A Multicenter Phase II Trial of Ipilimumab and Nivolumab in Unresectable or Metastatic Metaplastic Breast Cancer: Cohort 36 of Dual Anti–CTLA-4 and Anti–PD-1 Blockade in Rare Tumors (DART, SWOG S1609)

Metaplastic breast cancers (MpBC) are rare (∼1% of breast cancers) and very aggressive tumors, typically composed of both an adenocarcinoma and a metaplastic component (squamous, chondroid, spindle, rhabdoid, or osseous, typically of same clonal origin as ductal carcinoma component; refs. 1–3). MpBC has a poor response to standard cytotoxic therapies (4), and a median survival of 8 months for metastatic disease, which is significantly worse than that of nonmetaplastic triple-negative breast cancer (TNBC; ref. 5).

The molecular signature of MpBCs has similarities to the claudin-low and mesenchymal subtypes of TNBC. There is an enrichment of stem cell–associated genes including genes involved in epithelial‐to‐mesenchymal transition (6). Genomic studies have found that MpBC may have amplification of epidermal growth factor receptor, as well as alterations in genes involved in the PI3K/Akt pathway, Wnt/β-catenin signaling, and cell-cycle dysregulation (7).

Frequent overexpression of PD-L1 was recently demonstrated in primary MpBC, with PD-L1 positivity in tumor cells (8). Tumor-infiltrating lymphocytes (TIL) have also been demonstrated in MpBC, including PD-1–expressing TILs (8, 9), suggestive of an immunogenic cancer phenotype in some patients. These findings and preliminary evidence of clinical activity of PD-1 blockade in MpBC (10) led to the inclusion of MpBC into the DART study as cohort 36.

Here, we present the results of the metaplastic breast cancer cohort on DART (DUAL ANTI–CTLA-4 AND ANTI–PD-1 BLOCKADE IN RARE TUMORS, S1609), a prospective phase II study conducted through the National Cancer Institute (NCI)–supported SWOG Cancer Research Network.

DART is a multicenter (>800 sites), open label, phase II basket study (NCT02834013) of ipilimumab and nivolumab for rare malignancies. It is being conducted by the Early Therapeutics and Rare Cancers Committee under the auspices of SWOG and the NCI. The Cancer Therapy Evaluation Program (CTEP) provided study medication under an NCI Cooperative Research and Development Agreement (CRADA) with Bristol–Myers Squibb (BMS). The study was conducted in accordance with the Declaration of Helsinki. The trial design and eligibility criteria for DART were previously reported (11). All participants provided written informed consent authorized by each enrolling center's internal review board.

Eligible patients for cohort 36 must have had a histologically confirmed diagnosis of MpBC with disease measurable as per RECIST v1.1 (12). Enrollment was directed into study cohorts (baskets) based on the local pathology report (the DART trial did not mandate central review/verification of the rare histologies). All patients' cancers had progressed following at least one line of standard therapy and there must not have been other approved/standard therapy available that has been shown to prolong overall survival. Patients may have received either prior anti–CTLA-4 or prior anti–PD-1/anti–PD-L1 therapy (but not both), provided that it was completed at least 4 weeks prior to registration. Patients were required to have an Eastern Cooperative Oncology Group performance status (ECOG PS) 0–2, be at least 18 years of age, and have adequate organ function, within specific hematologic, renal, hepatic, adrenal, and thyroid parameters. Exclusion criteria included certain autoimmune diseases and ongoing grade 3/4 immune-related adverse events (irAE). For patients with brain metastases, central nervous system (CNS)-directed therapy must have been completed ≥28 days prior to registration and patients must have been off steroids for at least 7 days with stable disease at time of registration.

Patients received nivolumab 240 mg every 2 weeks and ipilimumab 1 mg/kg every 6 weeks (both intravenously, one cycle is 6 weeks). Disease assessments were performed at baseline and thereafter at weeks 8, 16, 24, and then every 12 weeks. Treatment continued until tumor progression, unacceptable toxicity, or withdrawal of consent.

The primary endpoint of the study was objective response rate [ORR; confirmed complete and partial response (CR and PR, respectively)] as assessed by RECIST 1.1 criteria (12). The regimen was considered of interest for further study if the true ORR was 12% or higher (2 responses out of 16 eligible patients). Subset analyses within the cohort were not prespecified.

As previously described, all cohorts used a two-stage design (13). If ≥1 response was observed in the first 6 eligible and evaluable patients, accrual to the second stage to a total of 16 patients would be opened. Two or more responses out of 16 patients were considered evidence that the treatment regimen merits further investigation, provided other data including AEs also appeared satisfactory. This design has 87% power (under an alternative response rate of 30%) with a one-sided alpha of 13% (null response rate assumed to be 5%) in each stratum.

