A Phase II Basket Trial of Dual Anti–CTLA-4 and Anti–PD-1 Blockade in Rare Tumors (DART SWOG 1609) in Patients with Nonpancreatic Neuroendocrine Tumors

Immune checkpoint blockade has transformed oncology with the potential for durable responses even in patients with metastatic disease. Approved indications for immune checkpoint blockade in rare tumors are limited to Merkel cell carcinoma, cutaneous squamous cancers, and microsatellite-unstable malignancies (1). Rare cancer histologies, collectively representing approximately a quarter of all cancers diagnosed, remain understudied and the efficacy of immune checkpoint blockade in these patient populations is unknown.

Neuroendocrine neoplasms represent a rare histologic subset of tumors with complex classification criteria dependent on the putative organ of origin, precluding a single taxonomy across anatomic sites (2). The World Health Organization (WHO) and European Neuroendocrine Tumour Society (ENETS) have developed a classification scheme reflected in the most recent staging guidelines (Supplementary Table 1) and was utilized in our study. Neuroendocrine neoplasms can develop throughout the body, with pancreatic neuroendocrine tumors (PNET) in particular having unique biological and clinical characteristics resulting in additional therapies being utilized for PNETs. Thus, PNETs are assessed in a separate cohort within S1609, currently accruing. Clinical trials for neuroendocrine neoplasms, particularly in high-grade neuroendocrine carcinomas, have been difficult to conduct due to the rarity of the disease, difficulties in precise classification, and a lack of robust predictive biomarkers for therapeutic efficacy. In this study, we evaluated the combination of ipilimumab and nivolumab in patients with diverse histologic sites and across tumor grades.

SWOG 1609 dual anti–CTLA-4 & anti–PD-1 blockade in rare tumors (DART), a basket immunotherapy trial studying ipilimumab plus nivolumab across multiple cohorts of rare tumor histologic subtypes, was designed to address the question of the efficacy of these agents in these understudied populations. The combination of ipilimumab and nivolumab was selected over nivolumab alone due to the signal-finding nature of this study with small cohorts of rare tumors, with lower-dose ipilimumab chosen to balance tolerability with potential efficacy. The trial is currently open across the United States at 861 sites. We present here the clinical data of the (nonpancreatic) neuroendocrine cohort of SWOG 1609 (S1609) DART.

The trial was conducted by SWOG, and the investigational agents were provided by the Cancer Therapy Evaluation Program (CTEP) of the NCI under an NCI Collaborative Research and Development Agreement (CRADA) with Bristol-Myers Squibb (BMS). All study subjects provided their voluntary, written informed consent using a document approved by the institution's human subject protection committee. The study was conducted in accordance with the Declaration of Helsinki. The protocol and all amendments were approved by SWOG, the NCI, the NCI central institutional review board (CIRB), and by the regulatory committees at the participating institutions. The Caris analysis of neuroendocrine specimens is IRB exempt as all data were analyzed utilizing deidentified aggregate data.

Rare cancers, for the purposes of this study, were identified typically with an incidence of less than 6 in 100,000 per year (3). Tumor grading was based on 2010 WHO criteria; pathology and grade were also determined by review of local pathology reports by the study principal investigators. No central pathology review was performed. This cohort (Cohort 23) of S1609 is comprised of refractory neuroendocrine neoplasms, independent of histologic grade and organ of origin, with the exception of pancreatic neuroendocrine neoplasms, which were stratified to a different cohort due to unique biology and alternate standard-of-care therapies. Well-differentiated, grade 3 neuroendocrine neoplasms were eligible for this cohort, and microsatellite instability (MSI) status was not available.

Eligible patients had (nonpancreatic) neuroendocrine neoplasms, had progressed following at least one line of standard systemic therapy, and did not have an approved or standard therapy available that had been shown to prolong overall survival. At enrollment, patients were required to be 18 years of age or older; have a Zubrod performance status of 0–2; and adequate hematologic, hepatic, thyroid, adrenal axis, and renal function, with absolute neutrophil count ≥ 1,000/mcL, platelets ≥75,000/mcL, hemoglobin ≥8 g/dL, creatinine clearance ≥50 mL/minute, total bilirubin ≤2.0 × institutional upper limit of normal (IULN), aspartate aminotransferase (AST) and alanine aminotransferase (ALT) ≤ 3.0 × IULN, TSH or free T4 serum ≤ IULN, and normal adrenocorticotropic hormone (ACTH) ≤ IULN. Women of childbearing potential were required to have a negative serum pregnancy test and subjects were required to practice adequate birth control during protocol participation.

