TBCRC023: A Randomized Phase II Neoadjuvant Trial of Lapatinib Plus Trastuzumab Without Chemotherapy for 12 versus 24 Weeks in Patients with HER2-Positive Breast Cancer

Breast tumors with overexpression or amplification of the HER2 gene (ERBB2 or HER2) comprise 15%–20% of breast cancer (1). These HER2-positive breast cancers have an aggressive clinical behavior, and used to be associated with poor prognosis. However, the introduction of the monoclonal anti-HER2 antibody trastuzumab has dramatically improved outcomes when added to combination chemotherapy (2–4).

HER2 belongs to a membrane tyrosine kinase family (HER1–4). This family works in a coordinated fashion to activate downstream signaling pathways that regulate cell proliferation, survival, angiogenesis, differentiation, invasion, and metastasis (5). Using two different combinations of anti-HER agents, trastuzumab together with either lapatinib or pertuzumab, our preclinical data were the first to demonstrate that HER2-positive xenograft tumors could be eradicated in an in vivo model (6–8). However, with both combinations of anti-HER2 agents, simultaneous blockade of estrogen receptor (ER) was required to enhance efficacy.

Results of several independent clinical trials validated the concept that potent inhibition of HER2 with combination therapy targeting HER receptors (commonly referred to as dual anti-HER2 therapy) without the use of chemotherapy yields a clinically meaningful rate of pathologic complete response (pCR) in the neoadjuvant setting (9–11). The NeoSphere trial showed a pCR rate of 17% in the arm with trastuzumab plus pertuzumab and no chemotherapy (11). In the TBCRC006 trial (NCT00548184), we showed that the combination of lapatinib and trastuzumab (with the addition of endocrine therapy in patients with ER-positive tumors) for 12 weeks resulted in a pCR rate of 27% in patients with a median tumor size of 6 cm (9). The PAMELA trial reported by Prat and colleagues, using the same dual anti-HER2 regimen administered for 18 weeks, reported an overall pCR rate of 30% (10).

These studies suggest that a subgroup of patients may not require chemotherapy, thus sparing these patients the unnecessary toxicity and cost of that treatment. Other studies in the neoadjuvant setting that combined chemotherapy with dual HER2 inhibition demonstrated increased rates of pCR with more complete blockade of the HER family, although in some cases it was modest and not statistically significant (11–14).

An intriguing finding from TBCRC006 was that 22% of patients had residual invasive disease of ≤1cm (ypT_1a-b) despite a median baseline tumor size of 6 cm. Most of these responses were observed in ER-positive tumors (9). This raised the questions of optimal treatment duration and whether a longer duration beyond 12 weeks may result in higher pCR rates, especially in ER-positive tumors. If so, this may help increase the proportion of patients achieving pCR who may not require chemotherapy.

We therefore hypothesized that neoadjuvant treatment for 24 weeks with dual anti-HER2 therapy without chemotherapy, plus endocrine therapy for ER-positive disease, would increase pCR rates in patients with HER2-positive breast cancer over the same treatment administered for only 12 weeks.

TBCRC023 (NCT00999804) was a randomized multicenter phase II study conducted in collaboration with the Translational Breast Cancer Research Consortium (Fig. 1A). Institutional review board and scientific committee approval were obtained at the lead site (Dan L Duncan Comprehensive Cancer Center at Baylor College of Medicine, Houston, TX) and other participating sites. Written informed consent was obtained from all patients.

Eligible patients were 18 years of age or older with histologically confirmed invasive breast carcinomas that were HER2-positive, assessed by IHC or FISH, assessed locally according to the 2007 American Society of Clinical Oncology (ASCO)/College of American Pathologists (CAP) guidelines (15). Breast tumors needed to be 2 cm or larger by clinical measurement, with any nodal status. Good performance status (Eastern Cooperative Oncology Group performance status 0–2) and adequate organ function were also required.

