Trastuzumab and Pertuzumab in Patients with non-Breast/Gastroesophageal HER2-amplified Tumors: Results from the NCI-MATCH ECOG-ACRIN Trial (EAY131) Sub-protocol J

Abstract

Purpose:

NCI-MATCH assigned patients with advanced cancer and progression on prior treatment, based on genomic alterations in pre-treatment tumor tissue. Arm J (EAY131-J) evaluated the combination of trastuzumab/pertuzumab (HP) across HER2-amplified tumors.

Patients and Methods:

Eligible patients had high levels of HER2 amplification (copy number [CN] ≥7) detected by central next generation sequencing (NGS) or through NCI designated laboratories. Patients with breast/gastroesophageal adenocarcinoma, and those who received prior HER2-directed therapy were excluded. Enrollment of patients with colorectal cancer was capped at 4 based on emerging data. Patients received HP IV Q3 weeks until progression or unacceptable toxicity. Primary endpoint was objective response rate (ORR); secondary endpoints included progression-free survival (PFS) and overall survival (OS).

Results:

Thirty-five patients were enrolled, with 25 included in the primary efficacy analysis (CN ≥7 confirmed by central lab, median CN=28). Median age was 66 (range 31–80), and half of all patients had ≥3 prior therapies (range 1– 11). The confirmed ORR was 12% (3/25 partial responses [colorectal, cholangiocarcinoma, urothelial cancers], 90% CI 3.4–28.2%). There was one additional partial response (urothelial cancer) in a patient with unconfirmed ERBB2 copy number. Median PFS was 3.3 months (90% CI 2.0–4.1) and median OS 9.4 months (90% CI 5.0–18.9). Treatment-emergent adverse events were consistent with prior studies. There was no association between HER2 CN and response.

Conclusion:

HP was active in a selection of HER2-amplified tumors (non-breast/gastroesophageal) but did not meet the predefined efficacy benchmark. Additional strategies targeting HER2 and potential resistance pathways are warranted, especially in rare tumors.

Translational Relevance (128/150)

High level amplification of the human epidermal growth factor receptor 2 (HER2) receptor across tumor types may indicate a heavier biological dependence on this receptor and pathway than in tumors without such amplification. Trastuzumab and pertuzumab (HP) has an established role in HER2-amplified breast cancer. We hypothesized that HP would be active across HER2- amplified tumors, other than breast and gastroesophageal adenocarcinoma, and conducted a single-arm phase II trial to characterize the efficacy and safety in this selected patient population. HP was active in a selection of rare tumors (confirmed overall response rate 12%, 90% CI 3.4–28.2%) and did not meet the predefined efficacy benchmark; there were no new safety signals. Translational biomarker exploration, as well as additional strategies targeting HER2 and potential resistance pathways, are warranted.

Introduction

Human epidermal growth factor receptor 2 (HER2)-directed therapies are approved for the treatment of patients with HER2-overexpressing breast and gastroesophageal adenocarcinoma, with associated survival advantages.(1,2) A variety of other tumor types also demonstrate amplification of the HER2 receptor, which may indicate a putative driving role and heavier biological dependence of these tumors on this pathway compared with those tumors without such amplification.(3) Resistance to trastuzumab, the first HER2-directed therapy approved for clinical use, is a common clinical challenge; with dysregulation of downstream signaling pathways, alternative receptor tyrosine kinase signaling and accumulation of the truncated active kinase p95-HER2 all being implicated.(4)

Based on extensive preclinical and clinical development, additional HER2-targeted therapies are now available and include the HER2/HER3 dimerization inhibitor pertuzumab.(5) Dual HER2-blockade with trastuzumab and pertuzumab (HP) has been shown to be more efficacious than trastuzumab alone in both early and late-stage breast cancer.(6–8) High tumor response rates (30–35%) have also been observed with HP in HER2-amplified colorectal cancer.(9) We thus hypothesized that HP would result in meaningful clinical benefit in patients with HER2-overexpressing advanced cancers where regulatory approval was not already obtained.

The ‘National Cancer Institute - Molecular Analysis for Therapy Choice’ (NCI-MATCH) trial is a signal-finding precision medicine study that incorporates genomic tumor testing to direct patients with refractory cancer to molecularly targeted treatments (NCT02465060).(10,11) Subprotocol Arm J evaluated the concurrent combination of HP across HER2-amplified tumors (copy number, CN ≥7), other than breast and gastric/GEJ adenocarcinoma.(12) A cutoff of 7 was chosen based on the fact that the use of next generation sequencing (NGS) to determine copy number variation was fairly new at the time the study began, and this was a conservative estimate.

