Leila Russo

High concordance of protein (by IHC), gene (by FISH; HER2 only), and microarray readout (by TargetPrint) of ER, PgR, and HER2: results from the EORTC 10041/BIG 03-04 MINDACT trial

BACKGROUND To investigate the correlation of TargetPrint with local and central immunohistochemistry/fluorescence in situ hybridization assessment of estrogen (ER), progesterone (PgR), and human epidermal growth factor receptor 2 (HER2) in the first 800 patients enrolled in the MINDACT trial. PATIENTS AND METHODS Data from local (N = 800) and central (N = 626) assessments of receptor status were collected and compared with TargetPrint results. RESULTS For ER, the positive agreement (the percentage of central pathology positive assessments that were also TargetPrint/local laboratory positive) for TargetPrint in comparison to centralized assessment was 98% with a negative agreement (the percentage of central pathology negative assessments that were also TargetPrint/local laboratory negative) of 96%. For PgR, the positive agreement was 83% with a negative agreement of 92%. For HER2, the positive agreement was 75% with a negative agreement of 99%. Even though the local assessment showed higher positive agreement for PgR (89%) and higher positive agreement for HER2 (85%), the range of discordant local versus central assessments were as high as 6.7% for ER, 12.9% for PgR, and 4.3% for HER2. CONCLUSION TargetPrint and local assessment of ER, PgR, and HER2 show high concordance with central assessment in the first 800 MINDACT patients. However, there are concerns about the higher discordance rates for some local sites. TargetPrint can improve the reliability of hormone receptor and HER2 testing for those centers with a lower rate of concordance with the reference laboratory, with the limitation of a positive agreement of 75% for HER2. TargetPrint consequently has important implications for treatment decisions in clinical practice and is a reliable alternative to local assessment for ER. CLINICAL TRIALS NUMBER NCT00433589.BACKGROUND To investigate the correlation of TargetPrint with local and central immunohistochemistry/fluorescence in situ hybridization assessment of estrogen (ER), progesterone (PgR), and human epidermal growth factor receptor 2 (HER2) in the first 800 patients enrolled in the MINDACT trial. PATIENTS AND METHODS Data from local (N = 800) and central (N = 626) assessments of receptor status were collected and compared with TargetPrint results. RESULTS For ER, the positive agreement (the percentage of central pathology positive assessments that were also TargetPrint/local laboratory positive) for TargetPrint in comparison to centralized assessment was 98% with a negative agreement (the percentage of central pathology negative assessments that were also TargetPrint/local laboratory negative) of 96%. For PgR, the positive agreement was 83% with a negative agreement of 92%. For HER2, the positive agreement was 75% with a negative agreement of 99%. Even though the local assessment showed higher positive agreement for PgR (89%) and higher positive agreement for HER2 (85%), the range of discordant local versus central assessments were as high as 6.7% for ER, 12.9% for PgR, and 4.3% for HER2. CONCLUSION TargetPrint and local assessment of ER, PgR, and HER2 show high concordance with central assessment in the first 800 MINDACT patients. However, there are concerns about the higher discordance rates for some local sites. TargetPrint can improve the reliability of hormone receptor and HER2 testing for those centers with a lower rate of concordance with the reference laboratory, with the limitation of a positive agreement of 75% for HER2. TargetPrint consequently has important implications for treatment decisions in clinical practice and is a reliable alternative to local assessment for ER. CLINICAL TRIALS NUMBER NCT00433589.

Highly reproducible results of breast cancer biomarkers when analysed in accordance with national guidelines-a Swedish survey with central re-assessment

Abstract Background. Biomarkers are crucial for decisions regarding adjuvant therapy in primary breast cancer, and their correct assessment is therefore of the utmost importance. Aims. To investigate the concordance between Swedish pathology departments and a reference laboratory, for routine analysis of oestrogen receptor (ER), progesterone receptor (PR), Ki67, and human epidermal growth factor receptor 2 (HER2), alone, and in combination (St Gallen subtypes). Methods. This survey included 27 of the 28 pathology laboratories in Sweden, covering 98% of cases of primary breast cancer surgery in Sweden. Paraffin-embedded tumour blocks (n = 270) were collected and sent to the central reference laboratory, together with the originally stained slides, for re-analysis. The primary evaluations were previously performed according to national Swedish guidelines, without any knowledge of the subsequent central assessment. Results. The agreement for ER, PR, and Ki67 was 99% [kappa value (κ) = 0.95], 95% (κ = 0.85), and 85% (κ = 0.70), respectively. The agreement for HER2 (0/1 + vs. 2+/3+) was 85% (κ = 0.64), but when equivocal tumours were further analysed with in situ hybridisation, only one discrepancy was observed. Discrepancies between results for ER and PR seem to be explained by analytical differences, whereas the interpretation of staining seems to be more critical for Ki67 and HER2 immunohistochemistry. The agreement between the results from the Swedish laboratories and the reference laboratory, based on the St Gallen subtypes, was 88% (κ = 0.81). Conclusions. When applying national guidelines, highly reproducible results were obtained in routine assessment of breast cancer biomarkers, and the results of this study confirm the clinical utility of these markers for decisions regarding the treatment of primary breast cancer.

