Clinical implications of discordance in HER2 reassessment from HercepTest to 4B5 in metastatic breast cancer Authors: Yuhei Yamaguchi, Mai Hoshino, Tatsunori Shimoi, Hiroki Tamura, Yutaka Natori, Takuro Saiki, Taro Yamanaka, Rui Kitadai, Ayumi Saito, Shosuke Kita, Aiko Maejima, Yuki Kojima, Kazuki Sudo, Akiko Miyagi Maeshima, Takeshi Murata, Shin Takayama, Yasuto Naoi, Kan Yonemori, Masayuki Yoshida Published: 2026 ABSTRACT AbstractPurposeHER2-low breast cancer has emerged as a therapeutic category following the DESTINY-Breast04 trial, where trastuzumab deruxtecan (T-DXd) improved survival. As the Ventana 4B5 assay was the companion diagnostic in this trial, many patients previously tested with the Dako HercepTest required reassessment. However, the frequency and clinical predictors of discrepancies between these assays remain unclear. This study aimed to evaluate discordance frequency between HercepTest® and 4B5 and identify associated clinicopathological factors.MethodsWe retrospectively analyzed patients who underwent prior HER2 testing with HercepTest and subsequent reassessment with 4B5 between May 2023 and November 2024 at our institution. The primary endpoint was reassessment discordance, and secondary analyses examined the clinicopathological predictors of discordance.ResultsEighty-four patients (median age: 58 years) were included. Specimens were obtained from primary tumors (28.6%) and metastatic lesions (71.4%); 15.5% were resections and 84.5% were biopsies. The discordance rate between HercepTest® and 4B5 was 60.7%, most commonly a shift from 2 + to 1+, followed by 1 + to 0. Multivariable analysis identified progesterone receptor negativity (odds ratio [OR] 3.84, 95% confidence interval [CI] 1.39–11.59, p = 0.01) and metastatic site sampling (OR 3.77, 95% CI 1.14–13.67, p = 0.03) as independent predictors. Approximately one-quarter of discordant patients lost their HER2-low status and potential eligibility for T-DXd.ConclusionsDiscordance between HercepTest® and 4B5 is frequent and clinically consequential, leading to reclassification based on 4B5 that affects eligibility assessment in a substantial subset of patients. These findings highlight the need for careful interpretation of HER2 results in real-world reassessment and increased awareness of diagnostic variability in HER2-low classification.Supplementary InformationThe online version contains supplementary material available at 10.1007/s10549-026-07994-6.Keywords: HER2, HER2-low, 4B5, HercepTest, Concordance Introduction Breast cancer is the most commonly diagnosed cancer and the leading cause of cancer-related deaths among women worldwide [1]. Although advances in early detection and adjuvant therapy have improved the outcomes of many patients, the prognosis of patients with human epidermal growth factor 2 (HER2)-negative metastatic breast cancer (MBC) remains poor. Chemotherapy is recommended for patients with visceral metastases; however, its therapeutic efficacy remains limited [2]. Therefore, novel therapeutic strategies are urgently needed. Historically, HER2-negative breast cancer has been defined as HER2 immunohistochemistry (IHC) scores of 0, 1+, or 2 + with negative in situ hybridization (ISH). However, results from the DESTINY-Breast04 trial demonstrated that trastuzumab deruxtecan (T-DXd) has a significant survival benefit in patients with HER2-low (IHC 1 + or 2 + with ISH-negative) MBC, leading to its approval and widespread adoption in clinical practice [3]. In the trial, T-DXd significantly prolonged both progression-free survival and overall survival compared with the physician’s choice of chemotherapy in patients with HER2-low breast cancer. These findings have introduced HER2-low as a novel therapeutic category and raised the need for accurate pathological assessments. Additionally, the DESTINY-Breast06 trial demonstrated that T-DXd improved outcomes compared with physician’s choice chemotherapy in patients with HR-positive, HER2-low or HER2-ultralow (IHC 0 with membrane staining) MBC, leading to its approval as first-line treatment [4]. Nevertheless, HER2-low classification relies on semi-quantitative IHC scoring, which is subject to variability, especially between different diagnostic assays. Pre-analytical factors such as fixation conditions may have a particularly significant impact on HER2-low expression [5], and it is essential to ensure that tissue processing complies with the 2018 ASCO/CAP HER2 Testing Guidelines [6]. Among the most widely used assays are Dako HercepTest® (HercepTest II, DAKO, Denmark) and the Ventana 4B5 assay (Ventana ultraView pathway HER2 [4B5], Roche Diagnostics, Japan). However, recent studies suggest that concordance between the two assays is suboptimal [7, 8]. This discrepancy is clinically relevant because 4B5 served as the companion diagnostic in DESTINY-Breast04, and thus served as the benchmark for T-DXd eligibility [3]. Most prior