Study participants were randomized in a ratio of 1:2 to receive 12 versus 24 weeks of lapatinib (Tykerb, supplied by GlaxoSmithKline) 1,000 mg orally every day and trastuzumab 4 mg/kg loading dose followed by 2 mg/kg per week. If the tumor was ER-positive and/or progesterone receptor (PR)-positive by IHC (according to the 2010 ASCO/CAP guidelines; ref. 16), patients were also treated with letrozole 2.5 mg orally once per day (combined with luteinizing hormone-releasing hormone (LHRH) agonist of physician's choice in premenopausal women).

Tumor tissue was collected at baseline (before treatment), and after 1 week and at the time of surgery (or completion of study treatment). Tumor biopsies were also performed at 12 weeks on patients on the 24-week arm, if they had residual tumor by ultrasound per the judgment of the local radiologist. Collected samples were divided with a portion placed in formalin for subsequent paraffin embedding (formalin-fixed, paraffin embedded tissue) and a portion flash frozen on dry ice.

Upon completion of study treatment, patients proceeded to surgery. However, further neoadjuvant treatment with chemotherapy was allowed, if deemed clinically necessary. Adjuvant therapy after surgery was at the discretion of the treating physician.

The study measured the rate of pCR (disappearance of all invasive tumor in the breast, ypT_0-is) at the time of surgery. Pathologic assessment of response was assessed according to the local institutional guidelines and practice. Patients who did not proceed to surgery, or received additional neoadjuvant therapy prior to surgery, were considered as not achieving the primary endpoint of pCR.

In this trial, while the protocol mandated the collection of only grade 3 and 4 toxicity data, lower grade toxicities reported by study participants were also recorded as part of the routine on-study assessment. Protocol-defined reportable adverse events consisted of toxicity of any grade that required a modification to treatment (e.g., dose modification, dose delay, or drug discontinuation); all liver toxicity regardless of grade; all ≥ grade 2 cardiac or pulmonary toxicity; abnormal laboratory values or diagnostic test results (if they induced clinical signs or symptoms or required treatment or further diagnostic tests); all grade 3 and 4 adverse events, regardless of causality; and all serious adverse events (SAE).

The design of this randomized trial combined an admissible Simon-like phase II design (17), which minimizes both the number of subjects enrolled on an ineffective treatment and minimizes the total sample size, for the extended (24 weeks) arm and a pick-the-winner-like requirement, at the conclusion of the trial, of numeric superiority of 24-week pCR rate over 12-week pCR rate. Randomization (2:1 in favor of 24 weeks) was stratified by ER status (ER-positive and ER-negative). The trial was designed to detect an increase in the rate of pCR in the 24-week arm from 27% observed at first report of TBCRC006 to 45%, which required up to 55 evaluable patients. The first stage of 31 patients (successful interim analysis conducted in November, 2013) required 10 or more responses to proceed for an additional 24 evaluable patients (total of 55). At the end of the trial, if 20 or more subjects responded, and the pCR rate with extended therapy was numerically superior to the control (12 week) pCR rate, we could conclude in favor of extended therapy. On the basis of simulations, we had 85% power overall to detect a true response rate of 45% in the extended arm and numeric superiority over control (assume control pCR rate = 28%), and no more than a 10% type I error when the extended arm pCR rate was 28% or less. The 12-week control arm did not have an early stopping rule. It accrued while the 24-week arm was open, with maximum enrollment of 28 evaluable subjects (half of the 24-week arm sample size). The control arm allowed an internal benchmark for the overall trial without the need to rely on historical data, including our own TBCRC 006 study that also tested a 12-week duration of therapy.

On April 12, 2013, a protocol amendment was approved to allow the enrollment of an expansion cohort of 40 participants, as long as the trial did not meet other stopping rules, to better characterize the correlative science objectives of the study. No change was made to the trial's primary endpoint or to the efficacy assumptions, and we report here the outcome data based on the prespecified efficacy cohort before the addition of the expansion cohort.

Between November 2011 and November 2013, eligible patients were enrolled at eight sites (Baylor College of Medicine, Dan L Duncan Comprehensive Cancer Center, Houston, TX; Vanderbilt University, Nashville, TN; University of Alabama in Birmingham, Birmingham, AL; Johns Hopkins Sidney Kimmel Cancer Center, Baltimore, MD; University of Chicago, Chicago, IL; Indiana University, Indianapolis, IN; Dana Farber Cancer Institute, Boston, MA; and Mayo Clinic, Rochester, MN).

