Ipilimumab and Gemcitabine for Advanced Pancreatic Cancer: A Phase Ib Study

Abstract

Background

Pancreatic ductal adenocarcinoma (PDAC) remains resistant to chemotherapy and immunotherapy individually because of its desmoplastic stroma and immunosuppressive tumor microenvironment. Synergizing cytotoxic T‐lymphocyte–associated antigen 4 (CTLA‐4) immune checkpoint blockade with chemotherapy could overcome these barriers. Here we present results of a phase Ib trial combining ipilimumab and gemcitabine in advanced PDAC.

Materials and Methods

This was a single‐institution study with a 3 + 3 dose‐escalation design. The primary objective was to determine the maximum tolerated dose (MTD). Secondary objectives included determining the toxicity profile, objective response rate (ORR), median progression‐free survival (PFS), and overall survival (OS).

Results

Twenty‐one patients were enrolled, 13 during dose escalation and 8 at the MTD. The median age was 66 years, 62% were female, 95% had stage IV disease, and 67% had received at least one prior line of therapy. The primary objective to establish the MTD was achieved at doses of ipilimumab 3 mg/kg and gemcitabine 1,000 mg/m². The most common grade 3 or 4 adverse events were anemia (48%), leukopenia (48%), and neutropenia (43%). The ORR was 14% (3/21), and seven patients had stable disease. Median response duration for the three responders was 11 months, with one response duration of 19.8 months. Median PFS was 2.78 months (95% confidence interval [CI], 1.61–4.83 months), and median OS was 6.90 months (95% CI, 2.63–9.57 months).

Conclusion

Gemcitabine and ipilimumab is a safe and tolerable regimen for PDAC with a similar response rate to gemcitabine alone. As in other immunotherapy trials, responses were relatively durable in this study.

Implications for Practice

Gemcitabine and ipilimumab is a safe and feasible regimen for treating advanced pancreatic cancer. Although one patient in this study had a relatively durable response of nearly 20 months, adding ipilimumab to gemcitabine does not appear to be more effective than gemcitabine alone in advanced pancreatic cancer.

Introduction

Pancreatic ductal adenocarcinoma (PDAC) is currently the fourth leading cause of cancer death in the U.S. and is expected to become the second leading cause of cancer‐related mortality by 2030 [1]. Treatment of PDAC remains challenging because of its advanced presentation and resistance to chemotherapy and immunotherapy [2].

The microenvironment of pancreas tumors is a major barrier to delivery of effective therapy. The dense peritumoral stroma impedes blood flow and prevents cytotoxic tumor‐infiltrating lymphocytes (TILs) from reaching tumor cells [3, 4]. Regulatory T cells make up a major component of the T‐cell infiltrate, which promotes fibrosis and downregulates cytotoxic TILs [5]. Patients with PDAC also have elevated serum levels of interleukin‐10 and transforming growth factor beta (TGF‐β) compared with healthy controls, creating an immunosuppressive state [4]. Thus, shifting the balance of the T‐cell infiltrate toward immune activation against tumor cells may be an important strategy in pancreatic cancer.

Inhibition of the cytotoxic T‐lymphocyte–associated antigen 4 (CTLA‐4) immune checkpoint has been proven to induce tumor regression and prolong survival [6, 7]. Ipilimumab is a CTLA‐4 inhibitor that has shown activity and survival benefit in melanoma [7–9]. Ipilimumab combined with the anti–programmed cell death‐1 (PD‐1) monoclonal antibody nivolumab has shown impressive response rates and improved survival in many malignancies, including melanoma and renal cell carcinoma [9–11]. In less immunotherapy‐sensitive histologies such as pancreatic adenocarcinoma [12, 13], ipilimumab could have a synergistic effect with chemotherapy by augmenting the antitumor T‐cell response.

Based on this hypothesis, we conducted a phase Ib clinical trial of CTLA‐4 blockade with ipilimumab in combination with gemcitabine in patients with advanced pancreatic cancer. We report here the maximum tolerated doses for the combination, safety data, and the antitumor activity observed in this trial.

