Phase 2 trial of ibrutinib and nivolumab in patients with relapsed CNS lymphomas

Lymphoma in the central nervous system (CNS) is a challenging disease with poor outcomes. Patients only have a few treatment options, particularly in relapsed disease, after high-dose methotrexate-based treatment, and the median overall survival (OS) is generally shorter than 6 months.^1-3 Relapsed CNS lymphoma may be a consequence of either primary CNS lymphoma (PCNSL) or recurrence of secondary CNS lymphoma (SCNSL), which is a relapse of a systemic lymphoma in the CNS that may have the distinct molecular characteristics of large B-cell lymphoma (LBCL).^4-8 PCNSL has a unique genomic profile with similarities to the activated B-cell subtype that may offer potential targets for therapy. The genetic aberrations commonly include MYD88 and/or CD79B mutations and are associated with an increased sensitivity to Bruton tyrosine kinase (BTK) inhibitors.⁹ In addition, translocation or copy number alterations of 9p24 are common, which may be associated with sensitivity to programmed death (PD)–targeted therapies.⁹ In contrast, SCNSL mirrors systemic diffuse LBCL (DLBCL), and cases may present with distinct biology when compared with PCNSL, particularly in cases with the germinal center B-cell (GCB) subtype, which has different molecular characteristics.¹⁰ All these studies support the importance of the development of biologic treatments that target the therapeutic vulnerabilities of CNS lymphomas.¹¹ 

Ibrutinib is an oral covalent BTK inhibitor that has demonstrated activity in patients with relapsed/refractory PCNSL and SCNSL as a single agent and in combination therapy. Ibrutinib as a single agent showed high efficacy with overall response rates (ORRs) of 50% to 80%, but the response duration was transient with a median duration of response often of <6 months.^12-18 Nivolumab is a monoclonal antibody PD-1 inhibitor, which has demonstrated evidence of single-agent activity in patients with CNS lymphoma in small case series.^19,20 Preclinical studies have shown possible synergy between ibrutinib and inhibition of the PD-1 pathway.²¹ BTK inhibition increases tumor-associated CD8⁺ T cells, leads to persistence of activated T cells, increases the T-cell repertoire, decreases the ratio of regulatory T cells to CD4⁺ T cells,^22,23 and may enhance the activity of nivolumab, in addition to a direct antilymphoma effect. Former trials have demonstrated the safety and efficacy of the combination of ibrutinib and nivolumab in various B-cell lymphoid malignancies.^24,25 Thus, we hypothesized that the combination of ibrutinib and nivolumab would demonstrate high efficacy with a tolerable toxicity profile in the treatment of CNS lymphoma. In this study, we report the results of a phase 2 single-center clinical trial (ClinicalTrials.gov identifier: NCT03770416) that evaluated ibrutinib in combination with nivolumab for the treatment of patients with relapsed or refractory CNS lymphoma.

This was an investigator-initiated, open-label, single-arm, phase 2 study that was conducted at The University of Texas MD Anderson Cancer Center. The trial had 2 sequentially enrolled cohorts. Cohort A received only 560 mg oral ibrutinib daily for 1 cycle (28 days per cycle), followed by ibrutinib in combination with 240 mg IV nivolumab every 14 days. Cohort B included the combination of ibrutinib and nivolumab from the initial cycle. Patients with a partial response or greater after 6 cycles of combined ibrutinib and nivolumab could continue therapy for up to 2 years or until progressive disease or unacceptable toxicity occurred. The initial plan was to enroll 20 patients in both cohort A and B, but enrollment closed early because of slow accrual owing to, in part, the COVID-19 pandemic.