Secondary endpoints were toxicity (per CTCAE version 4), overall survival (OS), and progression-free survival (PFS); PFS was equal for RECIST (12) and iRECIST (14) in this cohort. PFS and OS estimates were calculated utilizing the Kaplan–Meier method (15) and compared using log-rank tests. Confidence intervals (CI) for medians were built using the method of Brookmeyer and Crowley (16); CI for point estimates were calculated employing the log–log transformation. CIs for the primary ORR analysis accounted for the two-stage design and the observed sample size of 17 patients (17). All analyses were performed using R version 4.0.1.

Overall, 19 patients from 17 National Clinical Trial Network (NCTN) institutions were registered for cohort 36, with 17 patients meeting eligibility criteria and receiving protocol therapy (CONSORT Supplementary Fig. S1). Enrollment was rapid, with 8 eligible patients enrolled January to May 2018 (first stage) followed by a temporary study hold to analyze responses to determine proceeding to the second stage, followed by enrollment of an additional 9 eligible patients from October 2018 to April 2019.

Patient demographics and tumor characteristics are listed in Table 1. The median age was 60 years (range, 26–85 years); all patients were female and the majority Caucasian. As expected for MpBC, the majority of tumors were TNBC, high grade, and exhibited high proliferation. Patients had received a median number of two prior lines of systemic therapies, including standard chemotherapies (anthracycline, taxanes, eribulin mesylate, carboplatin, cisplatin, capecitabine), angiogenesis inhibitors, mTOR inhibitors, anti–PD-1 inhibitors, BET inhibitors, HDAC inhibitors, and other investigational agents.

Treatment-related AEs are summarized in Table 2. Overall, 11 patients (65%) experienced an AE, with 3 (18%) having a grade 3–4 AE, and 1 grade 5 AE (unknown cause, possibly related, further detail in Table 2). Altogether, 47% of participants experienced an irAE; the most common were liver function test (LFT) abnormalities, adrenal insufficiency, and rash.

Efficacy results are shown in Table 3. Of 17 enrolled and eligible patients, all of whom had measurable disease, three patients had confirmed objective responses by RECIST 1.1, resulting in an ORR of 18% (95% CI, 6%–40%). Notably, all three responses have been durable and have been ongoing at 28+, 33+, and 34+ months, respectively, and are therefore considered exceptional (Fig. 1). These responses were observed in spindle cell MpBC (n = 2) and chondromyxoid MpBC (n = 1). Nonresponders had poor outcomes; stable disease (SD) was seen in 18%, none lasting > 6 months. Median PFS and OS were 2 and 12 months, respectively (Fig. 1). Median follow-up among patients who are alive is 33 months. Cut-off date of data follow-up is February 4, 2021. Three patients had received prior anti–PD-1 therapy, none of whom had a tumor response.

Exceptional responders are described in more detail in Supplementary Table S1. All three responders had chemotherapy-refractory disease with significant tumor burden. The tumors' baseline target lesion sum ranged from 6.1 to 11.3 cm and cancers had recurred within one year of a taxane and/or anthracycline-containing regimen, or progressed on it. None had received prior immunotherapy. Reassuringly, responses persisted despite stopping ipilimumab in one patient and ipilimumab + nivolumab in another patient (Fig. 1). Of note, all three responders developed significant irAEs with adrenal insufficiency induced in all three.

Prespecified trial correlative studies have been delayed due to COVID-19 pandemic work restrictions. Molecular and immunohistochemistry (IHC) tumor characterization done as part of routine medical care is shown for all three responding patients in Supplementary Table S1. Tumors had low tumor mutational burden (TMB), were microsatellite stable, and two of three had negative or low PD-L1 expression/TILs.

To our knowledge, our study represents the first prospective trial of immunotherapy in MpBC, a rare subset of TNBC. This cohort of the DART trial met its primary endpoint: ipilimumab plus nivolumab was clinically active in advanced MpBC, with responses observed in 3 of 17 patients (ORR 18%). Importantly, all responses were durable (28+, 33+, and 34+ months), which is rarely observed in MpBC.

Advanced MpBC has a poor prognosis with median OS less than one year despite chemotherapy (18); therefore, new treatment strategies are urgently needed. MpBCs harbor a wide variety of genomic alterations, the most frequent being in the TP53 and PIK3CA genes (9). Therapies targeting the PI3K/AKT/mTOR pathway and antiangiogenesis agents have shown objective responses in a subset of patients (18–21). The limited treatment options and poor prognosis of MpBC along with the high tumoral PD-L1 expression and TIL presence in some patients (8) provided the rationale to study immunotherapies in this subtype.