Materials and Methods

Patient Inclusion Criteria

Patients aged 18 years or older with confirmed advanced or metastatic pancreatic adenocarcinoma were included. Patients were required to have an Eastern Cooperative Oncology Group performance status of ≤1 and adequate liver, renal, and bone marrow function. Prior gemcitabine given concurrently with radiotherapy for localized cancer was permitted. Prior gemcitabine as adjuvant therapy was allowed if ≥3 months had elapsed between the last dose of gemcitabine and detection of recurrent disease.

Patient Exclusion Criteria

Patients previously treated with ipilimumab, CD137 agonists, CTLA‐4 inhibitors or agonists, or anti–PD‐1 therapies were excluded. Prior systemic therapy for advanced pancreatic cancer with gemcitabine was prohibited. Other major exclusion criteria included central nervous system metastatic disease, active autoimmune disease, active infections including human immunodeficiency virus, hepatitis B or C infection, recent vaccine use, or significant comorbid conditions.

The trial was conducted at the Robert H. Lurie Comprehensive Cancer Center of Northwestern University and was approved by the institutional review board of Northwestern University Feinberg School of Medicine. It was registered with ClinicalTrials.gov as NCT01473940 on October 19, 2011. All patients provided written informed consent before enrollment. The study was conducted in accordance with the International Conference on Harmonization guidelines for Good Clinical Practice and the Declaration of Helsinki.

Study Design and Treatment

This was an open‐label study using a traditional 3 + 3 dose‐escalation design. Enrolled patients received gemcitabine weekly at the assigned dose intravenously over 30 minutes for 7 weeks followed by 1 week off and then subsequent cycles for 3 weeks on and 1 week off. Prior to gemcitabine infusion on weeks 1, 4, 7, and 10, patients received ipilimumab at the assigned dose intravenously over 90 minutes. Thereafter, ipilimumab was given every 12 weeks. Patients were continued on study treatment until unacceptable toxicity, progressive disease, or withdrawal from the study. Each patient was evaluated for dose‐limiting toxicity (DLT) at the end of week 12 of study treatment. Escalation to the next dose level occurred only if there were no DLTs at the previous dose level and thus did not establish the maximum tolerated dose (MTD). An additional expansion cohort at the MTD was included in the study design to generate additional safety and efficacy data. Dose levels included in the study are shown in Table 1.

Dose‐Limiting Toxicity and Maximum Tolerated Dose

Patients were assessed for dose‐limiting toxicity at the end of week 12. All toxicities were assessed according to the Common Terminology Criteria for Adverse Events version 4.0. For this study, an immune‐related adverse event (irAE) was defined as an adverse event (AE) of unknown etiology associated with drug exposure and consistent with an immune phenomenon. The MTD was defined as the highest dose at which zero of three patients or one of six patients experienced DLTs.

Major criteria for dose‐limiting toxicity included both hematologic and nonhematologic criteria. The hematologic criteria included grade 4 neutropenia for >7 days, febrile neutropenia with ≥ grade 3 neutropenia, grade 3 thrombocytopenia with ≥ grade 3 hemorrhage, or grade 4 thrombocytopenia. All grade ≥3 nonhematologic toxicities were defined as dose‐limiting toxicities except for hepatic dysfunction related to biliary tract obstruction, grade 3 toxicities lasting <24 hours (e.g., hypomagnesemia corrected with IV magnesium), grade ≥ 3 vomiting without appropriate supportive care, or toxicities clearly related to progressive disease. Any toxicities that resulted in inability to resume treatment for >14 days were considered DLTs.

Safety Assessments

Baseline and screening assessments included history and complete physical examination. Laboratory testing included complete blood counts, comprehensive metabolic profile, amylase, lipase, thyroid‐stimulating hormone (TSH), and carbohydrate antigen 19‐9 (CA19‐9), and baseline imaging and pathology reviews were conducted. Safety assessments were performed before each dose of gemcitabine and ipilimumab and included monitoring and recording all AEs, documentation of concomitant medications, routine laboratory tests, focused history, and physical examinations. Amylase, lipase, and TSH levels were also checked prior to each ipilimumab infusion.

