Long-Term Outcomes of Patients with Persistent Indolent B-cell Malignancies Undergoing Nonmyeloablative Allogeneic Transplantation

Abstract

Relapse is least common in patients with indolent B-cell malignancies (iB-NHL) who undergo nonmyeloablative allogeneic transplantation (NMAT) in complete remission (CR). However, for the many patients unable to achieve this state, outcomes are poorly described and methods to improve results are unknown. We sought to describe the long-term follow-up and predictive factors for these poor-risk patients unable to achieve CR prior to NMAT. We identified and evaluated patients with iB-NHL including chronic lymphocytic leukemia (CLL) treated with fludarabine/total body irradiation-based NMAT that had evidence of persistent disease prior to NMAT. From December 1998 to April 2009, 89 patients were identified, most commonly with small lymphocytic lymphoma/CLL (N = 62) and follicular lymphoma (N = 24). Pretransplant anti-CD20 radioimmunotherapy (RIT) using standard yttrium-90-ibritumomab tiuxetan was administered to 18 (20%), who more frequently had chemoresistant disease (81% vs 39%, P = 0.003), disease bulk >5 cm (61% vs 15%, P <0.001), thrombocytopenia <25k/µL (33% vs 7%, P = 0.002), and Hematopoietic Cell Transplant Comorbidity Index scores of ≥3 (72% vs 37%, P = 0.006). After adjusting for these imbalances, RIT-treated patients had improved PFS (HR = 0.4, 95% CI: 0.2–0.9, P = 0.02) and OS (HR = 0.3, 95% CI: 0.1–0.8, P = 0.008) compared to the non-RIT group. The 3-year adjusted estimates of PFS and OS for the RIT and non-RIT groups were 71% and 87%, vs 44% and 59% (respectively). The use of RIT was the only factor independently associated with improved PFS and OS. Rates of non-relapse mortality and graft-versus-host disease (GVHD) were similar between the two groups, though over 70% of patients developed clinically-significant acute or chronic GVHD. In conclusion, despite relatively high rates of GVHD, patients with persistent iB-NHL can derive durable benefit from NMAT.

INTRODUCTION

Nonmyeloablative allogeneic hematopoietic cell transplantation (NMAT) is frequently cited as the only potentially curative intervention for advanced-stage indolent B-cell non-Hodgkin lymphomas (B-NHL) and chronic lymphocytic leukemia (CLL). These results have been demonstrated consistently in a number of series, with long-term relapse-free survival rates ranging from about 40–80% [1–7]. Furthermore, low rates of regimen-related toxicity and non-relapse mortality (NRM) following NMAT make it feasible for older and/or more medically infirm patients. Improved methods of human leukocyte antigen (HLA) typing and medical prophylaxis have also reduced the rates of severe graft-versus-host disease (GVHD), the most common complication of this treatment [8,9].

Despite these successes, NMAT has limitations. Disease relapse remains the major concern following NMAT, particularly for patients with persistent/refractory disease at the time of transplantation with bulky disease sites [1,5]. Thus, it is felt that being in complete remission (CR) at the time of NMAT yields lower rates of relapse [2]. However, for patients with chemotherapy-refractory disease, the use of high-dose chemotherapy-based conditioning may be contemplated to achieve this goal. Unfortunately, identifying a regimen with effective antitumor activity in this setting remains challenging. Moreover, the risk of NRM may be prohibitive for such an approach, which can be estimated with tools such as the Hematopoietic Cell Transplantation Comorbidity Index (HCT-CI) [10].

One strategy to augment NMAT is with radioimmunotherapy (RIT), which delivers a therapeutic radionuclide to an antigen expressed on tumor cells via a monoclonal antibody. Anti-CD20 RIT-augmented NMAT using yttrium-90 (⁹⁰Y)-ibritumomab tiuxetan has been studied by our group and others, based on the hypothesis that chemoresistant tumors may remain radiosensitive, and achieving an early post-transplant CR with this approach could lead to better long-term outcomes [11–16]. These studies suggested the best outcomes were in patients with indolent histologies despite frequent high-risk clinical features, such as chemorefractory and bulky disease. However, to date, there has not been a formal comparative evaluation to determine if such an approach is superior to more standard conditioning regimens.

To address this question and to better understand the utility of NMAT in patients with persistent iB-NHL, we identified a cohort of patients treated at our center meeting eligibility requirements of our prior RIT-NMAT trial but for other reasons did not enroll. These patients could then serve as a fair standard-treatment comparator to the RIT group. Herein, we describe the long-term outcomes for this unique population of patients and evaluate the impact of a RIT-augmented conditioning regimen on results.