Objective responses to dual immunotherapy were seen in three patients, two with spindle cell and one with chondromyxoid histology. All three women had chemotherapy-refractory disease with significant tumor burden (range of baseline target lesion sum 6.1–11.3 cm). These responses are remarkable for several reasons. First, they were durable with all responses ongoing beyond 2 years, which is in stark contrast to the short-lived responses observed with chemotherapy (18) as well as with anti–PD-1 blockade in MpBC (10). Furthermore, three of the patients in our study (all nonresponders) who had received prior therapy with anti–PD-1 (combined with platinum chemotherapy, targeted agents, or a STING agonist) rapidly progressed on that therapy, all within 2 months. Second, responses were observed in patients with highly chemotherapy-refractory disease, as evidenced by disease recurrence within one year of a taxane and/or anthracycline combination regimen or progression on it, known to be an independent poor prognostic factor for TNBC (22–24). Third, responses were observed even in tumors with negative or low PD-L1 expression and low TIL. This may suggest a contribution of the anti-CTLA blockade to the efficacy of the regimen studied, as responses to anti–PD-1/anti–PD-L1 therapeutics are enriched in PD-L1–positive tumors in metastatic TNBC (23, 25). Fourth, all three responders developed significant irAEs including panhypopituitarism and adrenal insufficiency. While irAE toxicity has been observed across immune checkpoint inhibitor regimens and has been shown to correlate with antitumor responses in some studies (26), the duration of treatment is generally longer in responders, which could be a confounder due to longer exposure to therapy. Adrenal insufficiency, for instance, was diagnosed around cycle 3 to 4 at approximately 4 months in the responders, at which time most of the nonresponder patients had come off study for progression. Fifth, our responders had low TMB. TMB can be strongly correlated with responsiveness to checkpoint blockade, with only approximately 5% of patients responding to anti–PD-1/PD-L1 agents when TMB is low (≤5 mutations/mb); however, our prior studies suggest that responses to anti–CTLA-4/anti–PD-1/PD-L1 combinations, as given in the current study, are independent of TMB (27).

While the contribution of CTLA-4 blockade to the efficacy of the anti–CTLA-4/anti–PD-1 combination and in particular the duration of responses remains unknown in our trial, we would like to reference efficacy data of anti–PD-1/PD-L1 therapy in MpBC as well as TNBC in general. Data from a recently published case series of MpBC, which included 4 patients from an investigator-initiated trial of anti–PD-1 therapy (with capecitabine or paclitaxel) as well as one patient treated off-label with pembrolizumab and bicalutamide), demonstrated objective responses in 3 patients; however, PFS was only 5.3, 5.7, and 8.0 months (28). In another investigator-initiated trial of anti–PD-1 therapy with nab-paclitaxel, 1 of 2 patients with MpBC experienced an objective response; however, PFS was only 6.7 months (Adams et al, NCT02752685, unpublished data). In phase III trials of chemo-immunotherapy for metastatic (general) TNBC, durable responses were observed; however, the percentage of patients with ongoing responses years out is very small. For instance in our final analysis of Impassion130, among patients alive at 3 years, only 16 patients had not experienced progression of disease, representing 1.77% of the 902 patients enrolled and 2.66% of patients treated with the chemo-immunotherapy combination (12/451; ref. 29).

As discussed above, responses to chemo- plus anti–PD-1 therapy observed in MpBC are typically not durable (all ≤9 months) and combinatorial immunotherapies may be required to achieve durable responses. However, it is important to note the addition of anti–CTLA-4 to anti–PD-1/PD-L1 regimens has been associated with greater toxicity and higher mortality rates (30, 31). A meta-analysis of 112 trials involving 19,217 patients showed toxicity-related fatality rates of 0.36% (anti–PD-1), 0.38% (anti–PD-L1), 1.08% (anti–CTLA-4), and 1.23% (anti–PD-1/PD-L1 plus anti–CTLA-4; ref. 30).