Response Assessment

Imaging was obtained at baseline and response evaluation was done at week 12 by follow‐up imaging using immune‐related response criteria (irRC). All subjects who received at least one dose of study treatment were evaluated for response. Serum CA19‐9 was also measured at baseline and at week 12. For those who achieved a response or stable disease, study treatment was continued, and patients were evaluated for response using imaging and CA19‐9 measurement every 12 weeks until progression, unacceptable toxicity, or study withdrawal.

Correlative Studies

Prior to December 2013, the study included an optional peripheral blood collection. Studies to quantify and monitor regulatory T‐cell number and function, frequency of T‐cell subsets, and pro‐ versus anti‐inflammatory T‐cell function before, during, and after study treatment were planned. However, given that an insufficient number of samples were collected and analyzed, these studies were not completed.

Statistical Analysis

The primary endpoint of the study was to establish the MTD of the combination of ipilimumab and gemcitabine in advanced or metastatic pancreatic adenocarcinoma. Secondary endpoints included the toxicity profile of the combination, objective response rate (ORR) by irRC version 1.0 criteria, median progression‐free survival (PFS), and median overall survival (OS). Accrual began at dose level 1 using the standard 3 + 3 phase I dose‐escalation design [14].

The dose‐escalation schema resulted in a probability of dose escalation that was 91% (49%, 17%, 3%) when the toxicity rate was 10% (30%, 50%, 70%), respectively. For the expansion cohort, dose‐limiting toxicity was monitored by calculating the Bayesian predictive probability of a DLT given the data to date. This was done after each DLT was observed. Once a DLT was observed, patients were removed from the study. If there had been at least a 90% probability that the DLT rate exceeded 33%, then accrual to the expansion cohort would have been terminated.

Adverse events and safety data were analyzed using descriptive statistics. Response assessments were displayed using a waterfall plot, and PFS and OS were calculated using the Kaplan‐Meier method.

Descriptive statistics were calculated for patient characteristics (age, sex, stage of disease, and information about prior treatments). Progression‐free and overall survival were estimated using Kaplan‐Meier calculations. Analyses were performed using SAS version 9.4 (SAS Institute, Cary, NC).

Results

Baseline Patient Characteristics

From June 11, 2012, to March 23, 2016, 25 patients were screened for study enrollment. Two were ineligible based on lab values, one patient declined enrollment, and one patient never received study treatment for unknown reasons. Twenty‐one patients were enrolled in the study, of whom 13 were in the dose‐escalation phase and 8 in the expansion cohort treated at the established MTD. The median age was 66 years, 62% of patients were female, and 95% had stage IV disease. Fifty‐two percent of patients had one prior line of therapy, and 14% had two prior lines of therapy. Baseline characteristics for the study population are shown in Table 2.

Dose Escalation

In cohort 1, all three patients who were enrolled tolerated the dose well with no DLTs observed. In cohort 2, four patients were enrolled with one patient experiencing multiple serious adverse events (SAEs) but no DLT. The SAEs observed included grade 4 catheter‐related infection, grade 3 hemolytic uremic syndrome, grade 3 thrombocytopenia, and grade 3 retroperitoneal hemorrhage. Because of the number, severity, and rapid onset of SAEs shortly after study enrollment that could not be clearly attributed to study treatment, the patient became nonevaluable for DLTs, requiring an additional patient in this cohort. However, this patient was still included for general adverse event statistics and response assessment.

In cohort 3, three patients were initially enrolled with one experiencing a DLT of grade 3 elevation in alanine transferase (ALT). Per protocol, an additional three patients were enrolled (resulting in six patients in cohort 3). A second DLT of grade 3 diarrhea was recorded in this cohort. Multiple other SAEs were recorded in cohort 3, which are summarized in supplemental online Table 1. This resulted in the MTD being established at the cohort 2 doses, gemcitabine 1,000 mg/m² and ipilimumab 3 mg/kg. An additional eight patients were enrolled at the MTD as an expansion cohort. No DLT was reported in the expansion cohort.