Comparison of molecular (BluePrint and MammaPrint) and pathological subtypes for breast cancer among the first 800 patients from the EORTC 10041/BIG 3-04 (MINDACT) trial.

32 Background: Biology has become the main driver of breast cancer therapy. Intrinsic biological subtypes by gene expression profiling have been identified. Pathology can be used to define surrogates of these subtypes but these are not always concordant, which may lead to different treatment plans. We investigated the concordance between BluePrint (BP) + MammaPrint (MP) (micro array based) breast cancer subtypes and pathological surrogates (based on ER, PR, HER2 and Ki67). Contrary to the Perou gene set (evolved into PAM50), BluePrint was trained using pathological data. METHODS Using available data (centrally assessed pathology and genomic) from the MINDACT pilot phase (Rutgers et al 2011) 621 tumors were analyzed. Two pathology classifications were used: one with 4 categories and one with 5 categories (Goldhirsch et al 2011). Based on BP 3 subtypes are formed: Luminal, HER2 and Basal. The Luminal subtype is further split into Luminal A (MP low risk) and Luminal B (MP high risk). RESULTS See table. CONCLUSIONS All pathological Basal cases are BP Basal, apart from 1 BP HER2 case. Of the BP Basal cases, 15 are not pathological Basal: 1 is Luminal A, 11 are Luminal B (of which 8 are IHC ER/PR borderline (≥1% and < 10%)) and 3 are HER2. All pathological Luminal (A & B) that are BP HER2 are HER2- by TargetPrint. 25 of the 26 pathological HER2+ that are BP Luminal A are ER+. Most discordant cases are seen within the Luminal subtype, indicating that Ki67 discriminates Luminal A vs. B differently than MammaPrint does. The observed subtype discrepancies reveal potential important impact for treatment-decision making. MINDACT will provide important information. [Table: see text].

Abstract P4-11-08: Pathological assessment of discordant cases for molecular (BluePrint and MammaPrint) vs clinical subtypes for breast cancer, among 6,694 patients from the EORTC 10041/BIG 3-04 (MINDACT) trial