Of the 100 patients screened, 97 patients were enrolled and randomized in a 1:2 ratio to the 12- versus the 24-week arm (Fig. 1B, CONSORT diagram). Thirty-one of the 33 patients randomized to the 12-week arm completed study treatment. Two patients discontinued therapy early on this arm, one due to progressive disease, and one due to toxicity (Supplementary Table S1). The study treatment was otherwise well tolerated with grade 1–2 diarrhea and acneiform rash being the most common toxicities. Tumor biopsies were collected at baseline (before treatment), and at different time points thereafter, as described in the Patients and Methods section, from evaluable randomized patients (Table 1).

Three patients in the 24-week arm were found ineligible after they received treatment: 2 had metastatic disease at baseline, and 1 had disease that was FISH negative and was taken off the study by her doctor before completing one cycle of lapatinib. Forty-seven of 61 eligible patients on the 24-week arm completed study treatment. Fourteen patients discontinued treatment early (Supplementary Table S1). Seven were removed because of disease progression, 3 due to toxicity, and 4 withdrew consent; all of these were considered as not achieving pCR in the intent-to-treat analysis. The characteristics of patients with disease progression in the 12-week and 24-week arms are summarized in Supplementary Table S2.

Table 2 summarizes patient characteristics. Median age was 51, and 45% of the patients were premenopausal. Study participants came from diverse racial and ethnic backgrounds (listed in the table). Study participants had tumors with high-risk features, including a median tumor size of 5 cm (range 0–15 cm), nearly a third with stage III disease, and 71% with histologic grade 3. Almost two thirds of patients (65%) had tumors that were ER-positive.

The overall pCR rate in the breast was lower than expected. The 24-week arm passed stage 1 of the study design but in stage 2, did not meet the requirement for at least 20 pCRs out of the first 55 response-evaluable participants. Overall, the pCR rate was 12% in the 12-week arm and 28% in the 24-week arm (Table 3). This increase was seen entirely in the ER-positive subgroup, where the pCR was 9% in the 12-week arm and 33% in the 24-week arm. The pCR rate in the ER-negative subgroup was similar in both study arms, with 20% in the 12-week arm and 18% in the 24-week arm.

Figure 2 shows pathologic responses in each treatment arm by ER status. In ER-positive tumors, 30% of the patients in the 12-week arm had residual tumors of 1 cm or less, while only 13% of patients in the 24-week arm had that response. In the ER-negative group, there were no patients with residual disease of 1 cm or less in the 12-week arm, while 18% of patients in the 24-week arm had that response. A total of 8 patients (8.5%) were taken off study for disease progression, 1 (3%) in the 12-week arm, and 7 (11%) in the 24-week arm.

Study treatment was generally well tolerated. Table 4 summarizes reportable toxicities per study protocol. Toxicity was not common and included in the 24 weeks arm: grade 4 elevated liver function tests in 1 patient (2%), which was considered as a serious event possibly related to study treatment; grade 3 diarrhea in 1 patient (2%); and grade 3 mucositis in 1 patient (2%). The 12 weeks arm had 1 grade 3 anemia (3%) and 1 grade 3 renal calculi (3%). The latter was the other serious adverse event in the primary cohort and was not related to study medication.

The introduction of anti-HER2 therapy has led to dramatic improvements in the outcome of patients with HER2-positive breast cancer. In the neoadjuvant setting, dual HER2 targeting increases pCR rates with chemotherapy (9–11).

TBCRC006 demonstrated a 27% pCR rate despite the large median tumor size of 6 cm (9). Moreover, 22% of patients had residual in-breast disease of 1 cm or less (33% in patients whose tumors were ER-positive; ref. 9). It was therefore reasonable to hypothesize that longer treatment would result in higher pCR rate especially in the ER-positive subgroup, and this was the hypothesis tested in TBCRC023. The 24-week arm was powered to detect an increase in pCR based on a historical control (TBCRC006), and numeric superiority but not direct comparison with the 12-week arm. pCR rate was lower than expected in both study arms, and as such the study did not meet its primary endpoint. Despite that fact, our findings showed that longer duration of therapy (24 weeks) resulted in a numerically higher pCR rate in patients with HER2-positive and ER-positive disease (33% vs. 9%, a difference of 24%; 95% confidence interval, 5.3%–42.3%).

