Ibrutinib monotherapy in relapsed or refractory, transformed diffuse large B-cell lymphoma

Abstract

Background:

Histological transformation to diffuse large B-cell lymphoma (tDLBCL) occurs in a significant proportion of indolent lymphomas. However, few studies of novel agents inform its management, particularly when relapsed after or refractory (R/R) to prior treatment.

Patients and Methods:

We prospectively evaluated ibrutinib monotherapy in pathologically documented R/R tDLBCL in a single-arm study. The primary endpoint was overall response rate (ORR).

Results:

Twenty patients who had received a median 4 (range 2–9) prior lines of therapy overall [median 2.5 (range 1–9) for tDLBCL] were treated. ORR was 35%, including complete responses in 15%. Median progression-free survival and overall survival were 4.1 months (95% CI, 2.4 – 6.2) and 22.4 months (95% CI 7.5 – NR), respectively. Disease control > 2 months was seen in 75% and > 1 year in 15%. Response was associated with either low tumor bulk or low metabolic tumor volume (p = 0.05) but not with antecedent lymphoma histology (p = 1.0). Treatment-related adverse events were consistent with prior studies of ibrutinib.

Conclusions:

Ibrutinib showed low toxicity and meaningful efficacy in R/R tDLBCL, including short term disease control in most cases. Results demonstrate the potential utility of ibrutinib in this challenging clinical setting, including as a potential bridge to more definitive treatments.

MicroAbstract

Here we prospectively investigate ibrutinib monotherapy in 20 patients with relapsed or refractory diffuse large B cell lymphoma histologically transformed from indolent lymphoma (R/R tDLBCL). Objective responses were seen in 35%, with median progression free survival of 4.1 months (95% CI, 2.4 – 6.2). Ibrutinib shows favorable tolerability and promising activity in this setting that otherwise has limited established treatment options.

Introduction

Histologic transformation of indolent B-cell non-Hodgkin lymphoproliferative neoplasms (iB-NHL) to diffuse large B cell lymphoma (tDLBCL) occurs relatively frequently, especially in cases of early progression, and is associated with increased morbidity and mortality (1, 2). tDLBCL relapsed after or refractory to treatment (R/R tDLBCL) portends an especially poor prognosis, likely due to acquired chemoresistance and host features including immunoparesis and poor tolerance of therapy (3, 4). R/R tDLBCL thus constitutes an unmet medical need, and effective, novel agents with reduced toxicity are required. Despite this, few prospective studies have addressed this clinical scenario as industry-initiated protocols for R/R iB-NHL or R/R de novo DLBCL often exclude R/R tDLBCL (5, 6).

Ibrutinib is an orally administered covalent inhibitor of Bruton’s tyrosine kinase (BTK). While it has shown efficacy in multiple types of iB-NHL and subsets of de novo DLBCL, recent studies have shown it to be relatively less active in follicular lymphoma (FL) and DLBCL of germinal center B-cell (GCB) cell of origin (COO) (7–12). Here we evaluated the efficacy and safety of single-agent ibrutinib in patients with R/R tDLBCL.

Patients and Methods

This investigator-initiated, open-label, single-center, single-arm study was conducted in accordance with the general ethical principles in the Declaration of Helsinki and approved by the institutional review board at the Fred Hutchinson Cancer Research Center. All patients provided written informed consent. Patients were accrued at the Seattle Cancer Care Alliance / University of Washington. Janssen Research and Development (Titusville, New Jersey) provided study drug and funding. The trial was registered at www.clinicaltrials.gov (NCT02207062). Eligible patients were age 18 years or older with histologically confirmed tDLBCL measurable by standard criteria and R/R to at least 1 line of therapy given for the histologically documented transformed lymphoma (13). Patients were required to have an Eastern Cooperative Oncology Group performance status of ≤ 2 and acceptable organ function, including creatinine clearance > 25 mL/min, total bilirubin ≤ 1.5 times upper limit of normal, and absolute neutrophil count and platelet count of ≥ 1000 / mm³ and 50,000 / mm³, respectively (≥ 750 / mm³ and 30,000 / mm³, respectively, in the setting of marrow involvement by disease). Prior ibrutinib was not permitted.

