCL-L-CB11
mAb clone CB11,
1 Zytomed 0 0 1 0 - -
BMS014
rmAb clone EP3,
1 Biogenex 1 0 0 0 - -
AN726
rmAb clone EP3,
1 Diagnostic Biosystems 1 0 0 0 - -
RMPD049R
rmAb clone SP3,
2 Cell Marque 1 1 0 0 - -
237R
rmAb clone SP3,
1 Master Diagnostics 1 0 0 0 - -
MAD-000308QD
Ab clone MXR001,
2 Maixin 1 1 0 0 - -
RMA-0701
Total 372 304 49 7 12 - -
Proportion 82% 13% 2% 3% 95% -
1) Proportion of sufficient stains (optimal or good).
2) Proportion of sufficient stains with optimal protocol settings only, see below.
3) mAb: mouse monoclonal antibody, rmAb: rabbit monoclonal antibody, pAb: polyclonal antibody.
4) RTU system developed for the Agilent/Dako`s semi-automated systems (Autostainer Link48) but used by laboratories on different
platforms (Leica Bond and Dako Omnis)

Nordic Immunohistochemical Quality Control, HER-2 IHC run B23 2017 Page 2 of 7
Detailed Analysis
FDA/CE IVD approved assays

PATHWAY® rmAb clone 4B5 (790-2991, Ventana): 186 of 204 (91%) protocols were assessed as
optimal. Protocols with optimal results were typically based on heat induced epitope retrieval (HIER) in Cell
Conditioning 1 (CC1) (efficient heating time 8-64 min.) in BenchMark XT, GX or Ultra, 8 – 60 min.
incubation of the primary Ab and Iview, UltraView or OptiView as detection kit. Using these protocol
settings, 199 of 203 (98%) laboratories produced a sufficient staining result (optimal or good).

CONFIRM™ rmAb clone 4B5 (790-4493, Ventana): 12 of 15 (80%) protocols were assessed as optimal.
Protocols with optimal results were typically based on HIER in CC1 (efficient heating time 30-64 min.) in
BenchMark XT, GX or Ultra, 16 – 40 min. incubation of the primary Ab and Iview or UltraView as detection
kit. Using these protocol settings, 15 of 15 (100%) laboratories produced a sufficient staining result.

HercepTest™ pAb (SK001, Dako): 40 of 41 (98%) protocols were assessed as optimal. Protocols with
optimal results were typically based on HIER in HercepTest™ epitope retrieval solution at 97 - 99°C for 40
min. in a water bath or PT Link and 30 min. incubation of the primary Ab. Using these protocol settings, 28
of 29 (97%) laboratories produced a sufficient staining result.

HercepTest™ pAb (K5207, Dako): 2 of 2 protocols were assessed as optimal. Protocols with optimal
result were based on HIER in HercepTest™ epitope retrieval solution at 97-98°C for 40 min. in a water
bath or PT link and 30 min. incubation of the primary Ab. Using these protocol settings, 2 of 2 laboratories
produced a sufficient staining result.

Oracle™ mAb clone CB11 (TA9145, Leica): 4 of 9 (44%) protocols were assessed as optimal. Three
protocols were based on HIER in Bond Epitope Retrieval Solution 1 for 25-30 min., and 30 min., incubation
of the primary Ab. One protocol was based on HIER in Bond Epitope Retrieval Solution 2 for 15 min., and
15 min., incubation of the primary Ab. Using these protocol settings, 7 of 8 (88%) laboratories produced a
sufficient staining result.

Concentrated antibodies for laboratory developed (LD) assays

mAb CB11: 4 of 8 (50%) protocols were assessed as optimal. Optimal protocols were based on HIER
using Target Retrieval Solution (TRS) Low (Dako) (1/1)*, Bond Epitope Retrieval Solution 1 pH 6 (BERS1,
Leica) (1/1), Tris-EDTA/EGTA pH 9 (1/1) or Citrate pH 6 (1/2). The mAb clone CB11 was diluted in a range
of 1:150-400 depending on the total sensitivity of the protocol employed. Using these protocol settings, 5
of 5 (100%) laboratories produced a sufficient staining result (optimal or good).
* (number of optimal results/number of laboratories using this HIER buffer)

rmAb EP3: 3 of 5 (60%) protocols were assessed as optimal. Optimal protocols were based on HIER using
TRS pH 9 (3-in-1) (Dako) (1/1), BERS2 (Bond, Leica) (1/1) or Tris-EDTA/EGTA pH 9 (1/2). The rmAb
clone EP3 was diluted in the range of 1:70-200 depending on the total sensitivity of the protocol
employed. Using these protocol settings, 4 of 4 (100%) laboratories produced a sufficient staining result.

