Durvalumab as monotherapy and in combination therapy in patients with lymphoma or chronic lymphocytic leukemia: The FUSION NHL 001 trial

Abstract Background Studies suggest that immune checkpoint inhibitors may represent a promising strategy for boosting immune responses and improving the antitumor activity of standard therapies in patients with relapsed/refractory hematologic malignancies. Aims Phase 1/2 FUSION NHL 001 was designed to determine the safety and efficacy of durvalumab, an anti‐programmed death ligand 1 (PD‐L1) antibody, combined with standard‐of‐care therapies for lymphoma or chronic lymphocytic leukemia (CLL). Methods and Results The primary endpoints were to determine the recommended phase 2 dose of the drugs used in combination with durvalumab (durvalumab was administered at the previously recommended dose of 1500 mg every 4 weeks) and to assess safety and tolerability. Patients were enrolled into one of four arms: durvalumab monotherapy (Arm D) or durvalumab in combination with lenalidomide ± rituximab (Arm A), ibrutinib (Arm B), or rituximab ± bendamustine (Arm C). A total of 106 patients with relapsed/refractory lymphoma were enrolled. All but two patients experienced at least one treatment‐emergent adverse event (TEAE); those not experiencing a TEAE were in Arm C (diffuse large B‐cell lymphoma [DLBCL]) and Arm D (DLBCL during the durvalumab monotherapy treatment period). No new safety signals were identified, and TEAEs were consistent with the respective safety profiles for each study treatment. Across the study, patients with follicular lymphoma (FL; n = 23) had an overall response rate (ORR) of 59%; ORR among DLBCL patients (n = 37) was 18%. Exploratory biomarker analysis showed that response to durvalumab monotherapy or combination therapy was associated with higher interferon‐γ signature scores in patients with FL (p = .02). Conclusion Durvalumab as monotherapy or in combination is tolerable but requires close monitoring. The high rate of TEAEs during this study may reflect on the difficulty in combining durvalumab with full doses of other agents. Durvalumab alone or in combination appeared to add limited benefit to therapy.

assess safety and tolerability. Patients were enrolled into one of four arms: durvalumab monotherapy (Arm D) or durvalumab in combination with lenalidomide ± rituximab (Arm A), ibrutinib (Arm B), or rituximab ± bendamustine (Arm C). A total of 106 patients with relapsed/refractory lymphoma were enrolled. All but two patients experienced at least one treatment-emergent adverse event (TEAE); those not experiencing a TEAE were in Arm C (diffuse large B-cell lymphoma [DLBCL]) and Arm D (DLBCL during the durvalumab monotherapy treatment period). No new safety signals were identified, and TEAEs were consistent with the respective safety profiles for each study treatment. Across the study, patients with follicular lymphoma (FL; n = 23) had an overall response rate (ORR) of 59%; ORR among DLBCL patients (n = 37) was 18%. Exploratory biomarker analysis showed that response to durvalumab monotherapy or combination therapy was associated with higher interferon-γ signature scores in patients with FL (p = .02).
Conclusion: Durvalumab as monotherapy or in combination is tolerable but requires close monitoring. The high rate of TEAEs during this study may reflect on the difficulty in combining durvalumab with full doses of other agents. Durvalumab alone or in combination appeared to add limited benefit to therapy.  1,4 Clinical data also suggest that anti-PD-1/PD-L1 therapeutics, such as durvalumab, may improve the activity of lymphoma therapies including lenalidomide (an immunomodulatory drug), ibrutinib (a Bruton's tyrosine kinase inhibitor), bendamustine (an alkylating pro-apoptotic chemotherapeutic), and rituximab (a monoclonal anti-CD20 antibody) when given in combination. [5][6][7][8] Preclinical evidence in murine lymphoma models suggests synergistic antitumor activity with ibrutinib plus durvalumab. 3 Several reports have described a combination treatment approach with PD-L1/PD-1-targeted therapy, including ibrutinib and nivolumab in non-Hodgkin lymphoma (NHL) 9 and chronic lymphocytic leukemia (CLL) and lenalidomide with pembrolizumab in double-hit lymphoma. 10 The use of immune checkpoint inhibitors (ICIs) is also supported by the increased expression of PD-L1/PD-1 in tumorinfiltrating lymphocytes from patients with lymphoma and increased PD-1 expression in circulating T cells from patients with CLL. 11 Studies also showed benefits of using nivolumab or pembrolizumab in relapsed/refractory classical Hodgkin lymphoma (HL) 1,12 ; pidilizumab with rituximab in follicular lymphoma (FL) 13 ; and nivolumab in a small percentage of patients with diffuse large B-cell lymphoma (DLBCL), FL, or T-cell lymphomas. 14 Durvalumab, a monoclonal antibody that binds to PD-L1 to block its interaction with programmed cell death-1 PD-1, is approved in the United States and other countries for the treatment of several nonhematologic cancers. 15,16 Based on initial studies that suggested potential antitumor activity in hematologic malignancies with anti-PD-1/PD-L1 therapeutics, we hypothesized that the T-cell-mediated anti-tumor responses seen after treatment with durvalumab might act synergistically with standard-of-care treatments for lymphoma and CLL, ultimately resulting in enhanced efficacy without compromising safety. 17,18 The aim of the FUSION NHL 001 study was to determine the safety and tolerability of durvalumab when given in combination with lenalidomide ± rituximab, ibrutinib, or rituximab ± bendamustine in patients with lymphoma or CLL. The safety and efficacy of durvalumab were also assessed in patients receiving the recommended phase 2 dose (RP2D) of these treatment regimens as well as durvalumab monotherapy.
Previously reported data showed that PD-L1 expression is associated with poor prognosis in patients with DLBCL. 19 In addition, in patients with non-small cell lung cancer or urothelial cancer receiving durvalumab monotherapy, elevated PD-L1 expression and interferon (IFN)-γ scores in baseline tumor biopsy samples were associated with higher overall reaponse rate (ORR) or improved survival. 20,21 In the current study, an exploratory biomarker analysis was also performed to assess the expression of PD-L1, CD8, and an IFN-signature score signature 20 and how they associate with response to treatment with durvalumab.