Patients self-administered ibrutinib 560 mg daily on consecutive 21-day cycles until disease progression, unacceptable toxicity, or decision of the patient or treating physician to halt therapy (Figure 1). The primary endpoint of the study was investigator-assessed overall response rate [(ORR) = combined rates of partial response (PR) and complete response (CR)] (13). For this pilot study the predefined level for potential efficacy was ORR ≥ 30%. Secondary endpoints included standard assessments of progression-free survival (PFS), overall survival (OS), duration of response (DOR) (13), and safety. We also defined, as a post-hoc exploratory endpoint, the term bridging-benefit rate (BBR) as no progression of disease (PD) or unmanageable toxicity [severe adverse event (SAE) or toxicity requiring dose-reduction or hold] for 2 months after ibrutinib initiation for potential to support successful transition to engineered cellular therapy. Burden of disease at screening was defined and assessed post-hoc. As no common standards exist for cutoffs of high tumor burden by maximum dimension of largest tumor (MDT) or by metabolic tumor volume (MTV), we dichotomized at ≥ 4 cm and ≥ 100 cm³ for their approximately median values in this cohort (14). For MTV calculation, volumes of all lymphomatous lesions with SUV greater than liver were summed using PETVCAR software (GE Healthcare). The Kaplan-Meier method was used to estimate time-to-event in months with corresponding 95% confidence interval (CI). Clinical variables known to be correlated with outcomes of aggressive lymphoma were evaluated for association with efficacy in exploratory analyses. Statistical software R version 3.4.1 was used for analyses.

Figure 1.

Study Schema. Toxicity evaluations (small arrows) were performed each of the first 9 cycles and every other cycle thereafter. Treatment was held for toxicity (per Common Terminology Criteria for Adverse Events version 4.0) until improvement to grade ≤ 1 and then restarted at the original dose level. For subsequent episodes of the same toxicity, ibrutinib was held until improvement to grade ≤ 1 and then restarted at a reduced dose; ibrutinib was discontinued permanently if held for more than 21 consecutive days. Evaluation of disease (large arrows) was performed at screening and on completion of every 3 cycles of therapy during the first 9 cycles and on completion of every 6 cycles of therapy thereafter, or as indicated to evaluate clinical suspicion of disease progression. Abbreviations: LTFU = long-term follow-up; PD = disease progression.

Results

Patients

From October 2014 to March 2019 twenty patients with R/R tDLBCL were enrolled (Table 1). The median age was 68 and the antecedent iB-NHL was FL in 70%. There were no cases of high-grade B-cell lymphoma. Forty-five percent had an international prognostic index (IPI) of ≥ 3. Patients had received extensive prior treatment, including a median of 4 (range 2 – 9) regimens administered for their iB-NHL and a median of 2.5 (range 1 – 9) regimens specifically for tDLBCL. Therapies included anthracycline-based chemotherapy in 95%, platinum-based chemotherapy in 75%, and high dose therapy and autologous stem cell transplantation (SCT) in 45%; none had received prior BTK inhibitor. COO assessment by immunohistochemistry, as is established for de novo DLBCL, was performed for tDLBCL: germinal center B cell-like (GCB) predominated, identified in in 70% overall and 93% of cases with antecedent FL (15).

Table 1.

Patient demographics and baseline characteristics of all treated patients (N = 20).