Patients and Methods

Study Design and Patient Population

The NCI-MATCH master protocol is a multi-arm, molecular profile–driven, non-randomized phase II trial (EAY131, Supplementary Figure 1), developed by the ECOG-ACRIN Cancer Research Group and the NCI.(11) Progression following at least one line of standard systemic therapy and no other standard therapy available that has been shown to prolong overall survival (OS) OR where no standard treatment exists that has been shown to prolong OS was required for eligibility to the screening phase of the master protocol (Supplementary Methods). Consenting patients had their tumors molecularly characterized by NGS performed centrally or through NCI designated laboratories. Patients were required to have either a core biopsy or available archived tumor tissue obtained within the last 6 months for assessment. Targeted NGS was performed using the Ion Torrent Oncomine AmpliSeqTM panel of 143 genes and was supplemented with Phosphatase and tensin homolog (PTEN), MutS homolog (MSH) and MutL homolog (MLH) immunohistochemistry (IHC).(10) Patients in a particular molecularly-defined subgroup received treatment matched to that tumor tissue alteration and proceeded to the treatment phase (subprotocol arms).

Participants who met eligibility requirements from the master protocol, and with HER2 amplification as established in the analytical assay (CN ≥7)(10) detected by NGS, were considered for the Arm J subprotocol. Measurable disease by Response Evaluation Criteria in Solid Tumors (RECIST) version 1.1.(13), ECOG performance status 0–1, and left ventricular ejection fraction (LVEF) ≥ institutional lower limit normal were also required. Those with breast/gastroesophageal adenocarcinoma, and those who received prior HER2-directed therapy, were excluded. Enrollment of participants with colorectal cancer was capped at 4 based on emerging data.(9,14)

The study was performed in accordance with provisions of the Declaration of Helsinki and Good Clinical Practice guidelines. The study (ClinicalTrials.gov identifier: NCT06136897) was approved by the NCI Central Institutional Review Board and by the institutional review board of record for all participating institutions. All patients signed a written informed consent prior to participation.

Treatment and evaluation

Patients were treated with the standard intravenous dosing of trastuzumab and pertuzumab every 3 weeks for a 21-day cycle, until progression or unacceptable toxicity. Pertuzumab was administered before trastuzumab each cycle; 840mg loading dose and 420 mg thereafter. Trastuzumab loading dose was 8mg/kg then 6mg/kg thereafter. Restaging scans were performed every three cycles for the first 33 cycles and every four cycles thereafter. Echocardiography was required at baseline and every 3 months thereafter.

Common Terminology Criteria for Adverse Events (CTCAE, version 4.0) were used to grade treatment-related toxicity. Dose modifications of study agents were not permitted. Dose delays and discontinuation of study therapy were permitted as outlined in the study protocol (Supplementary Methods).

Statistical Considerations

The primary endpoint was objective response rate (ORR), defined as rate of complete or partial response per RECIST version 1.1. Allowing for a 10% ineligibility rate, the accrual goal was 35 patients for each subprotocol. The study design had the operating characteristics of at least 92% power to distinguish an ORR of 25% from a null of 5% with one-sided Type 1 error of 1.8%. If the ORR was ≥ 5/31 (16%), the study therapy would be deemed promising and worthy of further evaluation. Eligible and treated patients enrolled on basis of MATCH central NGS assay, or outside assays with molecular abnormalities confirmed by MATCH assay were included in the primary analyses. The null hypothesis is tested at one-sided significance level of 1.8% for arms with more than 31 patients in the primary analysis, and at 5% for arms with less than 31 patients in the primary analysis. Since there are 25 patients in the primary analysis for Arm J, a minimum of 4 responders were needed to consider HP worthy of further evaluation in patients with HER2 amplifications.

Secondary endpoints included progression-free survival (PFS), 6-month PFS rate, toxicity, and potential prognostic/predictive biomarkers. Exploratory endpoints included correlating response with HER2 CN by NGS. Data cutoff was November 3, 2021. All statistical analyses were performed with R, version 4.1.0 (R Foundation for Statstical Computing).

Data Availability Statement

The data underlying this article is available from the authors upon request and will be made available for request from the NCTN/NCORP Data Archive (https://nctn-data-archive.nci.nih.gov/) upon completing a Data Request Form for data from NCT06136897.