comparisons [7, 8] used simultaneous staining of the same tissue block rather than evaluating real-world clinical scenarios, where HER2 status is reassessed following disease progression or recurrence. Moreover, the clinicopathological factors influencing this discordance remain poorly characterized. Therefore, this study aimed to assess concordance between HercepTest® and 4B5 in a real-world cohort of patients with advanced breast cancer and to identify clinical and pathological factors associated with discordance. Materials and methods Study design This single-center, retrospective, observational study was conducted at the National Cancer Center Hospital (Tokyo, Japan). It was based on the data extracted from electronic medical records. Study period Data were collected from consecutive patients who underwent HER2 IHC reassessment using 4B5 between May 2023 and November 2024. Patient selection Eligible patients were those who had undergone prior HER2 testing with the Dako HercepTest® and subsequent reassessment with 4B5 following its reimbursement approval, primarily to evaluate clinical eligibility for T-DXd. Patients were excluded if HER2 IHC staining with 4B5 could not be performed because of insufficient tissue or poor sample quality. Patient specimens were processed per ASCO/CAP HER2 guidelines [6]: the tissue was fixed in 10% neutral buffered formalin for 6–72 h, blocks were stored in a cool, dark place, and metastatic lesions were subjected to ethylenediaminetetraacetic acid treatment without acid decalcification, which could affect staining. No cell blocks were included. For each case, HER2 IHC was performed on different sections from the same formalin fixed paraffin embedded block using the Dako HercepTest® (HercepTest II, DAKO, Denmark) and the Ventana 4B5 assay (Ventana ultraView pathway HER2 (4B5), Roche Diagnostics, Japan). Results were interpreted in accordance with the ASCO/CAP HER2 testing guidelines [6] in effect at the time of diagnosis. Immunostaining with HercepTest and 4B5 was performed according to the manufacturer’s instructions (Supplementary Table 1). HER2 IHC results for both assays were obtained from medical records. For these assays, HER2 IHC scoring was performed by consensus among multiple pathologists, with the same pathologist (MY) consistently participating in the evaluation. Outcome definitions and measurements The primary outcome was the discordance rate between HercepTest® and 4B5 results. Discordance was defined as cases in which both assays yielded differing HER2 statuses, whereas concordance referred to cases with identical results. Secondary analyses were conducted to identify the clinicopathological factors potentially associated with the discordance between HER2 IHC results. In addition, exploratory analyses were performed to identify factors associated with discordance that could alter eligibility for T-DXd. The number of factors was limited to four based on the sample size and selected from those considered to be related to discordance. For hormone receptors, the progesterone receptor (PgR) was chosen as an explanatory variable, as it was statistically significant in the univariate analysis. The factors evaluated were as follows: PgR status: defined as negative when the Allred proportion score was between 0 and 1. PgR status: defined as negative when the Allred proportion score was between 0 and 1. Specimen collection method (surgical resection vs. core needle biopsy): Surgical resection was defined as excisional sampling under general anesthesia, whereas needle biopsy was defined as percutaneous sampling under local anesthesia. Specimen collection method (surgical resection vs. core needle biopsy): Surgical resection was defined as excisional sampling under general anesthesia, whereas needle biopsy was defined as percutaneous sampling under local anesthesia. Specimen site (primary breast lesion vs. metastatic lesion). Specimen site (primary breast lesion vs. metastatic lesion). Time from specimen collection to 4B5 staining: <500 days vs. ≥500 days, with 500 days selected as the cutoff based on the median interval of 561.5 days. Time from specimen collection to 4B5 staining: <500 days vs. ≥500 days, with 500 days selected as the cutoff based on the median interval of 561.5 days. Statistical analysis Discordance rates between HercepTest® and 4B5 were calculated. Fisher’s exact test was used to evaluate the association between the discordance status and categorical clinicopathological factors, including PgR status, collection method, and specimen site. For multivariate analysis, logistic regression was performed to identify independent factors associated with discordance. The dependent variable was assay status (discordance vs. concordance) between the HercepTest and 4B5. Independent variables included PgR status (positive vs. negative), specimen collection method (surgical vs. biopsy), specimen site (primary