rmAb SP3: 12 of 23 (52%) protocols were assessed as optimal. Optimal protocols were based on HIER
using TRS pH 9 (3-in-1) (Dako) (1/3), BERS2 (Bond, Leica) (9/10), Tris-EDTA pH 9 (1/2) or Citrate pH 6
(1/2). The rmAb clone SP3 was diluted 1:50-250 depending on the total sensitivity of the protocol
employed. Using this protocol setting, 17 of 17 (100%) laboratories produced a sufficient staining result.

pAb A0485: 35 of 50 (70%) protocols were assessed as optimal. Optimal protocols were based on HIER
using either TRS low pH 6.1 (Dako) (20/29), TRS pH 9 (3-in-1) (Dako) (6/8), CC1 (BenchMark, Ventana)
(2/4), BERS1 (Bond, Leica) (3/3), Citrate pH 6 (2/3) or unknown (2/2). The pAb A0485 was typically
diluted in the range of 1:100-1,500 depending on the total sensitivity of the protocol employed. Using
these protocol settings, 43 of 46 (93%) laboratories produced a sufficient staining result.

Comments
In this assessment and in concordance with the previous NordiQC assessments of HER-2 IHC, insufficient
HER-2 staining results were characterised by too weak or false negative staining reactions. This was
particularly and most critically observed as 0/1+ IHC reaction in the low-level HER-2 gene amplified breast
carcinoma no. 2. This tumour was categorized as IHC 2+ in the NordiQC reference laboratories using three
FDA/CE-IVD HER-2 IHC assays: PATHWAY® (Ventana), HercepTest™ (Dako) and Oracle™ (Leica) and
showed a low level of HER-2 gene amplification (ratio 2.3 – 2.9) by ISH. False negative staining reaction of
the breast carcinoma no. 2 was seen in 58% of the insufficient results (11 of 19).
Nordic Immunohistochemical Quality Control, HER-2 IHC run B23 2017 Page 3 of 7
The remaining insufficient results were typically characterised by a poor signal-to-noise ratio, impaired
morphology, excessive counterstaining complicating interpretation or false positive 3+ IHC staining in the
HER-2 non-amplified tumours.
False negative results were seen both in laboratory developed (LD) and FDA-/CE-IVD approved assays,
while false positive results only were seen in LD assays.
False negative results were for the LD assays typically caused by too low sensitivity of the protocol applied
(e.g. too low concentration of the primary Ab, too short incubation time of the primary Ab and/or
insufficient HIER). For the FDA-/CE-IVD approved systems, used according to the official package inserts
for the respective systems, no single cause for insufficient and false negative staining reactions could be
identified from the protocols submitted. However, compared to previous assessments an extended use of
off-label protocols for the approved systems were applied by the participants. Three laboratories used the
Dako HercepTest™ SK001 assay on the Bond IHC platform, while 5 laboratories performed the assay
manually and as such not within the intended use and must consequently be considered as a LD assay. In
this assessment, all 8 laboratories using HercepTest™ SK001 on either Bond or manually produced a result
assessed as optimal. However, despite the encouraging results, off-label use must be carefully validated
by the end-users on a large cohort of breast carcinomas (n=100, ASCO/CAP 2013 guidelines).

The Ventana PATHWAY® /CONFIRM™ HER-2 IHC assay was also increasingly used off-label by the
participants, applying OptiView as detection system and not UltraView or iView as recommended by
Ventana. In this assessment no impact on the analytical sensitivity and specificity was revealed. In
contrast, internal studies previously performed in the NordiQC reference laboratory indicated a less precise
and robust HER-2 IHC assay if UltraView was substituted by OptiView PATHWAY® /CONFIRM™. OptiView
will typically amplify the analytical sensitivity of the IHC system 3-4x compared to UltraView.
Consequently if OptiView is applied, the HER-2 IHC assay must be adjusted at other parameters e.g
incubation time or the primary Ab, HIER settings to provide the analytical sensitivity level validated by
Ventana, which as mentioned can cause a less precise and robust assay.