METHODS

Study Cohort

We identified patients over the age of 18 years that presented to the University of Washington and Fred Hutchinson Cancer Research Center (FHCRC) between December 1998 and April 2009 for NMAT as treatment for an indolent B-cell malignancy, including CLL, follicular lymphoma (FL), small lymphocytic lymphoma (SLL), marginal zone lymphoma, and hairy cell leukemia. We further selected patients with detectable disease at the time of NMAT, as assessed by imaging studies according to standard international working group definitions [17,18] or via flow cytometric analysis of peripheral blood or bone marrow. We excluded patients without evidence of disease, with large-cell histologic transformation, and those who received NMAT from an HLA-mismatched donor. To control for potential imbalances, patients were only included in this analysis if they met the eligibility criteria for the RIT-based NMAT study, regardless of which treatment they received. For patients not treated on the RIT study, reasons that they did not enroll included referral for NMAT outside of the enrollment period of this study, insurance denial, or preference of the treating physician or patient. The majority of patients included in this analysis were also included in prior publications by our group [4,5,13], but this report includes an additional 2 years of patients treated with standard NMAT and up to 5 additional years of post-NMAT follow-up for those studied previously.

All patients received conditioning with fludarabine 30 mg/m² × 3 days followed by 200 cGy of TBI, as previously described [4]. Patients in the RIT group also received the following, as previously published [13]: On day −21 before transplantation, 250 mg/m² of rituximab was administered before an imaging dose of ¹¹¹In-ibritumomab tiuxetan to ensure expected biodistribution. On day −14, 250 mg/m² of rituximab was administered prior to of 0.4 mCi/kg of ⁹⁰Y-ibritumomab tiuxetan, with a maximum dose of 32 mCi. Post-grafting immunosuppression varied based on the specific protocol on which patients were treated, but all were based on the combination of a calcineurin inhibitor (i.e., cyclosporine or tacrolimus) and mycophenolate mofetil. All patients who were treated on an investigational study signed a consent form approved by the Human Subjects Committee of the University of Washington and/or the Institutional Review Board of the FHCRC in accordance with the Declaration of Helsinki. In addition, separate institutional approval was obtained for this analysis to retrospectively gather data from patient records and databases.

Study Variables

Baseline (i.e., pre-NMAT) demographic information, including age, sex, diagnosis (according to the World Health Organization classification for lymphoid neoplasms [19]), hematopoietic cell donor source, number of prior chemotherapy regimens, prior treatment with rituximab, prior autologous transplantation, and pre-NMAT platelet count was identically collected for the entire group. Patients were considered to have bulky disease if a tumor of >5 cm in largest diameter was noted on pre-NMAT computed tomography (CT) scan. HCT-CI scores were calculated as previously described [10]. Chemoresistant disease was defined as not achieving at least a partial remission to the last systemic therapy administered, according to international working group definitions [17,18]. Follow-up was updated as of September 2013.

Endpoints and Statistical Considerations

Progression-free survival (PFS), overall survival (OS), and NRM were calculated from the date of transplantation, where PFS was defined as a lack of relapse, progression, or death from any cause. Time of onset and severity of acute and chronic GVHD were defined according to standardized criteria [20,21]. Chronic GVHD included all patients treated with systemic immunosuppression for late acute or chronic GVHD by National Institutes of Health criteria [22]. Frequencies of characteristics between groups were compared using a chi-square test, while means were compared using a Wilcoxon rank sum text. Kaplan-Meier curves were used to estimate the probabilities of OS and PFS; cumulative incidence estimates were calculated for acute and chronic GVHD and NRM. The statistical significance of differences in event rates was evaluated with the proportional hazards regression model. All P-values are based on the Wald statistic and are two-sided. To account for imbalanced covariates in subgroups, adjusted estimates of PFS, OS, and cumulative incidences of NRM and acute and chronic GVHD were generated according to previously-published methods [23].