breast tumor vs. metastatic lesion), and time from specimen collection to 4B5 staining (< 500 days vs. ≥500 days). Statistical significance was set at p < 0.05. All statistical analyses were performed using R software (version 4.5.1). Results Eighty-four patients were included in this analysis. The median age of the patients was 58 years (range: 29–83 years), and 98.8% were female (n = 83). Estrogen receptor positivity was observed in 75.0% (n = 63) of cases and PgR positivity in 51.2% (n = 43) of cases. Specimens were obtained from primary breast tumors in 28.6% (n = 24) of cases and from metastatic lesions in 71.4% (n = 60) of cases. Surgical resection accounted for 15.5% (n = 13) of the specimens, whereas needle biopsy accounted for 84.5% (n = 71) (Table 1). Table 1Patient characteristicsOverall, N = 84 (%)Discordant, N = 51 (%)Concordant, N = 33 (%)P valueAge, median (range), years58 (29–83)56 (29–83)62 (32–80)0.19SexFemale83 (98.8)50 (98.0)33 (100)1.00Male1 (1.2)1 (2.0)0ERPositive63 (75.0)35 (68.6)28 (84.8)0.12Negative21 (25.0)16 (31.4)5 (15.2)PgRPositive43 (51.2)21 (41.2)22 (66.7)0.03*Negative41 (48.8)30 (58.8)11 (33.3)Sampling methodSurgery13 (15.5)6 (11.8)7 (21.2)0.35Needle biopsy71 (84.5)45 (88.2)26 (78.8)SpecimenPrimary24 (28.6)10 (19.6)14 (42.4)0.03*Metastasis60 (71.4)41 (80.4)19 (57.6)HercepTest: IHC score2+54 (64.3)38 (74.5)16 (48.5)0.03*1+27 (32.1)12 (23.5)15 (45.5)03 (3.6)1 (2.0)2 (6.1)4B5: IHC score2+16 (19.0)016 (48.5)< 0.01*1+53 (63.1)38 (74.5)15 (45.5)015 (17.9)13 (25.5)2 (6.1)Staining intervalmedian (range), days561.5 (6–4,825)631 (6–4,825)427 (10–3,112)0.46< 500 days41 (48.8)24 (47.1)17 (51.5)0.82≥ 500 days43 (51.2)27 (52.9)16 (48.5) Patient characteristics Using the HercepTest, 64.3% of the patients (n = 54) were classified as 2+, 32.1% (n = 27) as 1+, and 3.6% (n = 3) as 0. Using the 4B5 assay, 19.0% of the patients (n = 16) were 2+, 63.1% (n = 53) were 1+, and 17.9% (n = 15) were 0. Discordance occurred in 60.7% (n = 51) of patients, with the most frequent transitions being 2 + to 1+ (n = 37) and 1 + to 0 (n = 12). Concordance between the HercepTest and 4B5 results was observed in 39.3% of the cases (n = 33), including 2 + to 2+ (n = 16), 1 + to 1+ (n = 15), and 0 to 0 (n = 2) (Fig. 1). HercepTest and 4B5 IHC results. This diagram shows the transition from HercepTest staining categories to those of 4B5. The orange line represents the most frequent change among the discordant cases, namely a shift from HercepTest 2 + to 4B5 1+. Red lines indicate cases in which the 4B5 results would alter eligibility for T-DXd The mean interval from specimen collection to 4B5 testing was 631 and 427 days in the discordant and concordant groups, respectively. In the univariate analysis, while organ-level comparisons showed no significant differences in discordance (p = 0.21; Supplementary Fig. 1), significant associations with concordance status were observed for PgR expression (p = 0.03), specimen site (primary vs. metastatic, p = 0.03), and initial HercepTest score (p = 0.03) (Table 1). Discordance was found in 10 cases and concordance in 14 cases among primary breast tumors, whereas discordance was observed in 41 cases and concordance in 19 cases among metastatic lesions. Regarding the specimen collection methods, discordance was observed in 6 of 13 (46.2%) surgical resections and 45 of 71 (63.4%) needle biopsies (p = 0.35). Stratification by a 500-day cutoff revealed similar discordance rates between the groups, with 24 of 41 patients being discordant in the < 500-day group and 27 of 43 in the ≥ 500-day group (p = 0.82). Finally, a multivariable logistic regression analysis identified PgR negativity (OR 3.84, 95% CI 1.39–11.59, p = 0.01) and metastatic sampling site (OR 3.77, 95% CI 1.14–13.67, p = 0.03) as independent predictors of discordance (Fig. 2). PgR negativity was also identified as an independent predictor of T-DXd eligibility-changing discordance (OR 4.09, 95% CI 1.09–19.44, p = 0.05) (Fig. 3). Multivariable analysis of discordance between HercepTest and 4B5 staining. Specimens that were PgR-negative or obtained from metastatic lesions were significantly more likely to show discordant staining results between HercepTest and 4B5 Multivariable analysis of T-DXd eligibility changing discordance between HercepTest and 4B5 staining. Specimens that were PgR-negative were significantly more likely to show T-DXd eligibility-changing discordant staining results between HercepTest and 4B5 Discussion In this study, we systematically investigated the concordance between two HER2 IHC assays (HercepTest and Ventana 4B5) in a real-world cohort of patients with MBC, following the reimbursement approval of 4B5 in Japan. Our findings should not be interpreted as a direct analytical comparison between two assays performed under controlled conditions. Rather, they reflect real-world variability in HER2 reassessment across diverse clinical contexts, including