Patients with human epidermal growth factor receptor 2 (HER2)-positive early breast cancer (EBC) and residual invasive disease at surgery, after chemotherapy and HER2-targeted neoadjuvant systemic therapy (NAST), have higher rates of recurrence and death than those attaining a pathological complete response1,2. In the phase 3 KATHERINE study, adjuvant trastuzumab emtansine (T-DM1) reduced the risk of invasive disease recurrence or death by 50% compared with adjuvant trastuzumab in these high-risk patients3, and changed the standard of care for this patient population4,5.

Changes from HER2-positive to HER2-negative status in residual breast cancer after NAST have been documented6,7,8,9,10, but how this affects outcomes with subsequent adjuvant treatment remains unclear. This is of potential importance with T-DM1, a HER2-targeted antibody-drug conjugate. In this descriptive report we provide available outcome information in KATHERINE patients with HER2-negative status on re-testing of residual disease after NAST.

Paired pre-NAST (core biopsy) and post-NAST (surgical) tumor samples were available for 1002 of the 1486 patients enrolled in KATHERINE, and valid, centrally determined HER2 status results were available in 845 paired samples (invalid/unknown in 157) (Fig. 1a). Pre-NAST samples were preferentially used to assess eligibility and were prospectively confirmed to be HER2-positive. Of the assessable paired post-NAST residual disease samples, 775 (91.7%) were HER2-positive and 70 (8.3%) were HER2-negative. The HER2-negative subgroup consisted of 53 patients with HER2-negative status by both IHC and ISH. We also considered 17 additional patients with IHC 0–1+ and unknown ISH to be HER2-negative given the high likelihood of IHC 0–1+ being ISH-negative11. The rate of HER2 status conversion in KATHERINE is somewhat less than previous series reporting changes to HER2-negative status after HER2-targeted NAST in 13–32% of patients, however, the patient populations differ in some respects6,7,10.

Fig. 1: Analysis of outcomes in patients with HER2-negative vs HER2-positive residual disease upon retesting at surgery.
figure 1

a Patient data assessed for this analysis is shown. Of the 1486 enrolled patients, two in the trastuzumab arm were excluded from the analysis, one without centrally confirmed HER2-positive breast cancer and another who was inadvertently randomized twice. Paired pre-NAST and residual disease samples were available for 845 patients. Patients with HER2-negative residual disease (n = 70) include 53 with confirmed HER2-negative disease and 17 with unknown HER2-status (IHC 0–1+/ISH unknown). b Kaplan–Meier estimates of time to first IDFS event in each treatment arm are shown. Abbreviations: CI confidence interval, HER2 human epidermal growth factor receptor 2, IDFS invasive disease-free survival, NAST neoadjuvant systemic therapy, NE not estimable, and T-DM1 trastuzumab emtansine.

Full size image

To determine whether the 70 patients with HER2-negative residual disease could also be distinguished by baseline characteristics, we compared characteristics between these patients and those maintaining HER2-positive residual disease. Baseline age, sex, race, clinical stage at presentation, and tumor hormone receptor status were similar between the two groups (Table 1). However, the pre-NAST samples of patients with HER2-negative residual disease had increased percentages with HER2 1+ and 2+ staining intensity, gene ratios <4 and heterogenous HER2 protein expression than those who maintained HER2-positive residual disease (Table 1). Patients with HER2-negative residual disease were also more likely to have had residual disease corresponding to pathological stage I disease at surgery (47.1% vs 38.8%; Supplementary Table 1) and to have received dual HER2-targeted NAST (30.0% vs 17.9%; Table 1).

Table 1 Baseline characteristics by HER2 status in residual disease at surgery after neoadjuvant systemic therapy.
Full size table

HER2 gene expression, assessed by RNA sequencing of the surgical samples, was consistent with HER2 status determined by IHC and ISH, showing median HER2 gene expression of 10.8 and 8.5 in HER2-positive and HER2-negative residual disease, respectively (Table 1). While the mechanisms underlying an apparent change in HER2 expression after NAST are unclear, it has been postulated that HER2-negative cells are selected by HER2-directed NAST8. Indeed, higher rates of HER2-negative testing results have been demonstrated after NAST containing HER2-targeted therapy plus chemotherapy compared to chemotherapy alone7, and, in our study, patients with HER2-negative residual disease were more likely to have received dual versus single HER2-targeted NAST. As trastuzumab + pertuzumab-based therapy becomes a universal standard of care for NAST, it will be interesting to evaluate whether increased selection for HER2-negative cells may occur, with a higher rate of conversion to HER2-negative status in residual disease. Reduced reliability of testing in post-NAST specimens could also contribute to apparent HER2 status conversion.

Next, we assessed the potential impact of change in HER2 status on outcomes. In the overall population, regardless of treatment arm, there was no meaningful difference in IDFS between patients who converted to HER2-negative at surgery and those who remained HER2-positive (HR = 0.93; 95% CI: 0.50–1.71). Among those with HER2-negative residual disease, baseline characteristics were balanced between treatment arms (Table 1). There were 11 IDFS events in the 42 patients (26.2%) randomized to trastuzumab, and no events in the 28 randomized to T-DM1. Of the 11 IDFS events in the trastuzumab arm, seven were distant recurrences not in the central nervous system, one was a central nervous system recurrence, two were locoregional recurrences, and one was contralateral breast cancer.