RESULTS

Patient Characteristics

We identified 89 patients from our center that were treated with fludarabine/TBI-based NMAT for indolent B-NHL or CLL with persistent disease at the time of transplantation. Eighteen (20%) of these patients received RIT-augmented NMAT and are hereafter referred to as the “RIT group”; they constituted all of the patients with non-transformed indolent histologies that were treated on our previous phase II study of this approach [13]. The remaining 71 patients (80%) met eligibility criteria for this clinical trial (i.e., histologically-confirmed CD20+ B-NHL or CLL with persistent disease and an HLA-matched peripheral blood stem cell donor) and are hereafter referred to as the “control group”. All but 2 of these 71 patients (97%) were enrolled on an alternative prospective clinical trial at our center.

The baseline characteristics of these 89 patients with persistent pre-transplant disease are summarized and compared in Table 1. The majority of patients in this cohort were male, over 50 years of age, and had CLL/SLL. Beyond the presence of persistent disease, these patients generally had several other high-risk features. They were relatively heavily pre-treated, with a median of 4 prior chemotherapy regimens prior to NMAT, with a large majority having received rituximab as part of this prior treatment. Nearly half had chemoresistant disease, and one in four had bulky sites of disease. Among the 47 patients with CLL/SLL from whom cytogenetic data were available, 14 (30%) harbored high-risk cytogenetic abnormalities: 8 (17%) had a deletion at 11q, 5 (11%) had a deletion at 17p, and 1 (2%) had deletions at both loci.

Table 1.

Baseline characteristics of patents who underwent nonmyeloablative allogeneic transplantation for persistent indolent B-cell malignancies.

Patient characteristic	Total (N = 89)		RIT (N = 18)		Control (N = 71)		Pa
	N	%	N	%	N	%
Gender							0.69
Male	66	74	14	78	52	73
Female	23	26	4	22	19	27
Age (years)							0.95
Median	56		58		55
Range	30–68		30–68		35–67
Histology							0.08
CLL/SLL	62	70	10	56	52	73
FL	24	27	6	33	18	25
Otherb	3	3	2	11	1	1
Number of prior therapies							0.58
Median	4		4		4
Range	1–12		3–12		1–11
Prior rituximab	78	88	18	100	60	85	0.07
Prior autologous transplant	16	18	4	22	12	17	0.60
Chemoresistantc	38	48	13	81	25	39	0.003
Bulky disease (>5 cm)	22	25	11	61	11	15	< 0.001
HCT-CI score ≥3	39	44	13	72	26	37	0.006
Unrelated donor	42	47	12	67	30	42	0.06
Pre-NMAT platelet count <25k/µL	11	12	6	33	5	7	0.002

RIT, radioimmunotherapy; CLL/SLL, chronic lymphocytic leukemia/small lymphocytic lymphoma; FL, follicular lymphoma; HCT-CI, hematopoietic cell transplantation comorbidity index; NMAT, nonmyeloablative allogeneic transplantation.

^aP-values are from comparisons between the RIT and control groups.

^bRIT group: marginal zone lymphoma (1) and hairy cell leukemia (1); control group: marginal zone lymphoma (1)

^cResponse to the last systemic therapy administered before NMAT was not available from 9 patients (2 from the RIT group, and 7 from the control group).

Outcomes of iB-NHL with Pretransplant Persistent Disease

PFS and OS for these 89 patients, as well as PFS from selected subgroups, are depicted in Figure 1. The median PFS and OS for the group as a whole were 23.4 months and 76.5 months, respectively, while the 3-year PFS and OS were 46% and 63%, respectively, with a median follow-up of 6.8 years. Table 2 shows unadjusted comparisons between the various subgroups investigated. Among these, the largest apparent difference was observed between the patients with or without bulky disease >5 cm (median PFS: 14.8 months vs 30.8 months, respectively; hazard ratio [HR] = 1.5, 95% confidence interval [CI]: 0.9–2.7). However, this difference was not statistically significant (P = 0.15).

Figure 1. Survival of 89 patients who underwent nonmyeloablative allogeneic transplantation for persistent indolent B-cell malignancies.

Kaplan-Meier curves depict (A) progression-free and overall survival of all patients, followed by comparisons based on (B) disease bulk, (C) histology, (D) hematopoietic cell donor source, and (E) cytogenetic risk. Panel E depicts only those patients with chronic lymphocytic leukemia/small lymphocytic lymphoma.

Table 2.

Univariate analysis of progression-free survival between subgroups of patients with persistent indolent B-cell malignancies undergoing nonmyeloablative allogeneic transplantation.