Treatment consisted of nivolumab 240 mg i.v. every 2 weeks and ipilimumab 1 mg/kg i.v. every 6 weeks on a continuous schedule, with dose adjustments and brief breaks from therapy specified in the protocol for treatment-related toxicities. Patients were removed from study therapy for disease progression, symptomatic deterioration, treatment delay for any reason >56 days, unacceptable or immune-related toxicity with inability to decrease prednisone to < 10 mg daily, or per patient request.

Patients were evaluated with a history and physical, laboratory analyses (complete blood count, comprehensive metabolic panel, thyroid-stimulating hormone, free thyroxine, ACTH, cortisol, lipase), and toxicity assessment at least every 6 weeks at the beginning of each cycle. Imaging studies for disease assessment were performed prestudy, week 8, week 16, week 24, and then every 12 weeks until progression.

The primary objective of this phase II trial was to evaluate the overall response rate [ORR, confirmed complete and partial responses (CR and PR)] by RECIST v1.1 based on local site review. Our objective was to distinguish between a true ORR ≤5% (null hypothesis, as patients had failed all known active therapies) versus ≥ 30% (alternative hypothesis, a potentially clinically meaningful difference in tumor response in refractory solid tumors). A Simon two-stage design was used, which required an analysis on the first 6 eligible patients who received protocol therapy. If 1 or more of the 6 patients had a response (confirmed CR or PR), an additional 10 patients were to be accrued. The design specified 2 or more responses out of 16 patients would reject the null hypothesis (one-sided alpha = 13%, power = 87%). This cohort accrued more than 16 patients because, unexpectedly, accrual was faster than expected following the 2-week closure notification and several additional patients were enrolled onto incorrect cohorts in S1609 and restratified into this cohort after SWOG review of local pathology reports prior to knowledge of clinical benefit or toxicity. Analysis of ORR by tumor grade was not prespecified. The secondary objectives were to estimate progression-free survival (PFS), overall survival (OS), ORR by immune-related RECIST (iRECIST), PFS by iRECIST, and to assess toxicity.

PFS was measured from the start of protocol therapy to the first date of progression by RECIST v1.1 or death by any cause, with patients last known to be alive without progression censored at the date of last contact. OS was measured from the date of study registration to the date of death by any cause, with patients last known to be alive censored at the date of last contact. PFS and OS estimates were calculated using the Kaplan–Meier method and compared using log-rank tests. Confidence intervals for medians were constructed using the method of Brookmeyer and Crowley (4), and confidence intervals (CI) for point estimates (e.g., 6-month PFS) were calculated using the log–log transformation. CIs for the primary ORR analysis accounted for the two-stage design (5); exact binomial CIs were calculated for subgroups utilizing the R function “get_CI” from the package OneArmPhaseTwoStudy. Fisher exact test was used to compare subgroups. All analyses were performed using R version 3.4.3.

Thirty-five patients from 22 National Clinical Trial Network (NCTN) institutions were registered between March 13, 2017 and May 25, 2018, with 32 patients meeting eligibility criteria and receiving protocol therapy who are summarized in (Table 1). Three patients were excluded from analyses: one ineligible due to histology, one ineligible due to inadequate washout period prior to treatment initiation, and one eligible patient refused protocol treatment after giving initial consent. Of the 32 eligible patients who received protocol therapy, the median age was 60 years (range 36–81). The most common sites of primary tumor were lung and small intestine (both n = 6). Notably, 18 of the 32 patients (56%) had high-grade carcinoma. The median number of prior lines of therapy was 2 for both the entire cohort as well as for patients with high-grade disease.