Background Biology has become the main driver of breast cancer (BC) therapy. Biological subtypes have been recommended as a guide for treatment selection. Molecularly identified subtypes can be determined by gene expression profiling. Alternatively, pathology can be used to define surrogates of these subtypes. These methodologies are not always concordant, which can lead to different systemic therapies. The purpose of this preplanned translational research is to investigate the concordance between molecular based BluePrint and MammaPrint breast cancer subtypes and pathological surrogates (based on ER, PgR, HER2 & Ki67) and to characterize the discordant cases. Methods MINDACT is an international, prospective, randomized, phase III trial investigating the clinical utility of MammaPrint in selecting patients with early BC for adjuvant chemotherapy (CT), which enrolled 6,694 patients. Molecular subtyping data were obtained by MammaPrint and BluePrint (Agendia, Amsterdam, the Netherlands) on frozen samples (n=6,694) classifying patients in the following subtypes: Luminal A (Luminal-type/MammaPrint Low Risk); Luminal B (Luminal-type/MammaPrint High Risk); HER2-type; and Basal-type. ER, PgR, HER2 and Ki67 protein status were centrally assessed on FFPE blocks by IHC and/or FISH in the European Institute of Oncology, Milan, Italy (n=5,740; 86%). Patients were also classified according to the St. 2013 Gallen recommendations [Goldhirsch et al. 2013], which recognizes an additional category (Luminal B-like HER2+). Results Ki67 is often used as biomarker to distinguish Luminal A from Luminal B subgroups. The concordance between MammaPrint and centrally assessed Ki67 in Luminal-type patients is 60%, with a κ score of 0.26 (95% CI 0.24–0.28) indicating that Ki67 and MammaPrint cannot reliably substitute for each other. When using a cut-point of 20% instead of 14% the concordance increased to 78%, with a κ score of 0.44 (95% CI 0.41–0.47). There is a relatively large group of clinical HER2+ cases that are BluePrint Luminal-type (208 out of 541; 38%) indicating that tumor expression of the Luminal profile is dominant compared with expression of the HER2 profile. These patients have high IHC ER results and all except for 1 fall into the group that St Gallen separately defines as Luminal B-like HER2-positive. These patients may have lower response to trastuzumab [von Minckwitz et al. JCO 2012]. 98 out of 622 BluePrint Basal-type patients are clinical Luminal HER2-. 2/3 of these patients have low centrally assessed IHC PR expression and 1/3 have low centrally assessed ER expression (≥1% and Conclusions Marked differences are observed between BluePrint and MammaPrint (microarray based) breast cancer subtypes and centrally re-assessed pathological surrogates (based on ER, PR, HER2 & Ki67). The greatest discordance is seen in the sub-stratification of Luminal patients, and in the HR+/HER2+ patients. The observed subtype discrepancies may have an important impact on treatment decision making. Citation Format: Giuseppe Viale, Leen Slaets, Femke A de Snoo, Jan Bogaerts, Laura J van 9t Veer, Emiel J Rutgers, Martine J Piccart-Gebhart, Jeroen van den Akker, Lisette Stork-Sloots, Leila Russo, Patrizia Dell9Orto, Fatima Cardoso, On Behalf of the TRANSBIG Consortium & the MINDACT Investigators. Pathological assessment of discordant cases for molecular (BluePrint and MammaPrint) vs clinical subtypes for breast cancer, among 6,694 patients from the EORTC 10041/BIG 3-04 (MINDACT) trial [abstract]. In: Proceedings of the Thirty-Seventh Annual CTRC-AACR San Antonio Breast Cancer Symposium: 2014 Dec 9-13; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2015;75(9 Suppl):Abstract nr P4-11-08.

Abstract P3-10-36: Concordance of HER2 Central Assessment by Two International Central Laboratories: A Ring Study within the Framework of the Adjuvant HER2-Positive ALTTO Trial (BIG2-06/N063D/EGF106708)

In the Breast International Group (BIG) and North Central Cancer Treatment Group (NCCTG) co-led phase III adjuvant breast cancer trial, ALTTO (Adjuvant Lapatinib and/or Trastuzumab Treatment Optimisation), two central laboratories, the European Institute of Oncology (IEO; Milan, Italy) and Mayo Clinic (Rochester, MN), are responsible for confirming the human epidermal growth factor receptor-2 (HER2), estrogen receptor α(ER), and progesterone receptor status of the primary breast tumors for the Rest of World (excluding China, which conducts a separate central review) and North American patients, respectively, prior to patient study entry. As of December 2009, discordance in HER2 and ER testing between local and central laboratories was observed by both central laboratories. For IEO, 14.5% of 8,037 HER2 cases locally positive [by either immunohistochemistry (IHC) or fluorescence in situ hybridization (FISH)] were not able to be confirmed as centrally positive; whereas for Mayo, 5.8% of 412 HER2 cases locally positive were not able to be confirmed as centrally positive (Table 1). Central and local IHC ER discordance was 12.1% and 11.7% for 9,021 IEO screened cases and 419 Mayo screened cases, respectively (Table 2). In particular, IEO observed more false-positive (locally positive/centrally negative) HER2 findings than Mayo and Mayo observed more false-positive ER findings than IEO. Purpose: Motivated by the above findings, we launched a ring study to assess whether the central lab results of a subset of local/central discordant ALTTO cases could be confirmed in the other central lab. Methods: IEO and Mayo exchanged and retested a subset of FFPE breast tumors collected in ALTTO. IEO sent 20 HER2 false-positive, 5 ER false-positive, and 5 ER false-negative (locally negative/centrally positive) ALTTO breast tumors to Mayo. Mayo sent 5 HER2 false-positive, 20 ER false-positive, and 5 ER false-negative ALTTO breast tumors to IEO. IEO and Mayo performed IHC for ER according to their own methodology: DAKO cocktail of ER 1D5 and 2.123 monoclonal antibodies and monoclonal ER 1D5 antibody, respectively. The two laboratories performed IHC for HER2 according to the HercepTest® manufacturers’ instructions (Dako, Carpenteria, CA) and FISH for HER2 using the PathVysion HER2 DNA probe kit and the HER2/centromere 17 probe mixture (Abbott Molecular, Des Plaines, IL). Results: IEO and Mayo confirmed the central HER2-negative result in 100% of 25 cases. Analyses of ER are ongoing and ER results will be presented. Conclusions: In this subset of patients, enrollment ineligibility did not change when HER2 testing was performed by either IEO or Mayo Clinic central laboratories. Table 1. HER2 Local/Central Lab Results Table 2. ER Local/Central Lab Results Citation Information: Cancer Res 2010;70(24 Suppl):Abstract nr P3-10-36.