Characteristic	Patients, N	%
Median age, years (range)	68 (range, 35 – 89)
Sex
Male	9	45
Female	11	55
Antecedent lymphoma histology
FL	14	70
WM/LPL	2	10
CLL/SLL	2	10
iB-NHL, NOS	2	10
COO of tDLBCL
GCB type	14	70
non-GCB type	3	15
Unclassifiable	3	15
IPI at enrollment
0–1	5	25
2	6	30
3	6	30
4	3	15
Maximum tumor dimension (cm) at enrollment
< 4	11	55
≥ 4	9	45
Metabolic tumor volume (cm3) at enrollment
< 100	10	56
≥ 100	8	44
N/A	2
Median number prior therapies (range)
Overall	4 (range, 2–9)
Administered for tDLCBL	2.5 (range,1–9)
Prior rituximab	20	100
Prior anthracycline	19	95
Prior platinum	15	75
Prior autologous stem cell transplant	9	45
Prior CAR-T	2	10
Prior allogeneic stem cell transplant	0	0
Chemotherapy refractory	9	45
Rituximab refractory	10	50

FL = follicular lymphoma; WM/LPL = Waldenstrom’s macroglobulinemia / lymphoplasmacytic lymphoma; CLL = chronic lymphocytic leukemia / small lymphocytic lymphoma; iB-NHL NOS = indolent B-cell non-Hodgkin lymphoma not otherwise specified; COO = cell of origin; tDLBCL (transformed diffuse large B-cell lymphoma); GCB = germinal center B cell-like type; IPI = international prognostic index; N/A = not able to be assessed; CAR-T = chimeric antigen receptor.

Safety

Treatment-related adverse events (TRAE) were experienced by 95% (Table 2). The most common TRAEs were fatigue and bruising, each occurring in 55% and limited to grade 1–2 in all cases. Grade 3–4 TRAEs were observed in 45% and included atrial fibrillation or flutter in 10%, and nausea, pneumonia, esophagitis/mucositis, ventricular arrythmia, sepsis, neutropenia, and lymphopenia in 5% each. One patient died (grade 5 TRAE) on treatment. The death was sudden and without known cause; the patient had no prior history of cardiac disease or other remarkable medical history. An additional 2 patients died from disease progression within 30 days of the last dose. One patient discontinued therapy for TRAE grade 3 esophagitis/mucositis. Dose holds unrelated to elective minor procedures were required on 6 occasions in 25% patients for a median of 17.5 days (range 8 – 21); no dose reductions were made. Indications for dose holds included esophagitis/mucositis, neutropenia, transaminitis, ventricular arrythmia, and pneumonia.

Table 2.

Incidence of most common (≥ 10%) grade 1 and 2 treatment related adverse events and all grade ≥ 3 treatment related adverse events (N = 20).

Adverse Event, No. (%)	Grades 1 and 2	Grades ≥ 3
Bruising	11 (55)	0 (0)
Fatigue	11 (55)	0 (0)
Diarrhea / loose stools	7 (35)	0 (0)
Rash	6 (30)	0 (0)
Extremity swelling	5 (25)	0 (0)
Epistaxis	4 (20)	0 (0)
Oral mucositis / ulceration	4 (20)	0 (0)
Atrial fibrillation/flutter	1 (5)	2 (10)
Nausea	2 (10)	1 (5)
Muscle cramps	3 (15)	0 (0)
Pneumonia	1 (5)	1 (5)
Pruritis	2 (10)	0 (0)
Headache	2 (10)	0 (0)
Anorexia	2 (10)	0 (0)
Esophagitis/mucositis	0 (0)	1 (5)
Sepsis	0 (0)	1 (5)
Lymphopenia	0 (0)	1 (5)
Neutropenia	0 (0)	1 (5)
Ventricular arrythmia	0 (0)	1 (5)
Death, unknown cause	0 (0)	1 (5)

CTC, common terminology criteria for adverse events, version 4.03.