differences in specimen type, timing of biopsy, and pathological interpretation. We observed a high discordance rate of 60.7% in IHC results between HercepTest and 4B5; among the discordant cases, 27.5% led to a change in HER2-low classification that could affect eligibility for T-DXd. Furthermore, multivariable analysis identified PgR negativity and sampling from metastatic lesions as factors associated with discordance. These findings suggest that HER2-low classification is not a fixed biological entity, but rather a context-dependent and potentially unstable category, particularly in borderline cases between HER2 IHC 0 and 1+. Taken together, our results indicate diagnostic uncertainty in HER2 reassessment in the current therapeutic landscape. The observed discordance does not represent a comparison of the superiority of one assay over another; rather, it is likely to reflect a complex interplay of assay-related differences, spatial and temporal tumor heterogeneity, and interobserver variability. This complexity has important clinical implications for the interpretation of HER2 IHC results in treatment decision-making. Discordance in HER2 status between HercepTest and 4B5 was observed in 60.7% of the patients, while concordance was observed in only 39.3%. The discordance rate in our study substantially exceeds that reported in previous analytical studies [7–9]. The most frequent discordance pattern was a downward shift from HercepTest 2 + to 4B5 1+, which indirectly influenced therapeutic decision-making, especially regarding eligibility for T-DXd in HER2-low MBC. Among the 51 discordant cases, one case (2.0%) shifted from 2 + to 0, and 12 cases (23.5%) shifted from 1 + to 0. These cases became ineligible for T-DXd based on the 4B5 result. These findings highlight the possibility that switching assays in clinical practice may alter the eligibility for HER2-low-directed treatment, particularly in cases initially categorized as 2 + or 1 + by the HercepTest. If 4B5 is scored as 2 + or 1+, the tumor is classified as HER2-low and the patient remains a candidate for T-DXd; however, if 4B5 is scored as 0, the patient is reclassified as HER2-zero and consequently excluded from T-DXd therapy. Thus, assay discordance has the potential to change treatment options in routine clinical practice. Multivariable logistic regression identified PgR negativity (vs. positivity) and the metastatic specimen site (vs. primary) as independent predictors of discordance (OR 3.84, 95% CI: 1.39–11.59, p = 0.01; OR 3.77, 95% CI: 1.14–13.67, p = 0.03), with PgR negativity also identified as an independent predictor of T-DXd eligibility-changing discordance (OR 4.09, 95% CI: 1.09–19.44, p = 0.05). In hormone receptor-negative breast cancer, a mixture of basal-like and non-basal-like phenotypes has been reported, even among HER2-low tumors, with marked the spatial and clonal heterogeneity of HER2 protein expression [10]. Previous genomic analyses have shown that estrogen receptor (ER)-positive/PgR-negative breast cancers constitute a biologically and genomically heterogeneous subgroup, with distinct mutational profiles compared with PgR-positive tumors. This biological heterogeneity may partly explain the association between PgR negativity and HER2 immunohistochemical discordance observed in the present study [11]. Considering prior observations that metastatic lesions exhibit HER2 heterogeneity and downregulation owing to clonal evolution [12–14], staining and pathological evaluation using 4B5 may be challenging in HER2-low tumors. When 4B5 staining is performed on metastatic biopsy specimens, clinicians should be aware that discordance with prior HercepTest results may be more frequent compared with that from staining primary tumor samples. However, further validation using matched primary-metastatic pairs and ISH correlations is warranted. Contrary to previous reports suggesting that HER2 antigenicity degrades over time [15–20], we observed no significant association between the interval from specimen collection to 4B5 testing and discordance status, even when stratified using a 500-day cutoff. A specimen storage period of 500 days (approximately one and a half years) is generally not expected to compromise specimen quality. Moreover, organ-specific analysis revealed concordant or discordant patterns; however, formal statistical testing was not performed because the sample sizes for individual metastatic sites were limited. These findings suggest that the observed discordance is primarily attributable to assay-specific differences, rather than time- or site-related variables. Since HercepTest and 4B5 differ in antibody type (polyclonal vs. monoclonal), epitope recognition site, and staining protocol, variability in the staining outcomes is more likely to occur, particularly in the context of heterogeneous HER2 expression [8]. Indeed, it has been reported that HER2 status may vary