Abstract PD7-01: Can surrogate pathological subtyping replace molecular subtyping? Outcome results from the MINDACT trial

Background Molecular subgroups within early breast cancer (EBC), such as Luminal A, Luminal B, HER-2+, Basal-like may help to best to identify patients for specific treatment regimens. Controversy exists as to which methodology is best at identifying these molecular subgroups. Immunohistochemistry (IHC) may be used as a surrogate method to stratify patients. Molecular subtyping gene expression based tests, such as BluePrint, measure a greater number of genes than pathological criteria. ER, PgR, HER-2 and Ki67 are measured individually at the protein level, while BluePrint is designed to capture the functional underlying biologic pathway regulated by these receptors. Methods The MINDACT trial is an international, prospective, randomized, phase III trial which has proventhe clinical utility of MammaPrint in selecting EBC patients who can safely avoid chemotherapy. Here we present the results of a preplanned MINDACT sub-study to compare outcome based on molecular subtyping (MS) to surrogate pathological subtyping (PS) as endorsed by 2013 St. Gallen Consensus. MS data were obtained by MammaPrint (MP) and BluePrint classifying patients in the following subtypes: Luminal A (MP Low Risk); Luminal B (MP High Risk); HER2-type; and Basal-type. ER, PgR, HER2 and Ki67 protein status were centrally assessed by IHC/FISH. The primary hypothesis was that among PS Luminal patients, patients with HER-2+ or Basal-type tumors by MS would have a decreased DMFS compared to MS Luminal patients. At α=5% with 220 events, the study has 80% power to demonstrate this for HR=2.44. Results The table depicts classification of tumors according to PS versus MS for all patients (n=5,806). Most pronounced differences: MS classified 54% as Luminal A among the Luminal B by PS. MS classified 38% as Luminal (A and B) and 5% as Basal-type among the HER-2+ by PS. MS classified 5% as Luminal (A and B) among the TN cases by PS. MS identifies 63% of patients as Luminal A, while PS identifies 47%; 5yr DMFS for both methods was ≥ 96.0%. PS Luminal cancers that were classified as HER-2+ or Basal-type by MS had a lower 5yr DMFS (88.0% for HER-2+ and 90.2% for Basal), albeit non-significant, than those who were also Luminal by MS (95.9%): HR= 1.40, 95% CI = 0.75-2.60. In PS TN cancers, MS identified 24 out of 500 patients (5%) as Luminal-type with excellent prognosis (5yr DMFS of 100% versus 71.4% for MS HER-2+ or 90.1% for MS Basal-type). Among the PS Luminal patients, Ki67 cut at 20% identified patients with ki67 low (69%), with 5yr DMFS ≥ 96.0% (better compared to the 14% cut-off). Conclusions 1) MS was able to re-stratify 16% of patients to a low risk Luminal A-type group with an excellent outcome. 2) Among TN EBC, 5% were classified as Luminal by MS and had an excellent outcome. 3) Albeit limited by low numbers of patients in each subgroup, this study suggest that MS is better correlated with outcome. 4) The observed subtype discrepancies may have an impact on treatment decision making. 5) Centrally assessed Ki67 labeling index of 20% may be the best cut-off for surrogate differentiation between Luminal A and B. Citation Format: Cardoso F, Slaets L, de Snoo F, Bogaerts J, van 9t Veer LJ, Rutgers EJ, Piccart-Gebhart MJ, Stork-Sloots L, Russo L, Dell9Orto P, Viale G. Can surrogate pathological subtyping replace molecular subtyping? Outcome results from the MINDACT trial [abstract]. In: Proceedings of the 2016 San Antonio Breast Cancer Symposium; 2016 Dec 6-10; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2017;77(4 Suppl):Abstract nr PD7-01.

Immunohistochemical Performance of Estrogen and Progesterone Receptor Antibodies on the Dako Omnis Staining Platform: Evaluation in Multicenter Studies.