Safety

Treatment‐related AEs are summarized in Table 3. The most common AEs that occurred in ≥10% of patients were anemia, thrombocytopenia, aspartate aminotransferase (AST)/ALT elevation, leukopenia, fatigue, neutropenia, nausea, rash, alkaline phosphatase elevation, fever, diarrhea, dry mouth, creatinine increase, hyperglycemia, bilirubin elevation, peripheral sensory neuropathy, flu‐like symptoms, chills, hyponatremia, hypokalemia, and anorexia. Grade 3 or higher AEs were observed in 16 patients (76%) with one grade 3 elevation in ALT and one grade 3 diarrhea being the dose‐limiting toxicities at the gemcitabine 1,000 mg/m² and ipilimumab 6 mg/kg doses. 74% of all grade 3 or higher AEs were hematologic toxicities, including the only two grade 4 AEs observed (leukopenia and neutropenia). The most common nonhematologic grade 3 or higher AEs were diarrhea (10%), nausea (10%) and hypokalemia (10%).

Grade 3 or higher immune‐related AEs were observed in 19% of patients. Grade 3 diarrhea occurred in two patients (10%), and grade 3 AST/ALT elevation occurred in two patients (10%). Two of these AEs (one grade 3 ALT elevation and one grade 3 diarrhea) occurred in cohort 3 receiving ipilimumab 6 mg/kg, and two occurred in cohort 2 (one grade 3 AST elevation and one grade 3 diarrhea) receiving ipilimumab 3 mg/kg. There were no other grade 3 or 4 irAEs, including colitis, pneumonitis, rash, endocrine dysfunction, neuritis, and myocarditis. Serious adverse events are reported in supplemental online Table 1.

Response Assessment

Best response was evaluated by irRC criteria in all 21 patients enrolled. A total of three patients achieved a partial response, and seven patients had stable disease as the best response. There were no complete responses, and eight patients had a best response of progressive disease. The three responders had a median age of 66, 33% were female, all were white, and all had metastatic disease at study enrollment. Two had received two prior lines of therapy (both received gemcitabine followed by chemoradiation with 5‐fluorouracil [5‐FU]), and one had received one prior line of therapy (FOLFIRINOX).

The ORR was 14% (3/21). Two of the responses were in cohort 2 (MTD, gemcitabine 1,000 mg/m², ipilimumab 3 mg/kg), and one response occurred in cohort 3 (gemcitabine 1,000 mg/m², ipilimumab 6 mg/kg). Of the seven patients who achieved stable disease, two were in cohort 1 (gemcitabine 750 mg/m², ipilimumab 3 mg/kg), two were in cohort 2, and three were in cohort 3. Their median age was 68, 71% were female, and all seven had metastatic disease. Three were in the upfront setting, three had one prior line of therapy, and one had two prior lines of therapy. The median duration of stable disease was 2.37 months, and the median OS was 8.90 months. These data are displayed in Figure 1. Six patients are not represented in Figure 1 but were included for response assessment: three who died before the first response assessment and three who had progressive disease that was unmeasurable (e.g., peritoneal carcinomatosis). The ORR for patients treated at the MTD was 17% (2/12). All responding patients eventually discontinued study treatment because of progression of disease. A swimmer's plot illustrating these data is shown in Figure 2.

Responses were durable in the three patients who achieved partial responses with a median duration of 11 months. For the two responders in cohort 2, response durations were 8.5 months and 11 months. For the responder in cohort 3, response duration was 19.8 months. Median PFS was 2.78 months (95% confidence interval [CI], 1.61–4.83 months). Median OS was 6.90 months (95% CI, 2.63–9.57 months). Kaplan‐Meier curves for both PFS and OS are shown in Figure 3. For the 12 patients treated at the MTD, the median PFS and OS were 2.57 and 5.80 months, respectively.

Correlative Studies

Because of an insufficient number of peripheral blood samples collected for analysis, these studies were not performed, and sample collection was terminated in December 2013.