The side effect profile of the anti–CTLA-4/anti–PD-1 combination in our study was consistent with published combination studies (31, 32) and that observed in other cohorts of DART. No unexpected toxicities were observed, however, the fatal event and several potentially life-threatening AEs such as adrenal insufficiency highlight the importance of a thorough risk–benefit discussion with patients and their education about possible side effects before treatment start as well as careful monitoring of patients on treatment by oncologists at centers with immunotherapy experience. Furthermore, it is essential to raise awareness among emergency department physicians, critical care providers, and other specialists, especially as serious toxicities can have unusual clinical presentations and grade 5 toxicities can occur very early in the treatment course, as shown for ipilimumab combination therapies with a median 14.5 days (30). Subsequent studies should therefore carefully consider these risks and possibly modify dosing, as AEs have been shown to occur more often at higher doses of ipilimumab (30), or evaluate newer anti–CTLA-4 formulations such as probodies to widen the therapeutic window (33).

To better understand the molecular basis for responses, correlative samples were collected on study and will be analyzed as per pre-specified plan at Cancer Immune Monitoring and Analysis Centers (CIMAC) sites. Some patients had tumor NGS performed for clinical purposes outside of the DART trial and results were available as part of the medical record. Based on these local data, tumors of responders with NGS available showed low TMB and absence of microsatellite instability (MSI), which is consistent with our published large dataset of 192 MpBC demonstrating a low TMB across these tumors (median 2.7 mutations/Mb) along with microsatellite stability (0/192 MpBC were MSI high; ref. 9).

The strengths of our study include the enrollment of patients at both academic and community centers as well as support from the NCI and SWOG. Furthermore, DART served an unmet need with rare tumors and demonstrated that it was feasible to rapidly accrue even very rare tumors (11), especially because SWOG was able to mobilize >800 sites for the DART study. Limitations of the study include a relatively small sample size within a single arm design per disease cohort, and the lack of a randomized comparison with standard-of-care therapies. Central pathology review was not mandated, and therefore we relied on local site assessments that may be suboptimal for rare occurring cancers; however, review of all pathology reports was conducted by the study chair, confirming the presence of metaplastic components in specimens. Furthermore, for the three responders, digital pathology images were reviewed for assessment of TILs.

In conclusion, among 17 patients treated with nivolumab and ipilimumab, three exceptional responses were observed in chemotherapy-refractory, metastatic MpBC, all ongoing at 28+, 33+, and 34+ months. Further investigation of this combination is warranted. Of special interest, in this study, responses were dichotomized into the 18% that did remarkably well versus the others that had no significant benefit from ipilimumab and nivolumab. Such a dichotomization may suggest the presence of a unique biomarker for response in these patients and should be taken into consideration in the design of a future trial, possibly with an adaptive design. Interestingly, the exceptional responders included patients whose tumors had low TMB, low TIL, and no PD-L1 expression, indicating that the mechanism of response requires further in-depth interrogation, which is planned via collaboration with CIMAC sites.

S. Adams reports other support from BMS during the conduct of the study. M. Othus reports grants from National Institutes of Health during the conduct of the study, as well as personal fees from Merck, Biosight, Celgene, Glycomimetics, and Daiichi Sankyo outside the submitted work. S.P. Patel reports grants and personal fees from BMS during the conduct of the study, as well as grants and personal fees from Amgen, AstraZeneca, Bristol Myers Squibb, Certis, Eli Lilly, Genentech, Illumina, Merck, Pfizer, Rakuten, and Tempus outside the submitted work. K.D. Miller reports other support from Merck, Roche/Genentech, and AstraZeneca, as well as grants from Pfizer and Astex outside the submitted work. R. Chugh reports grants and personal fees from Epizyme, Inc.; personal fees from Ipsen and Deciphera; grants from Qilu Puget Sound, AADi, Novartis, Medivation, Advenchen, Plexxikonn, Springworks, Ayala, and Mundipharma; grants and non-financial support from GlaxoSmithKline; and non-financial support from Janssen and AstraZeneca outside the submitted work. A.P. O'Dea reports personal fees from Puma Biotechnology, Pfizer, Daiichi Sankyo, AstraZeneca, and Novartis outside the submitted work. H.R. Mirshahidi reports other support from Merck and Takeda outside the submitted work. J. Rodon Ahnert reports grants and personal fees from Novartis, Spectrum Pharmaceuticals Inc., and Pfizer; personal fees from Eli Lilly, Orion Pharmaceuticals, Servier Pharmaceuticals, Peptomyc, Roche Pharmaceuticals, Ellipses Pharma, Certera, NovellusDX, and IONCTURA SA; personal fees and other support from Molecular Partners, Merck Sharp & Dohme and Kelun Pharmaceuticals/Klus Pharma; grants, personal fees, and other support from Bayer; grants from Ipsen, Blueprint Medicines, Tocagen, Symphogen, BioAlta, GenMab, CytomX, Kelun-Biotech, and Takeda-Miillenium; grants and other support from GlaxoSmithKline; and other support from ESMO, Department of Defense, Louisiana State University, Huntsman Cancer Institute, Cancer Core Europe, Karolinska Cancer Institute, King Abdullah International Medical Research Center, WIN Consortium, Janssen, European Journal of Cancer, VHIO/Ministero De Empleo Y Seguridad Social, Chinese University of Hong Kong, SOLTI, and Elsevier outside the submitted work. R.M. Connolly reports grants from Pfizer, Genentech, Novartis, Puma Biotechnology, Merck, and Macrogenics outside the submitted work. R.J. Gray reports grants from National Cancer Institute during the conduct of the study. C.D. Blanke reports grants from NCI during the conduct of the study. Y.K. Chae reports grants and personal fees from BMS during the conduct of the study. R. Kurzrock reports grants, personal fees, non-financial support, and other support from Biological Dynamics, Boehringer Ingelheim, Debiopharm, Foundation Medicine, Genentech, Grifols, Guardant, Incyte, Konica Minolta, Medimmune, Merck Serono, Omniseq, Pfizer, Sequenom, Takeda, and TopAlliance; is a consultant or advisory board member and/or has received speaker fees from Actuate Therapeutics, AstraZeneca, Bicara Therapeutics, Biological Dynamics, Eisai, EOM Pharmaceuticals, Iylon, Merck, NeoGenomics, Neomed, Pfizer, Prosperdtx, Roche, TD2/Volastra, Turning Point Therapeutics, and X-Biotech; has an equity interest in CureMatch Inc., CureMetrix, and IDbyDNA; serves on the board of CureMatch and CureMetrix; and is a co-founder of CureMatch outside the submitted work. No disclosures were reported by the other authors.