The analysis of estrogen receptor (ER) and progesterone receptor (PR) expression levels by immunohistochemistry is an important part of the initial evaluation of breast cancer and critically important in treatment planning. Anti-ER&agr; (clone EP1) and anti-PR (clone PgR 1294) antibodies are in development for the Dako Omnis automated staining platform. These antibodies are not yet commercially available and are in performance evaluation, including the 4 international, multicenter studies reported here. For each antibody, a reproducibility study and a method comparison study was done in a randomized manner in order to test the antibodies under conditions closest to real-world user conditions. The reproducibility studies included 5 staining runs on the Dako Omnis with 20 formalin-fixed and paraffin-embedded human breast carcinoma specimens in 3 independent laboratories, and the method comparison studies included several hundred specimens stained on the Dako Omnis and on the Autostainer Link 48 platforms. Stained slides were evaluated for nuclear ER or PR expression according to American Society of Clinical Oncology/College of American Pathologists guidelines (≥1% cut-off for positive) by pathologists who were blinded from the staining method and specimen ID. For both anti-ER&agr; (clone EP1) and anti-PR (clone PgR 1294) on the Dako Omnis, high reproducibility agreement rates were obtained on the interrun, interlaboratory, and interobserver endpoints. High concordance rates were observed between the specimens stained on the Dako Omnis platform and the Autostainer Link 48 platform. Staining quality was excellent for both anti-ER&agr; (clone EP1) and anti-PR (clone PgR 1294) on the Dako Omnis. These results suggest that these antibodies are reliable and reproducible tools for immunohistochemistry analysis of ER and PR expression levels in formalin-fixed and paraffin-embedded breast carcinoma tissues on the Dako Omnis platform.

Abstract P1-01-16: Performance evaluation of two ready-to-use antibodies under development for the Dako Omnis automated staining platform on breast carcinoma specimens: Anti-estrogen receptor α clone EP1 and anti-progesterone receptor clone PgR 1294

Background: The expression of estrogen receptor alpha (ERα) and progesterone receptor (PR) in breast carcinomas is a strong predictor of the efficacy of hormonal therapy for breast cancer patients as well as providing a degree of prognostic information. Anti-ERα (clone EP1) and anti-PR (clone PgR 1294) configured as FLEX ready-to-use antibodies have been tested on the Dako Omnis automated staining platform. These products are in performance evaluation and are not commercially available. A series of concordance studies were performed to evaluate the performance characteristics of these monoclonal antibodies on breast cancer tissue specimens: anti-ERα clone EP1/Dako Omnis was compared to (a) anti-ERα clone EP1/Autostainer Link 48 (238 specimens) and to (b) anti-ERα clone SP1/Autostainer (116 specimens), and anti-PR clone PgR 1294/Dako Omnis was compared to (a) anti-PR clone PgR 636/Autostainer Link 48 (289 specimens) and to (b) anti-PR clone 16 (Leica Biosystems, Newcastle, UK) (144 specimens). In addition, the specificity of the ER and PR antibodies for Dako Omnis was evaluated on a set of normal tissue specimens. Methods: Formalin-fixed, paraffin-embedded (FFPE) human breast carcinoma specimens and normal tissues were obtained from commercial providers or local hospitals. The specimens had no associated personal information and were not traceable back to the tissue donors. Tissue pretreatment and immunohistochemical staining were performed using the recommended protocol for each antibody and staining platform. The stained slides were evaluated for nuclear ER or PR expression according to ASCO/CAP guidelines (≥1% cut-off for positive) by pathologists who were blinded from the staining method and specimen ID. The concordance studies included breast cancer specimens covering the clinical range of ER or PR expression with approximately half the specimens in the negative ( Results: High concordance rates were observed with both anti-ERα clone EP1/Dako Omnis and anti-PR clone PgR 1294/Dako Omnis compared to the other ER/PR antibodies, with overall agreement rates exceeding 95% in all of the comparative studies. On a set of normal tissues, specific positive nuclear staining was observed only in tissue types known to express ERα or PR. Conclusions: Monoclonal antibodies anti-ERα clone EP1 and anti-PR clone PgR 1294 configured as FLEX ready-to-use on Dako Omnis are sensitive and specific assays for detecting estrogen receptor and progesterone receptor in FFPE tissues. In comparison testing for assessment of hormonal receptor status on breast carcinoma specimens, anti-ERα clone EP1/Dako Omnis and anti-PR clone PgR 1294/Dako Omnis were highly concordant with commercially-available ER or PR antibodies. Citation Format: Viale G, Dell9Orto P, Falzon M, Falt A, Hicks D, Hoff K, Jakobsen K, Jensen LB, Levy YY, McMahon L, Miller K, Russo L. Performance evaluation of two ready-to-use antibodies under development for the Dako Omnis automated staining platform on breast carcinoma specimens: Anti-estrogen receptor α clone EP1 and anti-progesterone receptor clone PgR 1294. [abstract]. In: Proceedings of the Thirty-Eighth Annual CTRC-AACR San Antonio Breast Cancer Symposium: 2015 Dec 8-12; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2016;76(4 Suppl):Abstract nr P1-01-16.

