Treatment of patients with MYC rearrangement positive large B-cell lymphoma with R-CHOP plus lenalidomide: results of a multicenter HOVON phase II trial

Abstract

Patients with MYC-rearrangement positive large B-cell lymphoma (MYC+ LBCL) have an inferior prognosis following standard first-line therapy with rituximab, cyclophosphamide, doxorubicin, vincristine, and prednisolone (R-CHOP) as compared to patients without MYC rearrangement. Although intensive chemotherapy regimens yield higher remission rates, toxicity remains a concern. Lenalidomide is an oral immunomodulatory drug which downregulates MYC and its target genes thereby providing support using lenalidomide as additional therapeutic option for MYC+ LBCL. A phase II trial was conducted evaluating the efficacy of lenalidomide (15 mg day 1-14) in combination with R-CHOP (R2CHOP) in newly diagnosed MYC+ LBCL patients identified through a nationwide MYC-FISH screening program. The primary endpoint was complete metabolic response (CMR) on centrally reviewed 18F-fluorodeoxyglucose (18F-FDG) positron emission tomography (PET)-computer tomography (CT)-scan at end-of-treatment. Secondary endpoints were overall survival (OS), disease-free survival (DFS) and event-free survival (EFS). Eighty-two patients with stage II-IV MYC+ LBCL were treated with 6 cycles of R2CHOP. At EOT, 67% (confidence interval (CI) 58-75%) of the patients reached CMR. With a median follow-up of 25.4 months, 2-year estimates (95% CI) for OS, DFS, EFS were 73% (62-82%), 75% (63-84%) and 63% (52-73%) respectively. In this prospective trial for newly diagnosed MYC+ LBCL patients, we found that administering R2CHOP was safe, and yields comparable CMR and survival rates as in studies applying more intensive chemotherapy regimens. Hence, these findings offer new prospects for MYC+ LBCL patients and warrant comparison in prospective randomized clinical trials. This trial was registered at www.clinicaltrialsregister.eu (#2014-002654-39).

Introduction

Diffuse large B-cell lymphoma (DLBCL) comprises about 35% of all non-Hodgkin lymphomas (NHL) and is the most common lymphoma subtype.¹ The outcome of patients treated with rituximab, cyclophosphamide, doxorubicin, vincristine, and prednisolone (R-CHOP) is heterogeneous for which the IPI score and cell-of-origin (COO) are the most well-known denominators.^2,3 MYC rearrangement status is an independent prognostic factor, and is reported in 10-15% of DLBCL patients (hereafter MYC+ LBCL).^4-7 In about 30% of these patients, only a single MYC rearrangement is found (single hit [SH]), while in 70% MYC rearrangement is detected together with either a BCL2 or BCL6 rearrangement (double hit [DH]: MYC+/BCL2+ or MYC+/BCL6+) or with both (triple hit [TH]: MYC+/BCL2+/BCL6+).⁴ It has been shown that in patients with MYC+ LBCL, standard first-line therapy with R-CHOP results in an inferior prognosis compared to those without MYC rearrangement (2-year OS 35% vs. 61%8 and 5-year OS 31% vs. 66%6). Moreover, patients with MYC+ LBCL have an increased risk of central nervous system (CNS) relapse.^5,6 Recently, Rosenwald et al. demonstrated that the inferior prognosis of MYC rearranged patients is however largely observed in patients with DH/TH lymphoma. ⁷ In the revised World Helath Organisation (WHO) 2017 classification, SH is not recognized as a separate entity in contrast to DH/TH lymphoma.¹

In search for improvement, intensified chemotherapy regimens, such as hyper-CVAD and R-CODOX-M/RIVAC, have been investigated. Data mainly come from subanalyses of MYC rearrangement positive patients in trials designed for unselected DLBCL patients. These studies indicate that intensified treatment results in improvement of progression free survival (PFS), but not OS.^9-11 Only recently, a prospective, multicenter, single arm phase II study specifically designed for MYC+ LBCL patients showed that DA-EPOCH-R resulted in a promising CMR rate at end of treatment (EOT) of 74% and 4 year EFS and OS of 71% and 77%, respectively.¹²