Symmans and colleagues, demonstrated that patients with small amounts of residual disease (RCB-I) have a less favorable outcome than those with pCR (RCB-0) in response to chemotherapy plus anti-HER2 therapy (18). Our results suggest that longer duration therapy may convert some cases with small amounts of residual disease into pCRs. Our findings also argue that optimal duration of therapy for different subgroups (ER-positive vs. ER-negative) in HER2-positive breast cancer is different, and that the ER-positive subset appears to benefit from a longer duration of therapy. This is also supported by findings from other recent adjuvant therapy clinical trials in HER2-positive breast cancer. The ExteNET trial showed that the ER-positive subgroup appears to derive a larger benefit from longer duration of HER2-targeted therapy with the sequential use of neratinib after trastuzumab for a total of 24 months (19).

In the past few years, several chemotherapy-based neoadjuvant studies reported an increased rate of pCR with dual anti-HER2 therapy over single-agent anti-HER2 therapy (11–14). The second generation adjuvant HER2-targeted trials have thus far showed only a modest improvement in survival with the addition of a second anti-HER2 agent to trastuzumab, and this is in great part due to the inability to enrich traditional postoperative adjuvant trials with the patients most likely to benefit from such strategies. Hence, there are now significant efforts to design trial enrichment strategies that help more carefully select patients who might be candidates for escalation and deescalation strategies, such as those pursued in our studies of HER2-targeted therapy without chemotherapy and other trials (10, 13, 20, 21).

The increase in pCR with prolonged treatment needs to be weighed against a higher progression rate in the 24-week arm over the 12-week arm (11% vs. 3%). This supports the need to develop a strategy to select or enrich for patients who are likely to benefit from deescalation, and in whom chemotherapy may be safely omitted without compromising outcomes.

Identifying the patient population who may benefit from deescalated treatment may primarily depend on the biology of each tumor. However, equally important is the consideration of patient characteristics because some patients, particularly the elderly and patients with comorbidities, may be less tolerant to chemotherapy. These patients may be good candidates for deescalated treatment approaches. Such approaches may also be studied in lower risk tumors (e.g., stage I disease) where outcomes are very favorable and less toxic regimens may be desirable. These patients are generally considered good candidates for deescalation, as seen in the APT trial (22). Interestingly, the WSG-ADAPT trial testing T-DM1 in early HER2-positive/ER-positive breast cancer, with or without endocrine therapy and no systemic chemotherapy, reported an overall pCR rate of >40%, although adding endocrine therapy to T-DM1 did not offer added benefit (23). While these trials suggest the clinical prospect of deescalated treatment regimens, currently chemotherapy has a prominent and undeniable role in disease management, when considering the entire population of HER2-positive breast cancer. As such, patients who have higher clinical risk disease or unfavorable biology, who would probably benefit from chemotherapy, should receive it. Treatment strategies therefore need to be tailored to the clinical and biologic characteristics of each tumor and each patient. One size does not fit all.

HER2-positive breast cancer is increasingly being recognized as a heterogeneous collection of tumors with differential responses to HER2-targeted therapy. Characterizing the molecular and biologic heterogeneity of HER2-positive breast cancer through correlative studies using tumor specimens from chemotherapy-sparing clinical trials, such as ours, will help identify the subset of tumors that are truly dependent on HER2 and will therefore benefit the most from HER2-targeted therapy alone. Intrinsic subtype, deregulation of signaling pathways like PI3K, and immune infiltrates have been linked to treatment response and resistance (10, 13, 24, 25). Indeed, our recent TBCRC006 correlative biomarker study showed that high HER2 amplification levels combined with an intact PI3K pathway, defined by wild-type PIK3CA or high PTEN expression, identifies a subset of HER2-positive breast cancers that may be sensitive to HER2-targeted therapy alone, without chemotherapy (26). Furthermore, in a parallel correlative RNA-based study using tumors from three consecutive neoadjuvant trials in HER2-positive breast cancer with lapatinib plus trastuzumab, but without chemotherapy, including the TBCRC023, we have recently shown in tumors defined as HER2-positive by standard methods, that the HER2-enriched subtype (by nCounter PAM50) together with high ERBB2 mRNA levels predict higher pathologic response compared with HER2-positive tumors that are not HER2-enriched and/or do not have high ERBB2 mRNA levels (27, 28). Greater knowledge about these molecular and pathologic features is therefore expected to influence the design of future clinical trials.