Results

Patient Characteristics

From March 2017 to June 2019, thirty-five patients were enrolled in the study (Figure 1) with histopathologic tumor classification presented in Supplementary Table 3. All enrolled patients started protocol therapy, although three subjects were found to be ineligible due to low creatinine clearance (n=2) and presence of an intra-cardiac defibrillator (n=1). Of 32 analyzable patients (Supplementary Table 2), 25 had CN ≥7 confirmed by central lab (median CN=21, range 7.3 – 84.0) and were included in primary efficacy analysis of ORR per protocol (Primary Analysis Population, Table 1). Of these 25 patients, 11 were enrolled based on the central MATCH assay and 14 patients on the basis of results from NCI designated laboratories (outside assays). The MATCH assay case status is summarized in Supplementary Table 1.

Figure 1:

CONSORT Diagram

Table 1:

Patient Characteristics (Primary Analysis Population)

Patient Characteristic (n=25) N (%)
Median Age (range)	Years	66 (31–80)
Sex	Female	14 (56)
	Male	11 (44)
Race	White	22 (88)
	Black	3 (12)
ECOG	0	7 (28)
	1	18 (72)
Tumor Type	Gynecologic	11 (44)
	Gastrointestinal	11 (44)
	Genitourinary	2 (8)
	Head and neck	1 (4)
Lines of Prior Therapy	1	3 (12)
	2	9 (36)
	3	6 (24)
	≥4	7 (28)

In the primary analysis population, median age was 66 (range 31–80), 56% were female, and half of all patients had ≥3 prior therapies (range 1– 11) (Table 1). The most common tumor types included gynecologic (n=11) and gastrointestinal (n=11) cancers (Supplementary Table 3b). Gynecologic malignancies by histopathologic tumor classification were heterogeneous as presented in Supplementary Table 3b, and gastrointestinal cancers included colorectal adenocarcinoma (n=4) and cholangiocarcinoma (n=3).

Treatment & Safety

The median number of treatment cycles received was 4 (range 1–39) for the 25 patients in the primary analysis. At the data cutoff date (November 2021), one patient was still on treatment. The main reason for discontinuation of protocol therapy was disease progression (24/32, 75%), with only 2 patients discontinuing due to AEs. AE analysis included all treated patients. Tables 2 and A4 summarize the treatment-emergent adverse events, which were consistent with prior HP studies. There were two reported instances of Grade 3 heart failure, a known treatment-related toxicity for HP, but no treatment-related grade 4 or 5 events. The most frequent all grade treatment-emergent AEs included diarrhea (31%), infusion-related reaction (28% [majority grade 1/2]) and fatigue (23%, Supplementary Table 4).

Table 2:

Treatment Related Adverse Events

Toxicity (n=35)	Grade 3	Grade 4/5
Abdominal Pain	1	0
Alkaline Phosphatase increased	1	0
Dyspnea	1	0
Fatigue	1	0
Heart Failure	2	0
Hepatobiliary-other	1	0
Hypophosphatemia	1	0
Hypertension	1	0
Hypotension	1	0
Hypoxia	1	0
Infusion Reaction	1	0
Lymphocyte count decreased	1	0
Total	8	0

^*Toxicities at least possibly related to treatment reported via Case Report Forms

Efficacy

Eligible and treated patients enrolled based on the central MATCH assay OR outside assays with molecular abnormalities confirmed by MATCH assay were included in the primary efficacy analyses (n=25). The confirmed ORR in this population was 12% (3/25 partial responses [colorectal, cholangiocarcinoma and urothelial cancers], 90% CI 3.4–28.2%, Figure 2 and Supplementary Table 6), with a median duration of response of 5.2 months (Figure 3). There was one additional partial response (PR) in a patient with urothelial cancer whose ERBB2 copy number (outside assay) was not confirmed by MATCH assay. Nine additional patients in the primary analysis population had evidence of stable disease, another nine with progressive disease, and four patients who were not evaluable due to lack of follow-up assessment. For the overall cohort of 32 patients, the ORR was 12.5% (90% CI 4.4–26.4%). The median PFS for the 25 patients in the primary analysis was 3.3 months (90% CI 2.0 – 4.1), the 6-month PFS was 25.3% (90% CI 10.7 – 40.0), and the median OS was 9.4 months (90% CI 5.0 – 18.9) (Supplementary Figure 4). The median PFS amongst all 32 analyzable patients was 2.1 months (90% CI 1.8 – 4.1), the 6-month PFS was 22.9% (90% CI 10.4 – 35.4), and the median OS was 8.7 months (90% CI 5.0 – 13.8) (Supplementary Figure 5), with median follow up period for OS of 24.1 months (range 8.2–35.6).