depending on the antibody used [21]. This study had some limitations. First, all patients underwent 4B5 testing only once using specimens that had been previously evaluated with the HercepTest, and in those who underwent testing for recurrent lesions, the 4B5 results from the primary tumor were unavailable. Thus, whether HercepTest and 4B5 results differ between primary and metastatic lesions in the same patient remains unresolved. Second, the absence of a centralized pathology review introduces potential interobserver variability, although the pathological evaluation in our study was reviewed by at least two pathologists with expertise in breast pathology. Notably, several prior studies have demonstrated that HER2-low scoring, especially in distinguishing between IHC 0 and 1+, is associated with poor reproducibility and substantial interobserver and interinstitutional variability [9, 22]. This challenge may be partly attributable to the fact that the latest ASCO/CAP HER2 testing guidelines [6] do not provide detailed examples of extremely faint or barely recognizable staining patterns in tumor cells, which are critical for determining HER2-low or HER2-ultralow status. In addition, these guidelines do not specifically address borderline cases in which the proportion of stained tumor cells is approximately 10%, which may have hindered standardization. In this context, a recent expert consensus [23] provides practical guidance on categorizing rare staining patterns that may cause confusion in determining HER2-low or HER2-ultralow status and may help standardize clinical practice. The study findings underscore the need for greater standardization of the HER2-low assessment. Although interpretative variability cannot be completely excluded, the consensus-based assessments of both assays by multiple pathologists, with one of the pathologists consistently participating in all assessments, may have contributed to consistency across assessments. Third, this study did not fully account for potential variations in staining results owing to differences between tissue sections. Finally, we were unable to evaluate the correlation between the IHC and ISH results. Previous reports have shown discrepancies between IHC and ISH findings in certain cases, which may be attributed to atypical HER2 amplification, such as chromosome 17 polysomy or spatial heterogeneity of HER2 protein expression [8]. These findings underscore the need for greater standardization of the HER2-low assessment. In this context, such diagnostic uncertainty may influence clinical decision-making, particularly in borderline HER2-low cases where treatment eligibility depends on small differences in HER2 IHC interpretation, such as the distinction between IHC 0 and 1+. In conclusion, this study demonstrated substantial discordance between the HercepTest and 4B5 HER2 IHC assays in patients with MBC, with PgR negativity and metastatic disease significantly associated with higher discordance rates. These findings highlight the urgent need for standardized interpretation criteria and further prospective validation to ensure diagnostic consistency and equitable access to HER2-directed therapies. Electronic Supplementary Material Below is the link to the electronic supplementary material. Supplementary Material 1 Acknowledgements We thank the staff of the Department of Medical Oncology, Department of Diagnostic Pathology, and Department of Breast Surgery at the National Cancer Center Hospital for their valuable support in data collection and pathological assessment. Editorial support in the form of English language editing was provided by Editage. Author contributions The study was designed by Y.Y., M.H., T.S., and M.Y. Data collection was performed by Y.Y. and M.Y. Data analysis and interpretation were carried out by Y.Y., M.H., T.S., and M.Y. H.T., Y.N., T.S., T.Y., R.K., A.S., S.K., A.M., Y.K., K.S., A.M.M., T.M., S.T., Y.N., and K.Y. contributed to drafting the manuscript and provided critical review. All authors read and approved the final manuscript. Funding The authors declare that no funds, grants, or other support were received during the preparation of this manuscript. Data availability No datasets were generated or analysed during the current study. Declarations Ethical approval This study was approved by the Institutional Review Board of the National Cancer Center Hospital (IRB Number: 2014-092). Consent to participate and consent to publish The requirement for informed consent was waived due to the retrospective design of the study. Patient data were anonymized prior to analysis to ensure confidentiality. Competing interests The authors declare no competing interests. Footnotes Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. References Associated Data Supplementary Materials Supplementary Material 1 Data Availability Statement No datasets were generated or analysed during the current study.