Treatment-related adverse events are summarized in Table 2, with 84.4% of patients experiencing an adverse event (AE), and 50% developing a grade 3–4 AE. The most common AEs (across all grades and at least possibly related to treatment) were hypothyroidism (31%), fatigue (28%), nausea (28%), vomiting (25%), AST increase (25%), alkaline phosphatase increase (22%), and anorexia (22%). Six patients experienced grade 4 events, 2 with sepsis (6%), 2 with increased lipase (6%), 1 with retinopathy (3%), and 1 with hyperglycemia (3%). Overall, 72% of patients developed an immune-related AE (irAE) of any grade on treatment, with 38% (n = 10) developing grade 3–4 irAEs. The most common irAEs of any grade were hypothyroidism (31%) and AST increase (25%). The most common grade 3–4 irAEs were ALT increase (9%) and AST increase, lipase increase, and encephalopathy (all 6%). There were no treatment-related deaths.

Among 32 patients, the ORR was 25% (95% CI, 13%–42%), with 3% (n = 1) of patients achieving CR and 22% (n = 7) attaining a PR (Table 3; Fig. 1A). Altogether, 41% of patients had stable disease with 6% having stable disease >6 months and responses ongoing (Fig. 1D). Response rates were similar regardless of organ of origin (Fig. 1B). High-grade neuroendocrine carcinoma was present in 56% of patients (n = 18); 31% of patients had intermediate-grade and 12% had low-grade biology. Within the high-grade neuroendocrine cohort, 44% (95% CI, 22%–69% of patients; n = 8) had an objective response, with no responses (95% CI, 0%–23%) in the intermediate or low-grade tumors (Fig. 1C; P = 0.004 for ORR in high- vs. low/intermediate grade). The overall 6-month PFS rate was 31% (19%–52%), with a 6-month PFS rate of 44% (27%–75%) in high-grade disease versus 14% (4%–52%) in low-grade disease. The median PFS is 4 months 95% CI (3–6) with ongoing responses (Supplementary Fig. S1) and the median OS is 11 months 95% CI (6–∞; Supplementary Fig. S2).

We also assessed patients with iRECIST. There was only one patient that differed significantly. This patient had intermediate-grade disease of small intestine origin and achieved a confirmed iPR (instead of progressive disease per RECIST v1.1). By RECIST v1.1, this patient had progressed 59 days after treatment initiation, but currently remains on study with clinical benefit 326+ days after treatment initiation with confirmed iPR.

Neuroendocrine neoplasms represent a histologically and molecularly heterogeneous constellation of rare cancers that can arise across various organ types. Low- and intermediate-grade well-differentiated neuroendocrine neoplasms overexpress somatostatin receptors, which can be utilized both for functional imaging as well as therapeutic targeting with long-acting somatostatin analogues (6). In contrast, high-grade neuroendocrine carcinomas typically have more aggressive biology and minimal expression of somatostatin receptors and are typically treated with chemotherapy (7). ¹⁷⁷Lu-Dotatate has recently shown activity with an 18% response rate for somatostatin-positive (low/intermediate grade) midgut neuroendocrine tumors and has attained FDA approval (8).

To date, immune checkpoint blockade with anti–CTLA-4 and anti–PD-1 has not been prospectively studied broadly across rare tumors, or in combination for neuroendocrine neoplasms. Prior studies of anti–PD-1–directed monotherapy have had limited efficacy across the spectrum of neuroendocrine neoplasms (9–13). For this trial, the combination of ipilimumab with nivolumab was chosen to maximize response rates in signal finding cohorts relative to monotherapy, and the dose of ipilimumab of 1 mg/kg i.v. every 6 weeks was chosen to minimize toxicity while retaining combinatorial efficacy based on published comparative data in other tumor types (14). In the nonpancreatic neuroendocrine cohort of S1609 reported here, 25% of patients had a response to ipilimumab plus nivolumab, with no difference in ORR relative to organ of origin in our small cohort with a myriad of primary sites of origin (Fig. 1B). None of the lung tumors in our cohort were small cell lung cancer, for which anti–PD-1 and anti–CTLA-4 have previously been investigated (15–17). Of note, 44% of patients with high-grade neuroendocrine carcinomas had an objective response to therapy (Fig. 2). Overall, this regimen resulted in no grade 5 toxicities, a <10% rate of grade 3–4 immune-related colitis and hepatitis, and no reported pneumonitis in this cohort. However, serious treatment-related toxicity occurred in 37.5% of patients and treatment discontinuation due to grade 3–4 toxicities occurred in 31.5% of patients.