Lenalidomide is an oral immunomodulatory drug with direct antitumor effects and indirect effects on the tumor microenvironment.¹³ In vitro studies have demonstrated that lenalidomide exposure results in down-regulation of MYC and its target genes via cereblon and IRF4 in lymphoid cells, thereby providing the rationale for introdu-cing lenalidomide as a therapeutic option in MYC+ LBCL.¹⁴ Two phase II studies in ABC/non-GCB-subtype DLBCL have demonstrated that the addition of lenalidomide to R-CHOP (R2CHOP) is indeed feasible and may contribute to a favorable outcome by decreasing CNS relapse.^15,16 Against expectation, R2CHOP did not result in a survival advantage in ABC-subtype DLBCL, as has recently been shown in a phase III study (ROBUST).¹⁷

The present study reports the results of a prospective single- arm phase II trial forMYC+ LBCL patients treated with R2CHOP. Patients were identified through a nationwide molecular biomarker diagnostics program. We report outcome based on the primary endpoint, which was CMR by centrally reviewed a ¹⁸F-fluorodeoxyglucose (¹⁸F-FDG) positron emission tomography (PET)-computer tomography (CT) scan at EOT, as well as 2-year OS, DFS and EFS rates.

Methods

Screening program and patient eligibility

To support timely diagnosis of MYC+ LBCL and optimal enrolment in the present clinical trial, a nationwide diagnostic support program for MYC rearrangement assessment by fluorescence in situ hybridization (FISH) was implemented.¹⁸

Patients ≥18 years with newly diagnosed DLBCL or with B-cell lymphoma, unclassifiable, with features intermediate between DLBCL and Burkitt lymphoma (BLC-U) according to the WHO 2008 classification with a proven MYC rearrangement by FISH analysis including SH (not Burkitt lymphoma), DH or TH DLBCL were eligible.

During the screening period one cycle of R-CHOP, a short course of steroids, or irradiation to control local symptoms was allowed. Patients with Ann Arbor stage II-IV, a WHO performance status (PS) of 0-3, ≥ one lesion of ≥1.5 cm on a contrast-enhanced CT scan and ≥one positive lesion on PET-CT scan were eligible.

Patients diagnosed with any other subtype of aggressive B-cell lymphoma, a history of follicular lymphoma, proven CNS localization or HIV positivity were excluded.

Treatment

Treatment consisted of six cycles of standard R-CHOP every 3 weeks plus lenalidomide 15 mg orally on day 1-14 (R2CHOP; Online Supplementary Table S1), followed by two additional rituximab administrations.

Prophylactic intrathecal methotrexate or cytarabine (≥4 administrations), pegfilgrastim, venous thromboembolism prophylaxis (with aspirin or low-molecular-weight-heparin), and Pneumocystis prophylaxis were mandatory.

Safety assessments

Adverse events were graded according to National Cancer Institute Common Terminology Criteria for Adverse Events (AE), version 4.03. AE grade 1 were not reported.

Study overview

This multicenter, phase II study was designed by investigators of HOVON and was approved by the medical ethics committee of the Amsterdam UMC. All patients provided written informed consent. The trial was conducted in accordance with the Declaration of Helsinki and Good Clinical Practice guidelines. An independent data and safety monitoring board conducted a review during the planned interim analysis.

Endpoints

The primary endpoint of the study was CMR on EOT PET-CT scan as determined by central review plus EOT bone marrow (BM) examination in case of BM localization at diagnosis. In case a BM examination was not repeated at EOT in patients with baseline BM localization and the EOT-PET scan showed no BM uptake localization, the response was classified as CMR (based on recent findings that CMR on PET-CT has a high negative predictive value for BM localization).^23,24

Secondary endpoints were: OS defined as time from registration to death; DFS defined as time from achievement of first CMR on protocol until relapse or death whichever comes first; EFS defined as the time from registration to lack of CMR on EOT-PETCT, relapse or death; and positive predictive value (PPV) and negative predictive value (NPV) of iPET-CT for EOT result.