| Study design and treatments
The trial design is presented in Figure 1.  Lymphocytic Leukemia Response Criteria for CLL. 23,24 The study was conducted in compliance with International Council for Harmonization Good Clinical Practice, and in accordance with the general ethical principles outline in the Declaration of Helsinki and applicable national, state, and local laws of the pertinent regulatory authorities. The protocol was approved by an institutional review board/independent ethics committee prior to commencement.

| Safety
Primary endpoints for the dose-finding part of the study were to determine the RP2D of each combination therapy (including doselimiting toxicities [DLTs]) and to assess the safety and tolerability of durvalumab when given in combination with lenalidomide and rituximab or bendamustine and rituximab. To assess safety, the incidence of treatment-emergent adverse events (TEAEs) using National Cancer Institute Common Terminology Criteria for Adverse Events version 4.03 was evaluated. During dose confirmation, primary endpoints assessed the safety of durvalumab as monotherapy and when given in combination with lenalidomide and rituximab, ibrutinib, or bendamustine and rituximab by examining the incidence of TEAEs.

| Efficacy
During dose-expansion, the primary preliminary efficacy endpoint was ORR based on tumor-specific response criteria (e.g., Lugano Classification and International Workshop on Chronic Lymphocytic Leukemia response criteria). Secondary endpoints for dose-finding and confirmation parts of the study included ORR based on tumor-specific response criteria, time to response (TTR), duration of response (DoR), and progression-free survival (PFS). Overall survival was also examined.

| Exploratory biomarker analyses
Among patients with lymphoma, biopsies were collected within 28 days before cycle 1 day 1 (mandatory) and any time during cycle 2 (strongly recommended); in some cases, archival formalin-fixed paraffin-embedded biopsies were used. Biopsy slides were sent to Q 2 Solutions/EA Genomics (Durham, NC, USA) and RNA was extracted using the Qiagen micro RNeasy FFPE Kit (Qiagen, Hilden, Germany). Quality and quantity of recovered RNA was determined by both Bioanalyzer (Agilent Technologies, Santa Clara, CA, USA) and Qubit (Thermo Fisher Scientific, Waltham, MA, USA). The DV200 metric (Agilent Technologies) and ribogreen RNA quantitation (Thermo Fisher Scientific) identified the percentage of RNA fragments >200 nucleotides. Following extraction, the TruSeq RNA Exome kit (library prep, RNA enrichment, index adapters; Illumina, San Diego, CA, USA) was used to create barcoded libraries, which were quantified. Each sample was processed using the TruSeq RNA Exome Kit and library material was loaded onto a HiSeq 2500 (Illumina) or equivalent. Alignment was performed using a two-pass mode with STAR (v2.5.2b) on the full hg38 human genome, DESeQ 25 was used to normalize counts with the function "CalcNormFactors," followed by the "cpm" function to generate counts