Figure 2:

Waterfall plot of best percentage change from baseline in 21 patients with centrally confirmed HER2 result (primary analysis population) and evaluable measurements. This plot excludes patients who did not have follow-up tumor measurements available (n=4). PD, progressive disease; PR, partial response; SD, stable disease.

Figure 3:

Swimmers plot of duration of treatment in patients who achieved partial response (PR) or stable disease (n=12) and had centrally confirmed HER2 result (primary analysis population).

PD, progressive disease.

Although no complete responses were observed in this NCI MATCH treatment arm, the patient with colorectal adenocarcinoma who achieved PR had 99% reduction from baseline on imaging (Figure 2, Supplementary Figure 2). This patient was enrolled prior to the amendment excluding additional patients with colorectal cancer. Molecular profiling revealed intact mismatch repair and mutations in APC, SMAD4, FBXW7 and TP53, but not BRAF/KRAS, and HER2 copy number was 41. The patient with cholangiocarcinoma had 59% tumor shrinkage from baseline, remained on therapy for > 15 cycles and harbored SMAD4, TP53, CDK12 mutations, with HER2 copy number of 18. The patients with urothelial cancer who achieved PR harbored CCNE1 copy number variant (best tumor shrinkage 62.5%, HER2 copy number 12) and TP53 mutation (best tumor shrinkage 47.5%, HER2 copy number 82 but uncomfirmed). A number of other durable responses and stable disease were observed, including in patients with endocervical papillary adenocarcinoma and gallbladder cancer (Figure 3, Supplementary Figure 3).

Correlative Analysis

There was no significant association observed between HER2 copy number and treatment response, with results from the primary analysis population presented in Supplementary Figure 6 (p=0.87 for comparison of PR versus not). Co-occurring mutation data were available for the 25 patients included in the primary analysis (25/32). In this cohort, all 25 patients were identified as having an additional genetic alteration, alongside the HER2 amplification (CN ≥7) required for eligibility. The most common co-occurring mutation was a mutation in TP53, detected in 21 patients. SMAD4 mutation was present in two patients with PR (Figure 4), including the patient with colorectal adenocarcinoma noted above and a patient with cholangiocarcinoma. Four patients had co-occurring mutations in PIK3CA, and no tumors had deficient mismatch repair/high levels of microsatellite instability. Co-occurring mutation data for the overall cohort is also presented in Supplementary Figure 7.

Figure 4.

Co-occurring genomic alterations with HER2 using the central NCI-MATCH assay for 25 patients. Single nucleotide variant (purple), copy number variant (dark red), insertion-deletion (light red). Response data and histology for individual patients are listed at the bottom.; PD, progressive disease; PR, partial response; SD, stable disease; UE, unevaluable.

Discussion

The Arm J (EAY131-J) NCI-MATCH subprotocol investigating the combination of dual HER2-directed therapy with HP across HER2-amplified advanced tumors (non-breast/gastroesophageal) did not meet the predefined efficacy benchmark for ORR of 16% (4/25) required to reject the null ORR of 5%. Three patients achieved confirmed partial responses (colorectal adenocarcinoma, cholangiocarcinoma, and urothelial cancer), with a 4^th tumor response unconfirmed, thus, falling just short of considering HP worthy of further evaluation. Although the patient with colorectal carcinoma achieving partial response was noted to have a high HER2 copy number of 41, overall, there was no significant correlation between HER2 copy number and treatment response.

As this treatment arm of the NCI MATCH trial was starting to accrue, emerging data supporting the role of HER2-directed therapy in HER2-overexpressing or amplified advanced colorectal cancer was noted. In MyPathway, an industry-sponsored non-randomised phase 2a multiple basket study, the investigators reported an ORR of 32% in this cohort with HP alone.(9,15) The HERACLES trial also reported a 30% ORR in patients with heavily pre-treated HER2-positive and RAS wild type metastatic colorectal cancer with the combination of trastuzumab and lapatinib.(14) This emerging data led to capping of enrollment of patients with colorectal cancer to this Arm J (EAY131-J) of the NCI-MATCH subprotocol to ensure enrollment slots were available for cancer types not known to respond to HP, in keeping with the overall goal of NCI MATCH. This decision may also have contributed to this Arm of the MATCH trial not meeting its predefined efficacy objective in a heterogenous advanced solid tumor cohort where new efficacy signals were being sought. Differences in enrolment criteria between studies investigating HP in advanced solid tumors should also be considered. The MyPathway trial enrolled patients based on HER2 overexpression by immunohistochemistry 3+ staining or HER2 amplification based on in-situ hybridization (ISH) or NGS in archival or fresh tumor tissue samples (most recent biopsy recommended), supplemented with central NGS testing following enrollment where archival tissue was available; whereas the NCI MATCH eligibility was based on NGS from fresh tumor biopsies at the time of screening.