Discussion

This phase I study met its primary endpoint of establishing the MTD for the combination of gemcitabine and ipilimumab in advanced pancreatic adenocarcinoma. The irAE profile in this trial was comparable to prior trials of ipilimumab 3 mg/kg as monotherapy showing up to 15% grade 3 or higher irAEs [15, 16]. In this study, 19% of patients experienced grade 3 or higher irAEs, two with diarrhea and two with AST/ALT elevation. No grade 3 or 4 colitis, pneumonitis, endocrine dysfunction, or cutaneous toxicities were observed. The overall AE profile seen in this trial is also comparable to the toxicity profile of gemcitabine and nab‐paclitaxel [17, 18].

The absence of grade 3 or higher immune‐related colitis, pneumonitis, and endocrinopathies may be due to the addition of chemotherapy to ipilimumab. A similar decreased rate of immune‐related adverse events was previously demonstrated in clinical trials combining high‐dose ipilimumab with chemotherapy for metastatic melanoma [6, 19, 20]. The largest reductions in adverse events were in colitis and endocrine dysfunction [6, 19]. Whether chemotherapy truly tempers immune‐related adverse events from ipilimumab or by what mechanism this may occur remains largely unknown.

The objective response rate of 14% observed in this trial is similar to the response rate with gemcitabine alone and higher than the response rate with ipilimumab alone in advanced PDAC. Two phase II clinical trials of gemcitabine in advanced PDAC showed response rates of 6% and 11% [21, 22]. The single‐agent gemcitabine control arms in the MPACT (gemcitabine and nab‐paclitaxel) and PRODIGE ACCORD 11/0402 (FOLFIRINOX) trials also showed similar response rates of 7% and 9.4%, respectively [18, 23]. Ipilimumab monotherapy was also studied in a phase II trial of 27 patients with advanced PDAC. No objective responses were observed, although one patient achieved a delayed response after initial progression [12]. In this study, responses were relatively durable for the three responders with a median duration of 11 months and one response that lasted nearly 20 months.

The median OS observed here of 6.9 months was somewhat shorter than the median OS of 8.5 months seen in the MPACT trial using gemcitabine and nab‐paclitaxel or the median OS of 11.1 months observed with FOLFIRINOX [18, 23]. However, 66% of patients in this trial had received prior chemotherapy including 14% who had received at least two lines of treatment, whereas the MPACT and FOLFIRINOX trials included only untreated patients. Our results were very similar to the median OS of 6.1 months seen in the NAPOLI‐1 study of 5‐FU and nanoliposomal irinotecan in a similar advanced pancreatic cancer population who had received prior gemcitabine‐based chemotherapy [24].

As demonstrated in this study, the addition of immunotherapy to chemotherapy has not improved response rates compared with chemotherapy alone in advanced pancreatic cancer. A similar phase I trial that combined the CTLA‐4 antagonist tremelimumab with gemcitabine in untreated advanced pancreatic cancer patients showed a response rate of 10.5% with median OS of 7.4 months [25]. Another phase Ib trial that included 11 patients with advanced pancreatic cancer treated with the anti‐PD‐1 drug pembrolizumab combined with gemcitabine and nab‐paclitaxel showed a response rate of 18% [26], comparable to the response rate with gemcitabine and nab‐paclitaxel alone [18].

Combination immunotherapy and tumor vaccine strategies have also shown less than promising results to date. For example, the phase I/II ECHO‐203 study that evaluated the combination of the indoleamine 2,3‐dioxygenase 1 inhibitor epacadostat combined with durvalumab demonstrated no responses in the 15 patients with patients included [27]. Another phase I trial of the bispecific antibody M7824 targeting PD‐L1 and TGF‐β that included five patients with pancreatic cancer demonstrated only one partial response in a microsatellite instability–high patient [28]. Similarly, a number of vaccine approaches (GVAX, algenpantucel, CRS‐207) demonstrated no significant clinical benefit [29–31].