The primary objective was to determine the best ORR of ibrutinib and nivolumab in relapsed/refractory CNS lymphomas. Secondary objectives included the ORR of ibrutinib as a lead-in agent before the combination in cohort A, the best complete response (CR) rate for the combination, the 1-year progression-free survival (PFS) and OS, and the safety of the combination. The response was assessed by standard response criteria from the International PCNSL Collaborative Group and Lugano classification.^26,27 Magnetic resonance imaging of the brain (and spine if positive for disease at enrollment) was performed for all patients after cycle 2, 4, and 6, and after the initial cycle of ibrutinib monotherapy for patients in cohort A (cycle 0), and a positron emission tomography/computed tomography scan was performed after cycle 6.

Adult patients (age ≥18 years) were eligible if they had CNS lymphoma (primary or secondary) refractory to or relapsed after at least 1 previous line of therapy. Pathology was confirmed either by biopsy or by cerebrospinal fluid review. All patients underwent magnetic resonance imaging of the brain and/or spine, a lumbar puncture, and a positron emission tomography scan as part of the staging procedure during the screening period. Patients were required to have an Eastern Cooperative Oncology Group performance status of ≤2 and adequate organ and bone marrow function. Patients were ineligible if they previously received BTK inhibitor or PD-1 inhibitor therapy, had an active autoimmune disease, a requirement for high-dose steroids (>10 mg daily of prednisone or equivalent), an active HIV infection, or active hepatitis B or C infection.

Adverse events (AEs) were defined according to the National Cancer Institute Common Terminology Criteria for Adverse Events version 4.0. This study was approved by the institutional review board of the MD Anderson Cancer Center and conducted in accordance with the Declaration of Helsinki.

Descriptive statistics were used to define the baseline characteristics and treatment response/outcomes. Fisher exact tests and Wilcoxon rank-sum tests were used to evaluate the association between the end of therapy response and categorical variables and continuous variables, respectively. The Kaplan-Meier method was used to estimate the probabilities of PFS and OS, and the log-rank test was used to evaluate differences in the PFS and OS. PFS was defined as the time from study entry to objective disease progression or death from any cause, whichever occurred first. OS was defined as the time between study entry and death from any cause. Patients who completed the study and who were still alive at the time of the data cutoff date or who underwent stem cell transplant (SCT) for consolidation were censored at the last visit date or date of transplant.

A total of 18 patients were accrued (10 patients in cohort A and 8 patients in cohort B) from February 2019 to June 2022, and all patients were evaluable for response. The median age was 63 years (range, 43-88); 16 patients had PCNSL and 2 patients had SCNSL (Table 1). Five patients (28%) had a performance status of ≥2, 8 patients (44%) had lactate dehydrogenase levels above the upper normal limit. All patients had parenchymal disease, and no patients had radiographic spinal disease or concurrent systemic disease involvement. One patient had lymphoma involvement in the cerebral spinal fluid. Of those enrolled, 11 patients had non-GCB, 4 patients had GCB (2 patients each with PCNSL and SCNSL), and 3 were unable to be classified because of missing data of the Hans algorithm.²⁸ The median number of previous lines of therapy was 2 (range, 1-4); only 2 patients with SCNSL received 4 lines of treatment, including chimeric antigen receptor T-cell therapy (both for systemic DLBCL before SCNSL progression). Patients with PCNSL received 1 or 2 previous lines of treatment, and 3 patients previously underwent autologous SCT (auto-SCT). Ten patients had refractory disease. The median time from diagnosis to study treatment was 14.9 months.