Statistical analyses

An optimal Simon two-stage design was used with a response rate of 45% as the null hypothesis, and 60% as the alternative hypothesis. With a statistical significance level of 5% and a power of 80%, the required number of patients was 77, with an interim analysis for futility involving the first 26 included patients. In order to overcome dropouts due to ineligibility, 85 patients were enrolled. All efficacy analyses were restricted to eligible patients, while safety analyses included all enrolled patients. Data cut-off was June 28, 2019.

For the clinical protocol, central pathology review, central PETCT review and additional statistical information, see the Online Supplementary Data.

Results

Clinical characteristics

From April 2015 to February 2018, 85 patients were included from 20 hospitals in the Netherlands and Belgium. Three patients were declared ineligible (two because the MYC+ status was based on immunohistochemistry and not on FISH and one because of a transformed lymphoma), leaving 82 patients for efficacy and 85 patients for safety analyses.

Baseline patient and disease characteristics are shown in Table 1. The median age was 63 years (range: 28-82 years). 49 of 81 patients (60%) had a WHO performance status (PS) of 0; 58 of 71 patients (71%) had stage IV disease, and 42 of 82 patients (51%) had ≥2 extranodal localizations. The IPI score was high-intermediate and high in 65% of patients.

During treatment 12 of 82 patients went off protocol before completion (progressive disease [n=7], toxicity [n=2]; pulmonary embolism and diarrhea, other reasons [n=3]; new diagnosis of colon cancer, patient refusal, and vertebral fracture), see Figure 1.

Pathology review

Diagnostic biopsy samples of all 85 patients were available for pathology review. Results of all 82 eligible patients are summarized in Table 1 and the Online Supplementary Table S2. A diagnosis of DLBCL according to the WHO 2008 classification was confirmed in 65 of 82 patients (79%) and BCL-U in 12 of 82 patients (15%) and morphology was indecisive between DLBCL and BCL-U in 5 of 82 patients (6%). For classification according to the WHO classification 2017 see the Online Supplementary Table S2.

In 81 of 82 patients MYC rearrangement was confirmed at central review. Based on the intention to treat principle, the one patient in whom MYC rearrangement could not be confirmed was included in all analyses. In 9 of 82 cases, insufficient material was available to perform additional BCL2 and BCL6 rearrangement. In 73 of 82 cases, data on BCL2 and BCL6 rearrangement were available: 20 of 82 (26%) had a single MYC rearrangement (SH); 44 of 82 (54%) had DH lymphoma (31 patients had MYC/BCL2 rearrangements and 13 patients MYC/BCL6 rearrangements), and 9 of 82 (11%) had all three rearrangements (TH).

COO classification using a standard Hans algorithm showed GCB phenotype in 63 of 71 (89%) and non-GCB phenotype in 8 of 71 (11%). Lymph2Cx classification was performed in 38 cases showing GCB-subtype in 29 of 38 patients (76%), ABC-subtype in 7 of 38 patients (18%), and intermediate subtype in 2 of 38 patients (5%). Out of the 24 DH of TH patients, 21 showed GCB-subtype and 3 ABC-subtype. Out of the 12 SH patients, 8 showed GCB-subtype and 4 ABC-subtype.

Treatment

Most patients (n=68) started with lenalidomide in the second cycle and continued lenalidomide for 14 days after the sixth cycle of R-CHOP. When MYC FISH results were available at diagnosis, R2CHOP was started in the first cycle (n=14) (Figure 1).