Efficacy

Median time on therapy was 3.1 months (range 0.6 – 48.0) and the ORR was 35% (Figure 2). Complete response occurred in 15% and extended disease control (> 12 months) was also seen in 15%. Median OS was 22.4 months (95% CI, 7.5 – NR), median PFS was 4.1 months (2.4 – 6.2), and median DOR was 2.3 months (2.1 – NR) (Figure 3). With arrival of engineered cellular therapy including chimeric antigen receptor T-cell therapy (CAR-T) as a potentially curative approach for R/R tDLBCL, there is often a need for a well-tolerated bridging treatment to control disease during transition to CAR-T centers and through cell processing (16, 17). Here, the BBR was 75% overall and 60% even in high burden of disease by MDT or MTV. Of the 17 patients not on ibrutinib at last follow-up, 76% went on to further treatment, including CAR-T in 23% and allogeneic SCT in 6%. In patients that received subsequent cellular therapy no unexpected toxicity signals emerged. Of these patients, 3 remained in remission after CAR-T (median follow-up 11 months, range 2 to 47 months), which was the first post-ibrutinib systemic therapy in each; 1 other patient died due to lymphoma progression following post-ibrutinib treatment with allo-SCT and subsequent CAR-T.

Figure 2.

Outcomes of patients evaluable for response (N = 18). A) best response as measured by % change in sum of product of diameters (SPD) of index lesions. Color corresponds to histology of antecedent indolent B-cell non-Hodgkin lymphoproliferative neoplasm (iB-NHL). Bars are labeled with cell-of-origin of transformed diffuse large B-cell lymphoma by immunochemistry using Hans algorithm. B) Progression free survival (PFS) according to burden of disease. Two-month point indicated as measure of potential for bridging benefit. Duration of PFS at last follow-up noted in cases of > 12 months. Abbreviations: FL = follicular lymphoma; WM/LPL = Waldenstrom’s macroglobulinemia / lymphoplasmacytic lymphoma; CLL/SLL = chronic lymphocytic leukemia / small lymphocytic lymphoma; NOS = not otherwise specified; GCB = germinal center B cell-like; PFS = progression free survival.

Figure 3.

Kaplan Meier plots of A) duration of response, B) progression-free survival, and C) overall survival for cohort (N = 20) treated on study.

Excluding the 2 patients not assessable for response (1 who stopped treatment due to intractable esophagitis/mucositis and 1 who died without progression), univariate analyses showed low burden of disease by MTD or MTV was associated with a higher ORR (66% versus 11%, p = 0.05) and longer median PFS (6.2 versus 2.4 months, p < 0.01) (Table 3). Prior receipt of autologous SCT, possibly reflecting selection of patients who are more fit and/or have less treatment-resistant disease, was also associated with improved ORR (75% versus 10%, p = 0.01) and median PFS (6.2 versus 2.3 months, p < 0.01). Though the sample size does not support robust multivariate analysis, most patients (75%) who had a prior auto SCT also had low burden of disease at enrollment. Efficacy was not associated with antecedent lymphoma histology, COO of tDLBCL, or IPI at enrollment.

Table 3.

Univariate analysis of factors associated with overall response rate in evaluable patients (N = 18).

Characteristic	Response N (%)	P-value
IPI		0.64
1–2	5 of 11 (45)
3–4	2 of 7 (29)
Histology antecedent lymphoma		1.00
FL	5 of 14 (36)
Non-FL	2 of 4 (50)
Tumor burden1		0.05
High	1 of 9 (11%)
Low	6 of 9 (66%)
Cell of Origin		0.62
GCB	6 of 14 (43)
Non-GCB	1 of 4 (25)
Prior autologous SCT		0.01
Yes	6 of 8 (75)
No	1 of 10 (10)

¹High tumor burden defined by maximum dimension of largest tumor ≥ 4 cm or metabolic tumor volume ≥ 100 cm³. IPI = international prognostic index; FL = follicular lymphoma; GCB = germinal center B-cell; SCT = stem cell transplantation.

Discussion

Here we show that ibrutinib monotherapy achieved a response in 35% of patients with R/R tDLBCL. This included a 15% CR rate and a 15% rate of PFS exceeding 1 year. The ORR in this pilot study exceeded the predefined measure of success of ORR ≥ 30%, providing rationale for continued evaluation of ibrutinib for R/R tDLBCL.