For all patients, the best ORR was 77.8% (n = 14/18; 95% confidence interval [CI], 52-94) and the best CR rate was 50% (95% CI, 26-74). In cohort A, the ORR of ibrutinib as a lead-in agent after 1 cycle was 90% (95% CI, 55.5-99.7), and the CR rate was 60% (Figure 1). In cohort B, the ORR was 62.5% (95% CI, 24.5-91.5), the CR rate was 12.5% (95% CI, 0.3-52.7), and the CR rate was higher in cohort A (P = .008). In GCB patients (n = 4), the ORR and CR rates were 50%. In non-GCB patients, the ORR and CR rates were 91% and 45%, respectively, and there was no significant difference in the response rate by cell of origin. With a median follow-up time of 31 months, the median PFS and OS was 6.5 months (95% CI, 1.3 to not reached) and 21.0 months (95% CI, 8.3 to not reached), respectively. The 1-year PFS and OS rates were 42% (95% CI, 15-67) and 63% (95% CI, 36-82), respectively (Figure 1). Four patients withdrew from protocol therapy during CR, including 2 who underwent auto-SCT consolidation, and 2 who stopped because of fatigue. Three patients withdrew from protocol therapy during partial response; 2 because of mucositis and 1 because of arthralgias (patients 2, 4, and 8 in Figure 2). One patient each died of transplant-related complications (patient 15) and COVID-19 infection (patient 14) with CR after 4 cycles of therapy. There were 3 patients (17%) who have been in remission for >2 years (Figure 2); all these patients had PCNSL, and 1 had non-GCB, another 1 had GCB subtype, and 1 patient was unclassifiable because of limited sample availability.

The treatment was generally well tolerated. Treatment-related AEs occurred in all patients, and 9 (50%) patients experienced at least 1 grade 3 to 4 treatment-related AE. There was no grade 5 toxicity. Common any grade AEs that occurred in >25% of patients included fatigue (50%), nausea (33%), and mucositis (28%). Grade 3 to 4 AEs of interest included neutropenia and oral mucositis in 2 patients each and pneumonitis, arthritis, and maculopapular rash in 1 patient each. Two patients discontinued nivolumab because of flare-up of arthritis and rash. Details of AEs are summarized in Table 2. There were no atrial fibrillation or bleeding observed in the trial.

The combination of ibrutinib and nivolumab for the treatment of relapsed/refractory CNS lymphoma showed a high ORR of 77.8%, and responses were durable for >2 years without further treatment in 17% of patients. Cohort A, which started ibrutinib as a single lead-in agent, showed a numerically higher CR rate when compared with cohort B, which started with the combination from cycle 1, although the interpretation is limited by the small number of patients. Potential reasons for this difference include response evaluation timing (after 1 cycle in cohort A and after 2 cycles in cohort B; 2 responding patients in cohort A suffered disease progression after 2 additional cycles), the immune effects of single-agent ibrutinib (interleukin-2 inducible T-cell kinase inhibition in T cells, persistence of activated T cells, increases T-cell repertoire, decrease the regulatory T cells/CD4⁺ T-cell ratio through BTK-dependent and BTK-independent mechanisms, which leads to enhanced activity of nivolumab), and/or random chance.^23,29 Responses were seen in patients with both non-GCB and GCB subtypes with 2 of 4 patients with GCB (1 each for PCNSL and SCNSL) achieving CR. Intriguingly, these 2 patients had a durable remission. Unfortunately, 1 patient died during CR with COVID-19 pneumonia, but 1 patient remained in CR for >2 years without further treatment. The toxicities were generally manageable, and the combination did not seem to demonstrate excess toxicity, although we observed a few immune related toxicities, such as rash and flare-up of arthritis, that led to treatment discontinuation, which is similar to the observations in the LYM1002 trial of the same combination.²⁵ 