S. Adams: Conceptualization, resources, data curation, formal analysis, supervision, investigation, writing–original draft, writing–review and editing. M. Othus: Conceptualization, resources, data curation, software, formal analysis, supervision, funding acquisition, investigation, visualization, methodology, writing–original draft, project administration, writing–review and editing. S.P. Patel: Conceptualization, resources, supervision, funding acquisition, investigation, methodology, project administration, writing–review and editing. K.D. Miller: Conceptualization, writing–review and editing. R. Chugh: Investigation. S.M. Schuetze: Investigation. M.D. Chamberlin: Data curation, investigation. B.J. Haley: Data curation, investigation. A.V. Storniolo: Data curation, investigation. M.P. Reddy: Investigation. S.A. Anderson: Investigation. C.T. Zimmerman: Investigation. A.P. O'Dea: Investigation. H.R. Mirshahidi: Investigation, writing–review and editing. J. Rodon Ahnert: Conceptualization, resources, investigation, writing–review and editing. F.J. Brescia: Conceptualization, resources, investigation, writing–review and editing. O. Hahn: Conceptualization, resources, investigation, writing–review and editing. J.M. Raymond: Investigation. D.D. Biggs: Conceptualization, resources, formal analysis, supervision, investigation, methodology, writing–review and editing. R.M. Connolly: Conceptualization, resources, formal analysis, supervision, investigation, methodology, writing–review and editing. E. Sharon: Conceptualization, resources, formal analysis, funding acquisition, methodology, project administration, writing–review and editing. L.A. Korde: Conceptualization, resources, supervision, funding acquisition, investigation, methodology, project administration, writing–review and editing. R.J. Gray: Conceptualization, resources, formal analysis, supervision, funding acquisition, investigation, methodology, project administration, writing–review and editing. E. Mayerson: Conceptualization, resources, formal analysis, supervision, funding acquisition, investigation, methodology, project administration, writing–review and editing. M. Plets: Conceptualization, resources, formal analysis, funding acquisition, investigation, methodology, project administration. C.D. Blanke: Conceptualization, resources, supervision, funding acquisition, investigation, methodology, project administration, writing–review and editing. Y.K. Chae: Conceptualization, resources, supervision, funding acquisition, investigation, methodology, project administration, writing–review and editing. R. Kurzrock: Conceptualization, resources, supervision, funding acquisition, investigation, methodology, project administration, writing–review and editing.

Research reported in this publication was supported by the National Cancer Institute of the National Institutes of Health under grant award numbers U10CA180888, U010CA180819, U10CA180820, U10CA180821, U10180868, U10CA180794, and UG1CA233196, and in part by Bristol-Myers Squibb Company. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health or the Bristol-Meyers Squibb Company.

We would like to thank the patients who participated in this trial and their families, as well as the study teams of participating institutions. We also thank Dr. Roberto Salgado for scoring the TILs for the responders.

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.