With responses across different primary tumor sites and a signal toward improved response in high-grade neuroendocrine carcinomas, a potential predictive biomarker to help select for patients who may derive preferential benefit from immune checkpoint blockade is crucial. Biomarker analyses are underway for patients in the neuroendocrine cohort with a focus on PD-L1 expression by IHC, tumor mutational burden, and comprehensive transcriptomic profiling given relevance in other tumor types (18, 19). One prior study found that high-grade gastroenteropancreatic neuroendocrine carcinomas had a higher rate of PD-L1 expression relative to lower-grade tumors, and was associated with poorer survival (20). Poorly differentiated neuroendocrine carcinoma may also have a higher mutational burden than lower grade tumors (21, 22). Clinically, anecdotal response to anti–PD-1 has been reported in high mutational burden neuroendocrine carcinoma previously (23).

In addition, PD-L1 by IHC and tumor mutational burden in both tissue and blood have been associated with improved response to anti–PD-1 therapy across tumor types (18, 24–26). Host factors related to HLA-type, immune status, microbiome, and underlying etiology likely play a key role in influencing response to immune checkpoint blockade. To better understand the potential molecular basis for immunotherapeutic response, we assessed TMB and PD-L1 IHC in an independent cohort of neuroendocrine neoplasms not from S1609. While PD-L1 IHC was not different in high-grade versus intermediate/low-grade tumors, high-grade neuroendocrine tumors had a significantly greater rate of high TMB relative to intermediate/low-grade tumors, independent of site of origin (Supplementary Table S2). Thus, TMB merits additional evaluation as a biomarker to potentially discriminate response to combinatorial immune checkpoint blockade, which will be specifically assayed in this cohort through whole-exome sequencing. Overexpression of PD-L1 in high-grade neuroendocrine carcinomas of the lung has also been reported (27). The TMB landscape in large-cell neuroendocrine tumor of lung (n = 353) had been previously investigated and the median TMB found to be higher than in small cell lung cancer and non–small cell lung cancer (28).

Strengths of this study include a broad population of patients across various tumor types representing both academic and community site accrual across the United States, and support from the NCI, SWOG, and patient advocacy groups. Weaknesses of this study include its nonrandomized nature, small sample size, and heterogenous patient population, which limit outcome comparisons between subgroups. In addition, central pathology review was not mandated, and grading of tumor was done locally and most often with Ki67 measurement. Local pathology and imaging response assessments were utilized.

As studied in SWOG 1609, a basket rare tumor immunotherapy trial, ipilimumab plus nivolumab has clinical activity in nonpancreatic neuroendocrine neoplasms, in particular high-grade neuroendocrine carcinomas, across numerous primary originating organ sites. Ongoing studies focused on high-grade neuroendocrine carcinoma with rigorous correlative science to better understand host and tumor characteristics of immunotherapeutic response are underway.