Single-agent ibrutinib showed clinical activity in PCNSL and LBCL and has been used in the treatment of relapsed/refractory CNS lymphomas. The response rate and median PFS from single-agent ibrutinib in previous studies were 50% to 80% and ∼4 to 5 months, respectively, in CNS lymphomas.^13,16 Nivolumab caught major attention in the treatment of PCNSL based on the case series of 4 patients with relapsed/refractory PCNSL; all patients responded to nivolumab, and 3 of them achieved CR.²⁰ However, the formally evaluated response rate of nivolumab in relapsed/refractory PCNSL and primary testicular lymphoma in the phase 2 CheckMate 647 trial was rather disappointing with an ORR of 6.4%.³⁰ The activity of a checkpoint inhibitor as a single agent in the treatment of PCNSL may be limited. However, in the trial that treated patients who were deemed poor candidates for other consolidation, such as auto-SCT or radiation, nivolumab consolidation after a high-dose methotrexate regimen showed safety and potential benefit.³¹ In the LYM1002 trial that evaluated the nivolumab and ibrutinib combination in patients with relapsed/refractory non-Hodgkin lymphoma or chronic lymphocytic leukemia, the ORR was similar to single-agent ibrutinib in general,²⁵ except for Richter’s transformation,³² suggesting that the combination may provide benefit in certain diseases. In this trial, all patients had parenchymal disease and still we observed durable responses in a portion of patients without further treatment, which may be related to the impact of the combination based on previous studies that evaluated ibrutinib as single agent and that reported few if any long-term responses in patients with parenchymal/spinal disease.¹³ The results of both the previous trials and the current trial further accentuate the importance of discovering biomarkers for biology-driven treatment in CNS lymphomas.

Hodgkinson et al evaluated biomarkers of response to ibrutinib and nivolumab in the LYM1002 trial.³³ In DLBCL, elevated PD-L1 expression (≥5%), measured using immunohistochemistry, was associated with a higher ORR (62.5% vs 18.8%) and CR rate (37.5% vs 0%). Intriguingly, patients with mutations associated with B-cell receptor signaling, such as MYD88 or TNFRSF14, showed a lower likelihood of response to ibrutinib and nivolumab, although these mutations in the B-cell receptor signaling pathway provided a strong rationale for the treatment of PCNSL with BTK inhibitors.¹⁸ In addition, patients with the GCB subtype responded reasonably well (33.3%) to ibrutinib and nivolumab in the LYM1002 trial, similar to what was observed in this trial, which is unexpected based on ibrutinib single-agent trials that demonstrated a disappointing ORR of 5% in GCB DLBCL.³⁴ These studies indicate the complexity of biology and the potential risk of oversimplifying biomarkers. Gandhi et al recently identified strikingly distinct genetic and immunologic features in subtypes of PCNSL based on Epstein-Barr virus (EBV) status, which potentially suggest treatment implication.^5,35 EBV negative PCNSL typically showed a gene expression phenotype similar to activated B-cell DLBCL, including mutations in MYD88 and CD79B, and also showed genetic features of immune evasion, such as mutations in or deletions of HLA-A, HLA-B, or HLA-C.⁵ EBV⁺ cases were often observed among patients who were immunocompromised and exhibited a tolerogenic tumor microenvironment, including overexpression of genes associated with T-cell immune checkpoints (PD-L1, PD-L2, LAG-3, and TIM-3), macrophages (CD68 and CD163), and the antiviral cytokine tumor necrosis factor-α. Although it should be acknowledged that developing biomarkers based on previous studies is challenging, this simple marker of EBV can be valuable. These findings ideally should be validated in prospective trials to evaluate the implications of treatment decisions further, however, we were unable to validate these findings because of limited samples and the number of patients in the current study.

In conclusion, ibrutinib and nivolumab showed promising clinical activity and durability with manageable toxicity, highlighting the potential of nonchemotherapy-targeted therapy combinations for patients with refractory/relapsed CNS lymphoma. The trial was unfortunately closed because of slow recruitment, showing the significant challenge of clinical investigation in CNS lymphoma and accentuating the unmet need of biology-driven, well-designed studies that maximize the sample size of small patient numbers. In future trials, newer technology, such as circulating tumor DNA, should be used to overcome the challenge of sample collection to further characterize CNS lymphoma and to evaluate the mechanism of response and resistance to various treatments.³⁶ Multicenter collaboration is essential to facilitate research in this rare and challenging lymphoma.