Ibrutinib showed substantial activity despite extensive prior treatments and high-risk disease at enrollment. Post-hoc evaluation of variables for association with ibrutinib efficacy revealed the potential significance of clinical factors including disease burden by MTD or MTV and prior receipt of autologous SCT. Disease burden by MTV is broadly gaining traction as an independent predictor of outcome of de novo DLBCL and FL (18–20). While its role in risk stratification of R/R tDLBCL has yet to be established, these data indicate its possible utility.

Interestingly, ibrutinib efficacy was not correlated with histology of the antecedent iB-NHL. While our sample size is small, this finding contrasts with the wide range of observed ibrutinib response rates in relapsed WM/LPL (91%), CLL/SLL (71%), and FL (21%) and argues against dismissing the potential benefit of ibrutinib in cases of transformed FL (7, 10, 11). Additionally, while molecular characterization was limited to clinically standard analyses, we found no association between response rates and the COO assigned to tDLBCL, with 6/14 (43%) ORR in cases of GCB tDLBCL. In de novo DLBCL, ibrutinib is thought to be preferentially effective in activated B cell-like DLBCL based on published ORR of 37% versus only 5% in GCB DLBCL, and encouraging results in combination therapies for non-GCB DLBCL (12, 21, 22). Our data are consistent with a retrospective, real-world study by Winter et al. showing an ORR of 28% of R/R DLBCL to ibrutinib monotherapy, without a difference between GCB and non-GCB subtypes (23). The lack of association with COO might be due to the fact that transformation of iB-NHL can include dysregulation of B-cell transcription factors and activation of nuclear factor-κB, possibly yielding susceptibility to BTK inhibition (24, 25).

The relatively small sample size, attributable in part to requirement for pathologically documented transformed disease and subsequent failure of at least one regimen after transformation, may limit the precision of the response estimate. Nevertheless, this requirement ensured that ibrutinib was not treating the underlying indolent disease and likely selected for a higher risk group by excluding patients who had not received treatment for documented DLBCL. These data do suggest that ibrutinib is active in this setting, though in practice may be limited to DLBCL arising from diseases such as FL where prior use of BTK inhibitor is unlikely.

The introduction of modalities including CAR-T cell therapy is improving the treatment landscape for R/R DLBCL including tDLBCL (16, 17, 26). Still, well-tolerated, all-oral options offering disease control are needed after failure or during the preparation of more intensive treatments, the delay of which can hinder their successful administration (27). The 75% BBR shown here, including 60% BBR in cases of high tumor burden, supports the consideration and future research of ibrutinib as a bridge to therapies for R/R tDLBC; emerging data even suggest concurrent ibrutinib with CD19 CAR-T for CLL may enhance CAR-T viability and function and attenuate some immune-mediated toxicities (28, 29).

Conclusion

In conclusion, ibrutinib was well-tolerated and resulted in responses in over one-third of patients with R/R tDLBCL. Investigation of the underrepresented group of R/R tDLBCL in prospective studies can be accomplished in trials such as this, and data from FL or de novo DLBCL should not necessarily be directly extrapolated to tDLBCL. These results also support further investigation of ibrutinib in combination regimens and prior to CAR-T therapy in R/R tDLBCL.

Clinical Practice Points.

• Histologic transformation of indolent B-NHL to DLBCL occurs in a substantial number of patients and its relapse after treatment is associated with a poor prognosis.

• Few prospective studies have investigated R/R tDLBCL due to its frequent exclusion from studies evaluating indolent B-NHL or de novo DLBCL.

• In this investigator-initiated study of 20 patients with R/R tDLBCL, ibrutinib monotherapy was well tolerated and resulted in rates of overall and complete responses of 35% and 15%, respectively.

• Disease control extended > 1 year in 15%, indicating long-term activity for this strategy in a subset of cases.

• Two-month disease control, a period that may facilitate manufacture and delivery of CAR-T therapy, was observed in a majority (75%) of patients.