There is significant need to identify better predictive biomarkers of immune response in pancreatic cancer. Although PD‐L1, high microsatellite instability, and high tumor mutational burden have significant clinical utility in other cancers, their utility in pancreatic cancer may be low as they are rarely found in advanced PDAC and their predictive values are not established [13, 32–35]. Colony‐stimulating factor‐1 receptor (CSF‐1R) is one potential target that supports recruitment and maturation of immune‐suppressive macrophages in the pancreatic tumor microenvironment. The anti–CSF‐1R therapy cabiralizumab combined with anti–PD‐1 therapy showed early promise and a randomized phase II study is ongoing [36]. Another promising immune checkpoint is V‐domain Ig suppressor of T‐cell activation (VISTA), which is expressed by CD68+ macrophages. Interestingly, PDAC has significantly higher density of VISTA expression compared with melanoma, with preferential distribution in the tumor stroma [37]. This could be a novel strategy to target the stroma and immune system simultaneously. The best‐studied attempt to target the stroma directly with pegylated recombinant human hyaluronidase had disappointing results [38], but the stroma likely still represents an important resistance mechanism that must be overcome with novel approaches.

This study has several limitations, largely related to its nonrandomized design and small sample size. A larger, randomized study would be required to establish the definitive response rate, PFS, and OS with this combination. As the trial was initiated in 2011 when nab‐paclitaxel, FOLFIRINOX, and nanoliposomal irinotecan were not standard of care options, gemcitabine alone was used as the chemotherapy backbone. It is possible that greater synergy between chemotherapy and immunotherapy can be achieved with a more immunogenic chemotherapy regimen [39, 40]. The combination of FOLFIRINOX with either CTLA‐4 or PD‐1 inhibition would be limited by toxicity, but there are multiple ongoing trials using FOLFIRINOX and immune checkpoint blockade sequentially and combined with vaccine strategies (NCT01896869, NCT03806309). Using stereotactic body radiation therapy with immune checkpoint blockade to induce an abscopal effect is also an important area for further investigation with ongoing trials in PDAC (NCT02639026, NCT03245541) [41, 42].

Conclusion

This study established an MTD for gemcitabine and ipilimumab with a good overall safety profile. Although the observed response rate was similar to gemcitabine alone, the durability of responses suggests a component of immune activation that may warrant further investigation.

Acknowledgments

The study drug was provided with support from Bristol‐Myers Squibb.

Author Contributions

Conception/design: Suneel D. Kamath, Aparna Kalyan, Sheetal Kircher, Halla Nimeiri, Al Benson III, Mary Mulcahy

Provision of study material or patients: Aparna Kalyan, Sheetal Kircher, Halla Nimeiri, Al Benson III, Mary Mulcahy

Collection and/or assembly of data: Suneel D. Kamath, Aparna Kalyan, Angela J. Fought

Data analysis and interpretation: Suneel D. Kamath, Aparna Kalyan, Angela J. Fought

Manuscript writing: Suneel D. Kamath, Aparna Kalyan

Final approval of manuscript: Suneel D. Kamath, Aparna Kalyan, Sheetal Kircher, Halla Nimeiri, Angela J. Fought, Al Benson III, Mary Mulcahy

Disclosures

Aparna Kalyan: Bristol‐Myers Squibb (C/A, RF); Al Benson III: Bristol‐Myers Squibb, Guardant Health, Eli Lilly & Co., Exelixis, Purdue Pharma, inVentive Health Inc., Axio, Genentech, Bayer, Merck, Rafael Pharmaceuticals, Terumo, Taiho, Thera Bionic, LSK, Incyte Corporation (C/A), Acerta, Celegene, Advanced Acceleartor Applications, Novartis, Merck Sharp and Dohme, Taiho Pharmaceutical, Bristol‐Myers Suibb, Medimmune/AstraZeneca, Xencor (RF), Astellas, Axio, Infinity Pharmaceuticals, Bristol‐Myers Squibb, Amgen (other–data monitoring committee member); Halli Nimeiri: AbbVie (E, OI); Mary Mulcahy: BTG (RF). The other authors indicated no financial relationships.

(C/A) Consulting/advisory relationship; (RF) Research funding; (E) Employment; (ET) Expert testimony; (H) Honoraria received; (OI) Ownership interests; (IP) Intellectual property rights/inventor/patent holder; (SAB) Scientific advisory board

References

Author notes