Patients received a median (interquartile range [IQR]) dose of the planned drugs in the R-CHOP regimen as follows: cyclophosphamide 99.9% (99.0-101); vincristine 100% (72.5-100); doxorubicin 99.5% (97.7-101); prednisone 100% (100-100); rituximab 98.1 (95.1-100); pegfilgrastim 100% (100-100). Lenalidomide was given at a median dose intensity of 100% (range: 85.7-100). 57 of 82 patients (70%) received the planned ≥4 intrathecal prophylactic administrations.

Primary endpoint: CMR at EOT

At EOT PET-CT, 55 of 82 patients (67%) reached the primary endpoint of CMR (95% CI: 58-75, P<0.001), 5 of 82 patients (6%) reached a partial metabolic response (PMR), and 21of 82 patients (26%) had progressive metabolic disease (PMD) (Table 2).

One patient went off protocol due to toxicity after cycle 5 without EOT PET-CT (response unknown).

Univariate logistic regression analysis of baseline characteristics (BM localization, WHO performance status, stage, B symptoms, IPI, number of extranodal sites and age) did not reveal any significant predictors for reaching CMR.

Exploratory descriptive subgroup analyses revealed no differences between SH and DH/TH patients regarding achievement of the primary endpoint: CMR rate in both groups was 70% and 66% respectively (nine patients with unknown BCL2 and BCL6 rearrangement not included).

Secondary endpoints: survival analyses

With a median follow-up of 25.4 months (IQR 18.3-30.3), 1-year OS was 85% (95% CI: 76-91), DFS 77% (95% CI: 65-85) and EFS 66% (95% CI: 54-75). 2-year estimates for OS, DFS, EFS were and 73% (95% CI: 62-82), 75% (95% CI: 63-84) and 63% (95% CI: 52-73) respectively (Figure 2A-C).

Baseline patient characteristics (BM localization, WHO performance status, stage, B symptoms, IPI, number of extranodal sites and age) were not significantly predictive for prolonged OS in a univariate analysis at the 5% significance level.

Univariate regression analyses indicated that SH and DH/TH patients had comparable EFS and DFS, however DH/TH patients had a tendency for a higher risk of death compared to SH patients (Hazard ratio [HR]4.18, P=0.055; 95% CI: 0.97-18.02) (Online Supplentary Figure S1A-C). Separate analyses of DH MYC/BCL2 and DH MYC/BCL6 and TH in comparison to SH revealed no significant differences in OS (Online Supplementary Figure S2A-B).

In univariate analyses with response as time dependent covariate we found that patients who had achieved CMR at EOT PET experienced a reduced risk of death compared to patients who had not achieved CMR (HR 0.1, 95% CI: 0.03–0.33, P<0.001), (Figure 3). EOT PET-CT predicted relapse within 12 months, with a positive predictive value (PPV) of 81% and a negative predictive value (NPV) of 93% (Online Supplementary Table S3A).

In total, 29 patients showed progressive disease (11 without achieving CMR, 18 after achieving CMR [(at interim or EOT PET-CT)] including one patient with a CNS relapse.

Safety

Grade 2, 3 and 4 AE were seen in 27 (32%), 33 (39%) and 14 (16%) of all 85 registered patients respectively (Table 3). The most common grade 3–4 AE were neutropenia (18%), infections (14%) and gastrointestinal disorders (14%). Four patients experienced deep venous thrombosis (grade 2), and two patients pulmonary embolism (grade 3). Two of these patients (one with deep venous thrombosis and one with pulmonary embolism) had not received the mandatory thrombosis prophylaxis (protocol violation). One patient went off protocol due to grade 3 diarrhea.

71 serious AE were reported in 36 patients; 66 were due to hospitalization (42% infections, 26% gastrointestinal disorders), four to other conditions [two second primary malignancies, two recurrence of previously diagnosed (>5 year) malignancies]. One patient died during treatment due to progression. There were no treatment related deaths.