Our trial has limitations in that it was not powered for a direct comparison between treatment arms, and its results need to be validated in an adequately powered larger trial. However, if validated, our data suggest that longer duration of treatment with targeted therapy alone (along with endocrine therapy if ER positive) without chemotherapy increases the likelihood of pCR, especially in ER-positive patients. Correlative studies on tissue samples obtained over the course of this trial will help elucidate the mechanisms of sensitivity and resistance and will influence the next generation of clinical trials that employ a molecular triage approach for deescalation of treatment. Identifying upfront those patients who may be spared chemotherapy and those who may benefit from it will lead to more tailored treatment recommendations based on the biology of the disease, and appropriate therapy escalation or deescalation.

M.F. Rimawi is an employee/paid consultant for Genentech, Macrogenics, Daiichi Sankyo, and Novartis, and reports receiving commercial research grants from Pfizer and GlaxoSmithKline. R. Nanda is an employee/paid consultant for and reports receiving commercial research grants from Genentech. I. Krop is an employee/paid consultant for Genentech, Daiichi/Sankyo, and Macrogenics, and reports receiving commercial research grants from Genentech. M.P. Goetz reports receiving commercial research grants from Pfizer, Lilly, Sermonix, and Novartis. J.R. Nangia reports receiving other commercial research support to her institution from Paxman Coolers Ltd. R. Schiff is an employee/paid consultant for Macrogenics, and reports receiving commercial research grants from AstraZeneca, GlaxoSmithKline, Gilead Sciences, and PUMA Biotechnology. C.K. Osborne is an unpaid consultant/advisory board member for Eli Lilly, Genentech, AstraZeneca, Tolmar, and Pfizer, and reports receiving other remuneration from Omelveny and Myers. No potential conflicts of interest were disclosed by the other authors.

Conception and design: M.F. Rimawi, P. Niravath, A. Forero, A.M. Storniolo, I. Krop, R. Schiff, S.G. Hilsenbeck, C.K. Osborne

Development of methodology: M.F. Rimawi, A.C. Wolff, R. Schiff, C.K. Osborne

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): M.F. Rimawi, P. Niravath, B.N. Rexer, A. Forero, A.C. Wolff, R. Nanda, A.M. Storniolo, M.P. Goetz, J.R. Nangia, S. Jiralerspong, A. Pavlick, C. Gutierrez, C.K. Osborne

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): M.F. Rimawi, P. Niravath, T. Wang, A. Forero, R. Nanda, M.P. Goetz, S. Jiralerspong, J. Veeraraghavan, R. Schiff, S.G. Hilsenbeck, C.K. Osborne

Writing, review, and/or revision of the manuscript: M.F. Rimawi, P. Niravath, T. Wang, A. Forero, A.C. Wolff, R. Nanda, A.M. Storniolo, I. Krop, M.P. Goetz, J.R. Nangia, S. Jiralerspong, A. Pavlick, J. Veeraraghavan, C. De Angelis, R. Schiff, S.G. Hilsenbeck, C.K. Osborne

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): R. Nanda, A. Pavlick, C.K. Osborne

Study supervision: M.F. Rimawi, R. Nanda, J.R. Nangia, A. Pavlick, C.K. Osborne

This work was supported in part by NCI grants P50 CA186784 (SPORE) and P30 CA125123; DOD grants W81XWH-17-1-0579 (to M.F. Rimawi) and W81XWH-17-1-0580 (to R. Schiff); the Breast Cancer Research Foundation grants BCRF-16-142, 17-143, and 18-145 (to R. Schiff and C.K. Osborne), as well as grants from Komen Foundation for the Cure, Avon Foundation, Breast Cancer Research Foundation, and a research grant from GSK.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.