	N	Median PFS (months)	3-Year PFS	HR (95% CI)	P
No RIT	71	19.0	44%	1.0
RIT	18	53.0	56%	0.7 (0.4–1.5)	0.36
Disease Bulk ≤5 cm	67	30.8	49%	1.0
Disease Bulk >5 cm	22	14.8	36%	1.5 (0.9–2.7)	0.15
Other Histologies	27	18.1	37%	1.0
CLL/SLL	62	23.8	50%	0.8 (0.5–1.5)	0.56
Related Donor	47	37.6	53%	1.0
Unrelated Donor	42	19.0	38%	1.1 (0.7–1.8)	0.76
Non-High-Risk Cytogeneticsa	33	53.0	61%	1.0
High-Risk Cytogeneticsa	14	19.9	43%	1.2 (0.5–2.8)	0.63

PFS, progression-free survival; HR, hazard ratio; CI, confidence interval; RIT, radioimmunotherapy; CLL/SLL, chronic lymphocytic leukemia/small lymphocytic lymphoma; MRD, matched related donor;.

^a“High-risk cytogenetics” were those patients with CLL/SLL that harbored deletion 11q and/or deletion 17p; all others with CLL/SLL from whom cytogenetic data were available were considered “non-high-risk”.

Comparison of RIT vs Control Groups

Patients in the RIT group were more likely to have high-risk features than those in the control group (Table 1), with significantly more chemoresistant (81% vs 39%, P = 0.003) and bulky disease (61% vs 15%, P <0.001), high-risk HCT-CI scores (72% vs 37%, P = 0.006), and low pre-NMAT platelet count (33% vs 7%, P = 0.002). Multivariate analysis, adjusting for factors with greatest imbalance between groups (Table 3), demonstrated significant improvement in both PFS (HR = 0.4, 95% CI: 0.2–0.9, P = 0.02) and OS (HR = 0.3, 95% CI: 0.1–0.8, P = 0.008) in the RIT group. Chemoresistance was highly correlated with bulky disease and not included in the multivariate analysis. Model-based adjusted estimates of PFS and OS for the RIT group are shown in Fig. 2A and Fig. 2B, respectively, along with unadjusted Kaplan-Meier estimates. The adjusted estimates show the predicted PFS and OS for a group of RIT patients with the same covariate characteristics as the control group. The 3-year adjusted estimates of PFS and OS for the RIT group were 71% and 87%, compared to 44% and 59% for the control group, respectively.

Table 3.

Multivariable analysis of factors imbalanced between the radioimmunotherapy and control groups.

	Overall Survival (49 events)		Progression-Free Survival (59 events)
	HR (95% CI)	P	HR (95% CI)	P
RIT	0.3 (0.1–0.8)	0.008	0.4 (0.2–0.9)	0.02
Unrelated donor	1.1 (0.6–1.9)	0.76	1.2 (0.7–2.0)	0.54
Bulky disease (>5 cm)	2.4 (1.1–4.8)	0.03	1.9 (1.0–3.7)	0.06
Pre-NMAT platelet count <25k/µL	1.4 (0.6–3.2)	0.41	2.1 (1.0–4.4)	0.06
HCT-CI score ≥3	1.5 (0.8–2.7)	0.22	1.4 (0.8–2.4)	0.30

HR, hazard ratio; CI, confidence interval; RIT, radioimmunotherapy; NMAT, nonmyeloablative allogeneic transplantation; HCT-CI, hematopoietic cell transplantation comorbidity index.

Figure 2. Outcomes are significantly improved with the addition of anti-CD20 radioimmunotherapy (RIT) to nonmyeloablative allogeneic transplantation for persistent indolent B-cell malignancies, after adjusting for imbalanced covariates.

Kaplan-Meier curves depict (A) progression-free survival (PFS) and (B) overall survival (OS) in the control and RIT groups. The solid lines depict the observed outcomes for the control and RIT groups. The dashed lines reflect the adjusted estimates, showing the hypothetical PFS and OS for a group of RIT patients with the same covariate characteristics as the control group.