In this assessment, the FDA-/CE-IVD approved HER-2 IHC assays PATHWAY® /CONFIRM™ and
HercepTest™ from Ventana and Dako respectively were most successful and provided a higher pass-rate
superior to LD assays as illustrated in Fig. 1. PATHWAY®/CONFIRM™ IHC assays have provided a
consistent high pass rate throughout all HER-2 IHC runs in NordiQC.

The proportion of laboratories using FDA-/CE-IVD approved HER-2 IHC assays and LD assays is very
consistent. In this run, 27% of the participants (n=99) used LD assays compared to the range of 23 - 31%
in the last 11 assessments. Despite an overall improvement of the pass rate for LD HER-2 assays from run
B1 to B23, the pass rate and proportion of optimal results still is inferior to the FDA/CE-IVD approved
systems as PATHWAY® /CONFIRM™ and HercepTest™. In general, the three FDA-/CE-IVD approved HER-2
assays provided a high proportion of optimal results (89%; 237 of 265), whereas only 63% of LD HER-2
assays were assessed as optimal (67 of 107). As shown in Fig. 2, LD HER-2 assays provided a reduced
proportion of sufficient results but also demonstrates a shift towards lower assessment score (optimal to
good), typically caused by 2+ staining reaction in one or both of the HER-2 non-amplified tumours (no. 2
and 5) expected to show a 0/1+ staining reaction. The staining reaction of 2+ in these tumours would not
directly lead to a wrong diagnosis but require an additional ISH test due to the less precise IHC result.

The overall pass rate of 95% obtained in this assessment is very encouraging and is largely comparable to
the pass rates seen in the last 5 runs indicating a relatively stable level has been reached. A significant
improvement compared to the pass rate of 51% seen in run B1, 2006 has been obtained and maintained
at least for the five latest runs.

Nordic Immunohistochemical Quality Control, HER-2 IHC run B23 2017 Page 4 of 7
Figure 1. Pass rates of 23 HER-2 IHC assessments in the NordiQC breast cancer module

100

90

80

70
HercepTest™
60
PATHWAY®
50
Oracle™
40 Laboratory
Dev.
30

20

10

0
B1 B3 B5 B7 B9 B11 B13 B15 B17 B19 B21 B23

Figure 2. Proportion of assessment marks using FDA-/CD-IVD and LD assays

100% 5
2 7
21
90% 5

80%
28
70%

60%
Poor
50% Borderline
237
Good
40%
Optimal

30% 67

20%

10%

0%
FDA / CE-IVD HER2 IHC assays; LD HER2 assays
PATHWAY®, HercepTest™, Oracle™

Nordic Immunohistochemical Quality Control, HER-2 IHC run B23 2017 Page 5 of 7
Four different cell lines (Horizon Discovery, Cambridge UK) were included in this assessment to evaluate if
this material in combination with digital image analysis can be used to evaluate the accuracy and
interlaboratory reproducibility of HER-2 IHC assays and potentially function as standard reference material
for both EQA and internal quality control for HER-2 IHC assays. Subsequent analysis will be performed by
NordiQC and published at a later stage.

Scoring consensus B23

Laboratories were requested to submit scores (0, 1+, 2+, 3+) of their own HER-2 stained slides. This was
done by 86% (320 of 372) of the participants. For 284 of the 320 (89%) responding participants, scores
for all the tissues in the multi-tissue sections were in concordance with the NordiQC assessor group using
the ASCO/CAP 2013 interpretation guidelines. This was at the same level as in run B22, where 86% of the
scores were in consensus with the NordiQC assessor group. Among laboratories with sufficient staining,
96% (274 of 284) of interpretations were in agreement with the NordiQC assessors. Interpretation in
concordance with the NordiQC assessor group was seen in 78% (28 of 36) among participants with
insufficient staining. Typically, breast carcinoma no. 2 was interpreted as 2+ by the laboratories, but 0-1+
by the NordiQC assessor group.