| Statistical analysis
Sample-size determination established that a maximum of 60 patients was required for the dose-finding part, and a maximum of 100 patients was required for the dose-confirmation part. Celgene, together with AstraZeneca/MedImmune, decided that the dose-expansion part was not to be opened. As a consequence, the entire study was to enroll a maximum of approximately 160 patients. All primary and secondary statistical analyses were conducted with SAS ® version 9.1 or higher (SAS Institute, Inc., Cary, North Carolina, USA). The safety population included all patients who received at least one dose of study drug. The efficacy population included all patients who completed at least one cycle of their assigned treatment and had baseline and at least one post-baseline tumor response assessment. If multiple values were present for the same date, the mean of these values was reported (for character parameters like urinalysis, the worst value was reported). Confidence intervals (CIs) were presented as two-sided 95% CIs unless specified. For translational analyses, a t-test (using R version 3.6) was used to compare the nonresponders with the responders within each histology group for IFN-γ signature score. This specific score and corresponding analysis approach were prespecified before generating the gene expression data.

| Confirmation of RP2D
Across the clinical program, the RP2D for durvalumab was previously determined to be 1500 mg Q4W in patients weighing more than 30 kg and was not examined in the current trial. 15 The safety review committee did not confirm the RP2D for rituximab or lenalidomide in In         19 in Arm C, and 9 in Arm D); 27 were related to disease progression, two from unknown causes, and one from respiratory failure.
Among patients with MCL, there was one death in Arm B due to a treatment-related AE (pneumonitis) that was attributed to both durvalumab and ibrutinib and one in Arm D due to disease progression.
There were two deaths among patients with CLL/SLL in Arm C, one due to an AE (sepsis) that was unrelated to study treatment and one due to a lung infection that occurred after the treatment period. There were three deaths in patients with HL in Arm D after the treatment period; two patients receiving monotherapy died from unknown causes after the treatment period and one patient receiving combination treatment died from respiratory failure.

| Efficacy
Treatment responses are shown for this study in Table 2    ibrutinib 560 mg, 75%). The worst response seen during the dosefinding part in these patients was stable disease; no patient had progressive disease. During the dose-confirmation part, patients in Arm B with CLL/SLL who received durvalumab + ibrutinib 420 mg had an ORR of 100% and patients in Arm B with MCL who received durvalumab + ibrutinib 560 mg had an ORR of 70%. For patients in Arms A, B, and C with FL (n = 23), the ORR was 59% (27% CR); the ORR among patients with DLBCL (n = 37) was 18% (8% achieved CR).
Median TTR, DoR, PFS, and OS this study are provided in

| Exploratory biomarker analysis
The biomarker analysis is represented in Figures 5 and 6. Notably, in our study, patient samples from patients with FL who responded to therapy across all of the arms of the study exhibited a slightly higher   (3) 10.0% (1) 100.0% (10)  with R/R CLL or SLL were randomly assigned to receive daily ibrutinib or the anti-CD20 antibody ofatumumab. In this study, the investigator-assessed ORR was significantly greater in the ibrutinib group than in the ofatumumab group (85% vs. 23%, p < .001). 27 Although caution should be made when making cross-study comparisons, especially due to the small numbers in our study, during the dose-confirmation part of FUSION NHL 001, patients treated with durvalumab plus ibrutinib 560 mg (n = 10) had and ORR of 70%, of which 30% were complete and 40% were partial responses; neither median PFS or median overall survival was reached. The phase 3 AUG-MENT study found that patients with R/R FL or marginal zone lymphoma (N = 358) treated with lenalidomide plus rituximab had improved ORR versus those receiving placebo plus rituximab (78% vs. 53%, p < .0001). 28 Although the dose-finding part of FUSION NHL 001 also included patients with DLBCL (n = 4), those treated with lenalidomide (either 10 or 20 mg) plus rituximab (n = 8) had an ORR of 75%. While most studies of rituximab plus bendamustine focus on treatment-naive disease, in a retrospective multicenter analysis of 55 Italian patients with R/R DLBCL the ORR after rituximab plus bendamustine was 50%, including a 28% complete remission rate; median overall survival was 10.8 months. 29  Biomarker analysis showed that slightly higher IFN-γ signature scores were correlated with response to therapy in patients with FL who responded to therapy across all of the arms of the study. This may suggest that the use of an IFN-γ gene signature may serve as a biomarker by which to enrich for patients that may be more responsive to anti-PD-L1-based therapy and will require further investiga-

DATA AVAILABILITY STATEMENT
The Bristol Myers Squibb policy on data sharing may be found at https://www.bms.com/researchers-and-partners/independentresearch/data-sharing-request-process.html

ETHICS STATEMENT
The study was conducted in compliance with the International Council for Harmonization Good Clinical Practice and in accordance with the general ethical principles outline in the Declaration of Helsinki and applicable national, state, and local laws of the pertinent regulatory authorities. The protocol was approved by an institutional review board/independent ethics committee prior to study commencement.