Investigation of biomarkers of response and resistance to HER2-directed therapies is a research priority of the oncology community. In this small cohort, there was no association observed between HER2 copy number and treatment response, a biomarker of increasing interest in the breast cancer literature.(16) The selection of CN≥7 for HER2 gene was based on early feasibility data using the Oncomine assay that suggested excellent sensitivity and specificity using this CN cut-off, resulting in a consensus-based decision albeit without focused clinical validation at the time.(10,12) That traditional biomarkers of response to HER2-directed therapies were not assessed in this study, including protein overexpression by IHC, is a potential limitation of the study design.(17) It may be that tumor type specificity is more important than previously thought to determine sensitivity to dual HER2-directed therapies, when one considers the excellent outcomes with these therapies in responsive cancers. However, it should be acknowledged that even in cancer types felt to be responsive to dual HER2-directed therapies such as breast cancer, the ORR to HP is approximately 24%; an outcome not too different to what has been presented in our cohort. Additional limitations include the relatively small sample size and single arm non-randomized design, potential selection bias and confounders, heterogeneity of tumor types enrolled, variability of laboratory assay algorithms to assess HER2 CN, and inadequate molecular profiling of tumor tissue at the time of progression. Furthermore, differences in assay algorithms and adjustment for cellularity led to the need to adjust for HER2 CN with some of the NCI designated laboratories.

Other putative predictive biomarkers of response to HER2-directed therapies of interest include gene expression signatures such as intrinsic subtyping, novel “immune signatures”, the presence of tumor infiltrating lymphocytes, and other characteristics of the tumor microenvironment.(18,19) Dysregulation of the downstream PI3K pathway may also lead to constitutive activation of the PI3K/AKT pathway leading to treatment resistance, as one example of downstream mutational activation leading to lower response rates to anticancer targeted therapies.(20) Co-occurring PIK3CA mutations were detected in four patients in our study, all with gynecologic malignancies (adenocarcinoma cervix, carcinosarcoma uterus, vulvar cancer, uterine cancer) and none of whom responded to treatment. The most common co-occurring mutation overall was a mutation in TP53, with SMAD4 co-mutations detected in two patients with PR. Additional translational studies in this cohort (using tumor tissue and cell-free circulating tumor DNA) will focus further on potential mechanisms of response and resistance, acknowledging the small sample size and heterogeneous tumor types enrolled.

Additional strategies may be required where tumor heterogeneity and complex resistance mechanisms may preclude response to relatively well tolerated regimens such as HP. The addition of chemotherapy to a targeted therapy backbone may be required for optimal treatment effect in select cancer types.(6,21) Novel antibody-drug conjugates such as ado-trastuzumab emtansine (T-DM1) improve outcomes in both the early- and late-stage breast cancer setting.(22) Disappointing results were observed, however, in subprotocol Q of the NCI MATCH trial where the primary endpoint for ORR was not met for use of T-DM1 in a heavily pre-treated, mixed histology cohort.(23) More recently, trastuzumab deruxtecan has shown impressive results in the advanced disease setting across breast, lung, and GI cancers.(24,25) The impressive efficacy of such antibody-drug conjugates may relate to a “bystander effect” whereby nearby non-target cancer cells may be impacted by diffusion of free cytotoxic moiety into the surrounding tissue.

In conclusion, dual HER2-directed therapy with HP had activity in a minority of non-breast/gastroesophageal HER2-amplified tumors, noting that the majority of these responses occurred in rare tumor types. Additional strategies targeting HER2 and potential resistance pathways are warranted in cancers that may not respond to dual HER2-directed therapy. Based on our results, it may be most pertinent to focus on rare cancer subtypes which is an area of unmet medical need.

Abbreviations List:

CTCAE

Common Terminology Criteria for Adverse Events

CN

copy number

HER2

human epidermal growth factor receptor 2

HP

Trastuzumab and pertuzumab

IHC

immunohistochemistry

ISH

in-situ hybridization

LVEF

left ventricular ejection fraction

MSH

MutS homolog

MLH

MutL homolog

NGS

next generation sequencing

ORR

objective response rate

OS

overall survival

PR

partial response

PTEN

Phosphatase and tensin homolog

PFS

progression-free survival

RECIST

Response Evaluation Criteria in Solid Tumors