S.S. Patel is an employee/paid consultant for Bristol-Myers Squibb, AstraZeneca, Eli Lilly, Illumina, Tempus, Nektar, and Novartis, and reports receiving commercial research grants from Bristol-Myers Squibb, Eli Lilly, Fate, AstraZeneca, Merck, Pfizer, Roche, Xcovery and Genocea. Y.K. Chae is an employee/paid consultant for BMS, AstraZeneca, Genentech, Foundation Medicine, Guardant Health, Biodesix, and Lilly Oncology, reports receiving commercial research grants from BMS and Abbvie, and reports receiving speakers bureau honoraria from BMS, AstraZeneca, Genentech and Lilly Oncology. P.P. Singh is an advisory board member/unpaid consultant for Eisai Inc. and Novartis. M.H. Shah reports receiving commercial research grants from Merck. T. Al Baghdadi is an employee/paid consultant for and holds ownership interest (including patents) in Bristol-Myers Squibb. M. Matrana reports receiving speakers bureau honoraria from Bristol-Myers Squibb. W.M. Korn is an employee/paid consultant for Chris Life Sciences and Merck, Sharp & Dohme. C.W. Ryan is an employee/paid consultant for Exelexis, Pfizer, Eisai, Deciphera, and Genentech, reports receiving commercial research grants from Argos Therapeutics, Bristol-Myers Squibb, CytRx Corporation, Daiichi-Sankyo, Exelexis, Genentech, Novartis, Karyopharm Therapeutics, Merck, Pfizer, TRACON Pharma, Xynomic, and Eisai. R. Kurzrock is an employee/paid consultant for Gaido, LOXO, X-Biotech, Actuate Therapeutics, Roche, NeoMed, Soluventis, Pfizer, Merck, reports receiving commercial research grants from Incyte, Genentech, Merck Serono, Pfizer, Sequenom, Foundation Medicine, Guardant Health, Grifols, Konica Minolta, DeBiopharm, Boerhringer Ingelheim, and OmniSeq [All institutional]), reports receiving speakers bureau honoraria from Roche, has Stock and Other Equity Interests (IDbyDNA, CureMatch, Inc., and Soluventis); Consulting or Advisory Role (Gaido, LOXO, X-Biotech, Actuate Therapeutics, Roche, NeoMed, and Soluventis, Pfizer and Merck); Speaker&apos;s fee (Roche); Research Funding (Incyte, Genentech, Merck Serono, Pfizer, Sequenom, Foundation Medicine, Guardant Health, Grifols, Konica Minolta, DeBiopharm, Boehringer Inegelheim, and OmniSeq [All institutional]); Board Member (CureMatch, Inc., and CureMetrix, Inc.). No potential conflicts of interest were disclosed by the other authors.

The content is solely the responsibility of the authors and does not necessarily represent the official views of the NIH or Bristol-Myers Squibb Company.

Conception and design: S.P. Patel, M. Othus, Y.K. Chae, F.J. Giles, H.X. Chen, E. Sharon, C.D. Blanke

Development of methodology: S.P. Patel, Y.K. Chae, F.J. Giles, D.E. Hansel, E. Sharon, M. Plets

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): S.P. Patel, Y.K. Chae, D.E. Hansel, P.P. Singh, A. Fontaine, M.H. Shah, A. Kasi, T.A. Baghdadi, M. Matrana, Z. Gatalica, J. Hayward, E. Sharon, C.W. Ryan, C.D. Blanke

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): S.P. Patel, M. Othus, Y.K. Chae, F.J. Giles, D.E. Hansel, W.M. Korn, C. McLeod, E. Sharon, E. Mayerson, M. Plets, C.D. Blanke

Writing, review, and/or revision of the manuscript: S.P. Patel, M. Othus, Y.K. Chae, F.J. Giles, D.E. Hansel, P.P. Singh, M.H. Shah, A. Kasi, T.A. Baghdadi, M. Matrana, W.M. Korn, H.X. Chen, E. Sharon, C.W. Ryan, M. Plets, C.D. Blanke

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): S.P. Patel, M.H. Shah, J. Hayward, C. McLeod, H.X. Chen, E. Sharon, E. Mayerson, C.W. Ryan, M. Plets, C.D. Blanke

Study supervision: S.P. Patel, Y.K. Chae, F.J. Giles, J. Hayward, H.X. Chen, E. Sharon, C.W. Ryan, M. Plets, C.D. Blanke

This work was supported by NIH/NCI grant awards CA180888, CA180819, CA180821, CA180820, CA189870, CA180850, CA189873, CA189809, CA180858, CA189830, CA189821, CA180834, CA189971, CA180801, CA189953, CA189856; legacy grant awards CA73590, CA12644; and in part by Bristol-Myers Squibb Company. The authors wish to thank Ms. Marcia Horn, JD, SWOG Patient Advocate and President/CEO, International Cancer Advocacy Network; Mr. Dion Holmes, Protocol Coordinator, SWOG Operations Office; Dr. Heloisa P. Soares, MD, PhD, University of New Mexico; Dr. Howard Streicher, MD, National Cancer Institute, Investigational Drug Branch, Cancer Therapy Evaluation Program, and Dr. David Arguello, M.D., and Ms. Michelle Winerip, Caris Life Sciences, for their invaluable assistance with this trial.

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.