Contribution: J.R.W. designed the trial; D.C. and L.F. analyzed the data; D.C. wrote the draft of the manuscript; R.E.S., R.N., S.A., P.S., L.M., F.S., M.A.R., F.B.H., L.E.F., S.P.I., L.J.N., and S.S.N. provided patient care in the trial; D.P.G., S.A.M., W.M., and G.M. supported conduct of the trial; and all authors reviewed and contributed to the final manuscript.

Conflict-of-interest disclosure: D.C. reports receiving research funding from Genentech, Bristol Myers Squibb (BMS), Genmab, MorphoSys, and ONO Pharmaceutical. R.E.S. reports receiving research funding from BMS, GlaxoSmithKline (GSK), and Rafael Pharmaceuticals. R.N. reports serving as advisory board member/consultant for 280 Bio, Inc. S.A. reports receiving research funding from Nektar, Merck, Xencor, Chimagen and Genmab, Kite/Gilead, Janssen, and Caribou; serving as a member of the scientific advisory committee for Chimagen; serving on the data safety monitoring board for Myeloid Therapeutics; and serving as an advisory board member or consultant for ADC Therapeutics and Kite/Gilead. P.S. reports receiving research funding from Sobi, AstraZeneca-Acerta, ALX Oncology, and ADC Therapeutics, and serving as an advisory board member/consultant for Roche-Genentech, AbbVie-Genmab, Ipsen, Kite/Gilead, Hutchison MediPharma, AstraZeneca-Acerta, ADC Therapeutics, Sobi, and TG Therapeutics. L.M. reports receiving research funding from Dizal Pharma and BMS. S.P.I. reports receiving research funding from Merck, Innate, AstraZeneca, Dren Bio, CRISPR Therapeutics, Legend, Pfizer, ONO Pharmaceutical, Acrotech, and Corvus; serving as an advisory board member or consultant for Salarius, Electra, Seagen, Dren Bio, Electra, and Secura Bio; and owning stock in IMPaRT.ai. L.J.N. reports receiving research funding from BMS, Caribou Biosciences, Daiichi Sankyo, Genentech, Genmab, Gilead/Kite, Janssen, Incyte, Ipsen, Merck, Novartis, and Takeda, and serving as an advisory board member or consultant for AbbVie, BMS, Caribou Biosciences, Genentech, Genmab, Gilead/Kite, Janssen, Incyte, Ipsen, Merck, Novartis, Regeneron, and Takeda. S.S.N. reports receiving research support from Kite/Gilead, BMS, Allogene, Precision Biosciences, Adicet Bio, Sana Biotechnology, and Cargo Therapeutics; serving as an advisory board member or consultant for Kite/Gilead, Merck, Sellas Life Sciences, Athenex, Allogene, Incyte, Adicet Bio, BMS, bluebird bio, Fosun Kite, Sana Biotechnology, Caribou, Astellas Pharma, MorphoSys, Janssen, Chimagen, ImmunoACT, Orna Therapeutics, Takeda, Synthekine, CARsgen, Appia Bio, GSK, Galapagos, ModeX Therapeutics, and Jazz Pharmaceuticals; having stock options from Longbow Immunotherapy, Inc; and owing intellectual property related to cell therapy. J.R.W. reports receiving research funding from ADC Therapeutics, Allogene, AstraZeneca, BMS, Genentech, Janssen, Kite/Gilead, MorphoSys/Incyte, Novartis, and Nurix; and serving as an advisory board member or consultant for AbbVie, ADC Therapeutics, Allogene, AstraZeneca, BMS, Genentech, Genmab, Janssen, Kite/Gilead, MorphoSys/Incyte, Novartis, Nurix, Pfizer, Regeneron. The remaining authors declare no competing financial interests.

Correspondence: Jason R. Westin, Department of Lymphoma and Myeloma, The University of Texas MD Anderson Cancer Center, 1515 Holcombe Blvd, Houston, TX 77030; email: jwestin@mdanderson.org.