Observational analysis: predictive value of iPET-CT

At iPET-CT after three cycles of R2CHOP, 57 of 82 patients (70%) were in CMR; of these 45 of 57 (79%) were still in CMR and 11 of 57 (19%) showed PMD at EOT PETCT, and one missed EOT evaluation (Table 2). 23 of 82 patients (28%) were in PMR at iPET-CT; 10 of 23 (43%) of these converted to CMR, 4 of 23 (17%) remained in PMR, 9 of 23 (39%) showed PMD at EOT. The PPV of iPET-CT for predicting EOT PET-CT result was 60% (15 of 25), the NPV 79% (45 of 57) (Online Supplementary Table S3B).

Discussion

From retrospective series it is clear that first-line R-CHOP therapy is not sufficiently effective for patients with MYC+ LBCL with CR rates of 40-50% and 3-year OS rates of 35% only.^6,8 Intensified chemotherapy regimens such as R-CODOX-M/R-IVAC or autologous stem cell transplantation have not improved OS, and result in increased toxicity.^5,9-11

We designed a prospective clinical trial for MYC+ LBCL patients, in which a time window of one cycle of R-CHOP was allowed to perform molecular diagnostics. This approach permitted high risk patients to start treatment immediately and overcame the bias of inclusion of mainly lower risk patients due to enrolment delays.²⁵ In this trial we show that the addition of lenalidomide to R-CHOP resulted in EOT CMR rate of 67% CMR and 2-year survival rates of 73%, 75%, and 63% for OS, DFS and EFS respectively. To our knowledge, this is the second prospective trial especially designed for MYC+ LBCL patients. Recently, Dunleavy and colleagues reported a single arm phase II study in which the efficacy of DAEPOCH- R for MYC+ LBCL patients was explored.¹² Results for EOT CMR and survival rates are largely comparable between both approaches with EOT CMR rate of 74%, 4-year OS rates of 77% and EFS of 71% in the study of Dunleavy. When compared to the trial of Dunleavy, our patient population was larger (82 vs. 53 patients), comparable in age (median 63 vs. 61 years) but included more patients with IPI ≥3 (65% vs. 49%) and more patients with DH/TH (65% vs. 45%). Regarding safety, grade 3/4 infections were seen in 24% of cycles with DA-EPOCH-R versus grade 3 (and no grade 4 infections) in only 2,8 % of cycles (18 episodes) with R2CHOP. DA-EPOCH-R resulted in three treatment related deaths vs. none with R2CHOP.

Table 1.Patient demographics and disease characteristics.

Figure 1.Disposition of the patients. Eighty-two patients were included. In 14 patients, MYC fluorescence in situ hybridization (FISH) was performed immediately at diagnosis, these patients started with R2CHOP (lenalidomide in combination with rituximab cyclophosphamide, doxorubicin, vincristine, and prednisolone) in cycle 1. In 68 patients, MYC results became available during the first cycle of R-CHOP; these patients were registered after the first cycle of R-CHOP and started with R2CHOP in the second cycle and continued lenalidomide for 14 days after the sixth cycle of R-CHOP. During treatment 13 patients went off protocol (progressive disease [n=7], toxicity [n=2; pulmonary embolism and diarrhea], other reasons [n=3; new diagnosis of colon cancer, patient refusal, and vertebral fracture]). R: rituximab, R2: rituximab + lenalidomide

Figure 2.Survival analyses. (A) Overall survival (OS; time from registration to death, n=82); (B) disease-free survival (DFS; time from achievement of first complete metabolic response [CMR] on protocol until relapse or death whichever comes first, n=69); (C) event-free survival (EFS; defined as the time from registration to lack of CMR on end of treatment [EOT] positron emission tomography [PET]-computer tomography [CT] scan, relapse or death, n=82) of MYC+ LBCL patients.

Table 2.Response rates on interim and EOT PET-CT scan.

Lenalidomide penetrates the CNS, and thereby may aid to prevent CNS relapses as has been suggested for nongerminal center B-cell (GCB) subtype lymphoma patients treated with R2CHOP.¹⁵ Indeed, in this study, which combined lenalidomide and intrathecal prophylaxis, a remarkably low rate of CNS relapse at a median follow-up of 25.4 months was seen (n=1).