Conclusion
The FDA-/CE-IVD approved HER-2 IHC assays PATHWAY®/CONFIRM™ (Ventana) and HercepTest™
(Dako) were in this assessment the most precise assays for the semi-quantitative IHC determination of
HER-2 protein expression. Laboratory developed assays produced a lower pass-rate and were less precise
for the HER-2 status requiring an additional ISH test for final evaluation.
Inclusion of 2+ tumours with and without HER-2 gene amplification in the control material for both EQA
and internal quality control is essential to evaluate precision and performance stability of the IHC HER-2
assays used by laboratories.

Figs 1a and 1b – optimal staining results, same protocol

Figs 2a and 2b – insufficient staining results - false negative, same protocol
Figs 3a and 3b – insufficient staining results – false positive, same protocol

Fig 1a. Fig 1b.

Left: Optimal staining result for HER-2 of the breast Left: Optimal staining result for HER-2 of the breast
ductal carcinoma no. 1 with a ratio of HER-2 / chr17 of > ductal carcinoma no. 4 with a ratio of HER-2 / chr17 of
6.0. 1.3 – 1.7.
> 10% of the neoplastic cells show an intense and > 10% of the neoplastic cells show a weak to moderate
complete membranous staining reaction corresponding to membranous staining reaction corresponding to 2+.
3+.
Right: Optimal staining result for HER-2 of the breast
Right: Optimal staining result for HER-2 of the breast ductal carcinoma no. 5 with a HER-2 / chr17 ratio of 1.2–
ductal carcinoma no. 2 with a ratio of HER-2 / chr17 of 1.4.
2.3 – 2.9. > 10% of the neoplastic cells show a faint membranous
> 10% of the neoplastic cells show a weak to moderate staining reaction corresponding to 1+.
and complete membranous staining reaction
corresponding to 2+.

Nordic Immunohistochemical Quality Control, HER-2 IHC run B23 2017 Page 6 of 7
 Fig 2a. Fig 2b.
Left: Staining result for HER-2 of the breast ductal Left: Staining result for HER-2 of the breast ductal
carcinoma no. 1 with a ratio of HER-2 / chr17 of > 6.0. carcinoma no. 4 with a ratio of HER-2 / chr17 of 1.3 –
> 10% of the neoplastic cells show a strong and 1.7.
complete membranous staining reaction corresponding to > 10% of the neoplastic cells show a faint perceptible
3+. membranous staining reaction corresponding to 1+.

Right: Insufficient and false negative staining result for Right: Staining result for HER-2 of the breast ductal
HER-2 of the breast ductal carcinoma no. 2 with a ratio carcinoma no. 5 with a HER-2 / chr17 ratio of 1.2 – 1.4.
of HER-2 / chr17 of 2.3 – 2.9. < 10% of the neoplastic cells show a faint membranous
> 10% of the neoplastic cells show a faint perceptible staining reaction corresponding to 0.
membranous staining reaction corresponding to 1+, but
does not meet the criteria to be classified as 2+ and will
not be referred to ISH.

Fig 3a. Fig 3b.

Left: Staining result for HER-2 of the breast ductal Left: Insufficient and false positive staining result for
carcinoma no. 1 with a ratio of HER-2 / chr17 of > 6.0. HER-2 of the breast ductal carcinoma no. 4 with a ratio
> 10% of the neoplastic cells show an intense and of HER-2 / chr17 of 1.3 – 1.7.
complete membranous staining reaction corresponding to > 10% of the neoplastic cells show an intense and
3+. complete membranous staining reaction corresponding to
3+.
Right: Staining result for HER-2 of the breast ductal
carcinoma no. 2 with a ratio of HER-2 / chr17 of 2.3 – Right: Staining result for HER-2 of the breast ductal
2.9. carcinoma no. 5 with a HER-2 / chr17 ratio of 1.2 – 1.4.
> 10% of the neoplastic cells show a strong and > 10% of the neoplastic cells show a moderate
complete membranous staining reaction corresponding to membranous staining reaction corresponding to 2+.
3+.

SN/LE/RR 13.04.17

Nordic Immunohistochemical Quality Control, HER-2 IHC run B23 2017 Page 7 of 7