NRM and GVHD

We found no significant difference in rates of NRM between the RIT and control groups (HR = 0.5, 95% CI: 0.2–1.8, P = 0.32; Fig. 3A). In addition, no difference in the cumulative incidence of clinically-significant (i.e., grade II–IV) acute GVHD (HR = 1.0, 95% CI: 0.6–1.9, P = 0.88) or chronic GVHD (HR = 1.1, 95% CI: 0.6–2.0, P = 0.76) between these two groups (Fig. 3B and 3C, respectively) was noted. Among patients that were alive and disease-free 1 year after NMAT, 14 of 15 patients (93%) in the RIT group and 37 of 44 (84%) in the Control group developed chronic GVHD. Similar to above, we compared the rates of these events following adjustment for potentially confounding factors between the RIT and control groups: for NRM, we adjusted for age ≥50, HCT-CI ≥3, number of prior therapies ≥5, and donor relation; for acute and chronic GVHD, we adjusted for age and donor relation. Following this, the adjusted HR's were very similar to the unadjusted HR’s for NRM (HR = 0.4, 95% CI: 0.2–1.5, P = 0.18), grade II–IV acute GVHD (HR = 0.9, 95% CI: 0.5–1.7, P = 0.73), and chronic GVHD (HR = 1.1, 95% CI: 0.6–2.0, P = 0.82). While not specifically assessed here, details regarding engraftment and toxicity among these patients have been reportedly previously and were similarly comparable between RIT and control groups [4,5,13].

Figure 3. The addition of anti-CD20 radioimmunotherapy does not significantly increase the rate of serious toxicity following nonmyeloablative allogeneic transplantation.

Kaplan-Meier curves depict the cumulative incidence rates of (A) non-relapse mortality, (B) grades II–IV acute graft-vs-host disease (GVHD), and (C) chronic GVHD in the control and RIT groups.

DISCUSSION

We describe the long-term outcomes (6.8 years median follow-up) after NMAT for patients with persistent indolent B-NHL or CLL and demonstrate two key findings: such an approach can yield long-term remissions in a large proportion of these seemingly high-risk patients, and improved results may be obtained after addition of standard doses of ⁹⁰Y-ibritumomab tiuxetan. Even though attempts are typically made to induce a CR with salvage therapies prior to NMAT, many patients cannot achieve this goal, supported by the fact that the largest cohorts described to date include a substantial proportion of patients not in CR [1–7]. Further, a subset of these patients with refractory disease are likely being denied transplant for this reason. Our data, which to our knowledge are the first to specifically examine this relatively common scenario, assist in providing an estimate of outcomes for such patients.

In our analyses, we attempted to identify the key factors that may predict outcome for this inherently high-risk population. Interestingly, the only factor that was associated with significantly better PFS and OS in a multivariate analysis was the addition of RIT (Table 3). From our results, we have identified one of the first disease-specific manipulations to a NMAT platform that can significantly improve outcome. This conclusion was suggested previously in a similar study [16]: In patients with FL undergoing NMAT, these authors concluded that the addition of anti-CD20 RIT seemed to overcome the negative prognostic impact of chemorefractory disease. Unlike our findings here, no formal statistical analyses were performed to objectively demonstrate benefit from such an approach.

The potential benefit of RIT is particularly notable in our study given the preponderance of patients with CLL/SLL and the relatively low level of CD20 expression typical of this disease. It has long been known that even low doses of radiation can be exquisitely effective in CLL/SLL [24]. However, evaluation of RIT in patients with active disease outside the transplant setting has been limited due to the risk for prolonged hematopoietic toxicity in the setting of extensive marrow involvement or baseline cytopenias. Our study would suggest that the pretransplant use of RIT may represent the optimal use of this modality in CLL/SLL with marrow involvement, particularly since the degree of marrow involvement was not factored into eligibility for our RIT-based approach [13].

We concede that the limited sample size and non-randomized nature of our study does not definitively establish that anti-CD20 RIT should be used prior to NMAT in all patients. It is notable that in over a decade of NMAT at our center, we were only able to identify 89 such patients, potentially suggesting that many patients unable to achieve CR are not referred for transplant due to perceived poor outcomes. This premise is particularly important in the current era of novel B-cell receptor signaling pathway targeted agents such as ibrutinib and idelalisib, which typically induce high overall response and disease control rates but rarely yield CRs and are not curative [25–28]. Our data would suggest that these drugs could still be used as a successful bridge to transplant while donors are being identified despite resultant persistent disease.

In conclusion our data indicate that despite the negative prognostic impact of persistent disease at the time of NMAT, a lack of CR should not be a contraindication to transplant, with 46% of our patients with iB-NHL or CLL/SLL being alive and progression-free at 3 years. The addition of ⁹⁰Y-ibritumomab tiuxetan appears to further improve both PFS and OS without additional appreciable toxicity. These findings are particularly relevant in the current era of novel therapeutics and set the stage for a prospective randomized trial of standard vs RIT-augmented NMAT in non-remission patients with indolent B-cell malignancies.