Previous studies have suggested that HER2-negative status of residual disease after NAST is associated with poor prognosis6,7,10, and in this analysis of KATHERINE, patients with HER2-negative residual disease in the trastuzumab arm had a 3-year IDFS of only 70% (Fig. 1b). In contrast, no IDFS events were reported in those receiving T-DM1, suggesting this adverse prognostic effect may be offset with T-DM1. These data are consistent with other data from KATHERINE showing that biomarkers assessed in the surgical sample affected outcomes in the trastuzumab arm, but not in the T-DM1 arm12, however they are limited in that paired sample data were not available from all patients because of insufficient tumor material and/or invalid or unknown HER2 status results.

While the analysis of this small subset of patients must be considered exploratory and descriptive, the data do not support a hypothesis that subsets of patients presenting with HER2-positive EBC found to have HER2-negative residual disease on retesting after NAST may not derive benefit from adjuvant T-DM1. These results and the low rate of HER2 status conversion further suggest that HER2 retesting of residual disease has no clinical utility and should not be a prerequisite for T-DM1 therapy in this setting.

Methods

The KATHERINE study (NCT01772472, registered January 21, 2013) evaluated T-DM1 in patients with HER2-positive EBC who had residual invasive disease and had received taxane/trastuzumab-based NAST. The study design, including patient eligibility criteria and patient disposition, have been published previously3. In brief, patients were eligible for the study if they had histologically confirmed, centrally confirmed, HER2-positive, non-metastatic, invasive primary breast cancer (T1–4, N0–3, M0 [excluding T1aN0 and T1bN0]) at presentation and residual invasive disease detected pathologically in the surgical specimen of the breast or axillary lymph nodes after completion of taxane-based neoadjuvant chemotherapy administered with trastuzumab. Patients were required to have completed at least six cycles (≥16 weeks) of NAST including ≥9 weeks of trastuzumab and ≥9 weeks of taxane-based chemotherapy (or, if receiving dose-dense chemotherapy regimens, ≥6 to 8 weeks of taxane-based therapy and ≥8 weeks of trastuzumab). HER2-directed therapy and chemotherapy could be given concurrently, and patients could have received more than one HER2-directed therapy, and anthracyclines and alkylating agents as part of preoperative therapy. Within 12 weeks of surgery, patients were randomized to adjuvant trastuzumab (6 mg/kg intravenously every 3 weeks) or T‐DM1 (3.6 mg/kg intravenously every 3 weeks) for 14 cycles. Adjuvant radiotherapy and adjuvant hormonal therapy were permitted, as indicated. The KATHERINE study methods were performed in accordance with relevant guidelines and regulations and approved by the institutional review board at each participating center (e.g., Houston Methodist Institutional Review Board). KATHERINE was conducted in accordance with the International Council for Harmonisation E6 Good Clinical Practice Guideline and the principles of the Declaration of Helsinki; and followed local laws and regulations. All patients provided written informed consent.

HER2 status for study eligibility was performed preferentially on specimens collected pre-NAST and, in the analysis described herein, was also assessed on paired residual samples submitted for correlative studies. Samples were submitted to the central laboratory in the form of a formalin-fixed paraffin-embedded tumor block or partial block or slides obtained from the pretreatment primary tumor biopsy material (or residual tumor tissue from definitive surgery post-NAST). HER2 status of all samples was centrally assessed in the same laboratory (Targos Molecular Pathology, GmbH [Kassel, Germany]) and according to interpretation guidelines of the Ventana assays (PATHWAY® anti-HER-2/neu [4B5] assay with rabbit monoclonal primary antibody, Ventana Medical Systems, Inc., cat#790–2991 and INFORM HER2 Dual ISH assay, DNA Probe Cocktail, Ventana Medical Systems, Inc., cat#800–4422). Breast cancer was considered HER2-positive with an immunohistochemistry (IHC) score of 3+ or amplification of HER2 by in situ hybridization (ISH), defined as a ratio of ≥2.0 for the number of HER2 gene copies to chromosome 17 copies. HER2 expression by IHC was recorded as focal (<30%), heterogeneous (30–79%), or homogeneous (≥80%), based on the percentage of cells stained with IHC2+/3+ intensity. RNA expression of the whole transcriptome was measured using RNA sequencing (RNA-seq) with TruSeq RNA Access (Illumina, Inc., San Diego, California) at Expression Analysis (Morrisville, North Carolina) on macro-dissected tumor samples. Results from RNA analysis, adjusted for tumor content, were used to quantify HER2 (ERBB2) expression in support of the IHC and ISH analyses. Gene expression normalization and transformation to log2(nCPM) was performed with the edgeR R package (version 3.32.1). To regress out the tumor content effect the limma R package (version 3.42.0) was used. Median and IQRs of HER2 gene expression were estimated with the quantile function from the stats R base package (R version 4.0.5).

The endpoint for this exploratory analysis was IDFS, defined as the time from randomization until first occurrence of: recurrence of ipsilateral invasive breast cancer, recurrence of ipsilateral locoregional or contralateral invasive breast cancer, distant recurrence, or death from any cause. Unstratified hazard ratios and 95% confidence intervals were estimated using Cox proportional hazards models. Three-year IDFS rates were estimated with the Kaplan–Meier method. P values were not computed since analyses were exploratory. IDFS was evaluated by randomized treatment arm in patients who had paired specimens with central HER2 status results.

Reporting summary

Further information on research design is available in the Nature Research Reporting Summary linked to this article.