Several remarks regarding our study can be made. First, although correlation CMR at EOT PET-CT with survival in MYC+ LBCL has been described in a retrospective study,²⁶ one might argue that it is not an ideal primary endpoint. Given the high FDG-avidity of MYC+ LBCL and the fact that CMR at EOT PET-CT in our study was highly predictive for DFS (NPV of 93%), we feel that using CMR at EOT PET-CT as a surrogate endpoint for highly aggressive B-cell lymphomas such as MYC+ LBCL is justified.

Second, clinical prognostic markers, including age, stage, IPI score, as well as COO were not significantly correlated to CMR on EOT PET-CT and survival, which might be explained by the inclusion of high risk patients; 65% of our patients have an IPI score of ≥3, versus only 27% in Ziepert’s meta-analysis of the value of IPI in the rituximab era.²

Figure 3.Survival according to end-oftreatment PET-CT scan result. Patients who have achieved complete metabolic response (CMR) at the end of treatment (EOT) positron emission tomography (PET)-computer tomography (CT) scan experienced a reduced risk of death compared to patients who have not yet achieved CMR (Hazard ratio [HR] 0.1, 95% Confidence interval [CI]: 0.03– 0.33, P<0.001). Response was simplified to “CMR” versus “no-CMR”.

Table 3.Adverse events.

Furthermore, our patient population included patients with SH lymphoma (24%) based on previous reports demonstrating poor prognosis of these patients following R-CHOP.^5,6,8,27 However, in the revised WHO 2017 classification, SH is not recognized as a separate entity in contrast to DH/TH lymphoma. Recently, Rosenwald et al. demonstrated that the inferior prognosis of MYC rearranged patients is largely observed in patients with DH lymphoma.⁷ However, SH patients still have a worse prognosis compared to patients without a MYC rearrangement, although this is not statistically significant when regarding OS (P=0.077). Our trial was not powered to study the prognostic impact of SH versus DH/TH in the R2CHOP setting.

Finally, we explored the role of iPET-CT scanning as a tool for early identification of refractory MYC+ LBCL. In non-selected cases of DLBCL, CMR on iPET-CT after two to four R-CHOP cycles has a high NPV for 2-year PFS, but the PPV varies widely.²⁸ In our study, the PPV of iPET-CT for achievement of CMR on EOT PET-CT was only 60% and therefore does not support the use of an interim PETCT scan as interpreted with the current standard criteria to identify primary refractory cases treated with R2CHOP.

In this prospective trial for newly diagnosed MYC+ LBCL patients, we found that administering R2CHOP was safe, and yielded comparable CMR and survival rates as in studies applying intensive chemotherapy regimens. Moreover, R2CHOP can be delivered on an outpatient basis in contrast to Burkitt schemes and is easier to deliver than DA-EPOCH-R, since it does not require placement of a central line. These findings offer new prospects for MYC+ LBCL patients and warrant comparison in prospective randomized trials.

Footnotes

• Received September 25, 2019
• Accepted December 6, 2019

Correspondence

MARTINE E.D. CHAMULEAU m.chamuleau@amsterdamumc.nl/m.chamuleau@vumc.nl

Funding

This work was supported by Celgene who funded the research, provided lenalidomide free of charge, approved the original protocol and all amendments as written, but had no role in either the design of the study, the collection, analysis, and interpretation of the data, or the preparation of the manuscript. The Dutch Cancer Society provided funding for the screening program and the clinical trial.

Acknowledgments

The authors wish to acknowledge: the trial managers and central data managers of the study (N. Lamers and A. Elsinghorst, R. Sewsaran and H. Hofwegen), HOVON data Center in Rotterdam; Dr Staudt’s Laboratory at NCI/NIH Bethesda, MD, USA, for online analysis of Lymph2Cx raw data for COO characterization; and the DSMB members M. Jerkeman, K. Dunleavy and E. van Werkhoven for their advice following the interim analysis