Abstract P5-02-02: Central testing of ER, PR, HER2, and Ki67, routinely analyzed at 27 pathology departments in Sweden – Potential consequences for the choice of adjuvant therapy

Background Immunohistochemical (IHC) techniques are still gold standard for measurement of estrogen receptors (ER), progesterone receptors (PR), the proliferation marker Ki67, and human epidermal growth factor receptor 2 (HER2) (confirmation with in situ hybridization for 2+) in routine breast cancer pathology. These biomarkers are also included in the St Gallen 2013 four-marker panel, as a surrogate for the intrinsic subtype classification; Luminal A-like, Luminal B-like (HER2-), Luminal B-like (HER2+), HER2+ (non-luminal), and Triple negative. Correct analyses are of outmost importance since these biomarkers are underlying treatment decisions in breast cancer. It is mandatory for departments performing breast cancer pathology to participate in quality assurance programs. Aims In 2013, SweQA-breast cancer (Swedish Quality Assurance) performed a study to investigate the agreement of ER-, PR-, Ki67-, and HER2-scoring between pathology departments in Sweden and a central reference laboratory (European Institute of Oncology (EIO), Milan, Italy). To explore the clinical consequences, the aim was also to investigate to what extent the central testing would potentially have altered the adjuvant medical therapy. Methods Twenty-seven of the 31 pathology departments in Sweden participated by collecting the first breast carcinoma diagnosed after the 15th each month during 2012, except for July and December (n=270). Paraffin embedded tumor tissue were collected and sent to EIO, where new sections, stainings, and evaluations were performed. These results were compared with the results from the routine analyzes performed in Sweden. Expression scores of ER, PR, Ki67, and HER2 were dichotomized according to predefined thresholds from the St Gallen consensus statement of 2013 (Goldhirsch et al. Ann Oncol 24(9);2206-13;2013). Results The pairwise agreement figures for ER, PR, and Ki67 were; 99% (kappa-value (κ)=0.95), 95% (κ=0.84), and 85% (κ=0.70) respectively. Eight of the PR discrepant cases were locally positive but centrally negative, whereas six showed the opposite pattern. In 34 of the 39 discrepant cases for Ki67, one or both assessments were located near cut-off (15-25%). When categorizing HER2 as 0/1+ vs. 2+/3+, agreement was 86% (κ=0.65). The HER2-IHC 2+/3+ cases in Sweden were further analyzed with in situ hybridization as part of clinical routine, and when these results were used to obtain the final HER2 status, only 1 discrepant case of 265 evaluable cases was observed. When using the St Gallen 2013 surrogate definition for the five intrinsic subtypes, the overall pairwise agreement between Sweden and Italy, for all tumors with complete data (n=256), was 86% (κ=0.78). Conclusions The agreement was very good (κ>0.80) or good (κ 0.61-0.80) for all four biomarkers (ER, PR, Ki67, and HER2). Almost 90% of the discrepant cases for Ki67 were explained by values near cut-off. When strictly applying the St Gallen 2013 surrogate definition of the five intrinsic subtypes, without regard to TNM stage, the results of the central testing would potentially have altered the medical adjuvant treatment for 37 of the 256 patients (14%). For 28 of these 37 patients, the discordance was explained by Ki67. Citation Format: Maria Ekholm, Dorthe Grabau, Per-Ola Bendahl, Goran Elmberger, Hans Olsson, Leila Russo, Guiseppe Viale, Marten Ferno. Central testing of ER, PR, HER2, and Ki67, routinely analyzed at 27 pathology departments in Sweden – Potential consequences for the choice of adjuvant therapy [abstract]. In: Proceedings of the Thirty-Seventh Annual CTRC-AACR San Antonio Breast Cancer Symposium: 2014 Dec 9-13; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2015;75(9 Suppl):Abstract nr P5-02-02.