Brentuximab vedotin is associated with improved progression-free survival after allogeneic transplantation for Hodgkin lymphoma

Abstract

We previously reported that BV enabled successful reduced-intensity allogeneic hematopoietic cell transplantation (RIC-alloHCT) in patients with relapsed Hodgkin lymphoma, after a median follow-up of 14.4 months. We now provide an updated report on 21 patients who were treated from 2009–2012 with BV prior to RIC-alloHCT with a uniform fludarabine/melphalan conditioning regimen and donor source after a median follow-up of 29.9 months. We have also retrospectively compared the patient characteristics and outcomes of these BV pre-treated patients to 23 patients who received fludarabine/melphalan RIC-alloHCT without prior BV, in the time period before the drug was available (2003–2009, pre-BV era). Patients who were treated with BV prior to RIC-alloHCT had a lower median HCT-specific comorbidity index (HCT-CI) and a reduced number of peri-transplant toxicities. There were also improvements in 2-year PFS (59.3% versus 26.1%) and cumulative incidence of relapse/progression (23.8% versus 56.5%).

Introduction

Brentuximab vedotin (BV) is an antibody-drug conjugate of anti-CD30 antibody and the microtubule-disrupting agent, monomethyl auristatin E [1]. BV is approved for use in Hodgkin lymphoma (HL) patients who have failed autologous hematopoietic cell transplantation (autoHCT). Phase II studies report an overall response rate of ~75% with an acceptable safety profile [2, 3]. We previously published our findings on the use of reduced intensity allogeneic hematopoietic cell transplantation (RIC-alloHCT) in 24 relapsed/refractory patients with HL [4], yielding a 2-year progression-free survival of 27% (95% CI 22–32). We have also reported early data showing that BV salvage prior to RIC-alloHCT results in a 1-year OS of 100% and PFS of 92.3% (95%CI 61.3–98.8) in patients with relapsed Hodgkin lymphoma [5]. We now report on a more homogenous Hodgkin patient population, with extended follow-up for outcomes of RIC-alloHCT following BV salvage. Additionally we have retrospectively compared the outcomes of these patients to a consecutive case-series of BV-naïve patients who underwent RIC alloHCT in the pre-BV era. Our hypothesis is that BV salvage therapy could deliver patients who are better candidates for transplantation, via higher response rates and lower toxicity, thus contributing to improved outcomes after RIC-alloHCT.

Patients and methods

The City of Hope Institutional Review Board approved the retrospective analysis of data from a consecutive case-series of 23 HL patients who underwent RIC-alloHCT with no prior BV exposure (no-BV group) between January 2003 and July 2009 (pre-BV era) and a consecutive case-series of 21 HL patients who received BV prior to RIC-alloHCT (BV group) from July 2009 to December 2012. Sixteen out of the 21 HL patients who received BV were enrolled on prospective clinical trials (4 separate trials). None of the 23 HL patients without prior BV exposure received BV at relapse following RIC-alloHCT. Eligible patients were ≥18 years with histologically-confirmed HL expressing CD30, who had relapsed after prior autoHCT, or were not autoHCT candidates. Patients were excluded if they had received a previous alloHCT. All patients received fludarabine/melphalan (fludarabine 25 mg/m2×5 days followed by melphalan 140 mg/m2×1day) as their transplant conditioning regimen. Only matched related sibling donor (MRD) and matched unrelated donor (MUD) transplants were included; haploidentical and cord blood transplants were excluded. Comorbid conditions at the time of alloHCT were scored using the HCT-CI [6]. The Bearman scale [7] was used to capture toxicities associated with RIC-alloHCT. Baseline patient characteristics for the 44 patients are summarized in Table 1.

Table 1.

Patient, Disease and Treatment Characteristics

Characteristics	BV N = 21	No-BV N = 23
Median age, years (range)	31 (22, 55)	37 (16, 63)
Disease stage at diagnosis (%)
I–II	9 (43)	11 (48)
III–IV	11 (52)	11 (48)
Unknown	1 (5)	1 (4)
Response to induction (%)
Refractory	5 (24)	7 (30)
Relapsed	16 (76)	16 (70)
Median # prior regimens (range)	4 (3, 6)	4 (3, 6)
Prior Regimens (%)
Induction-ABVD	19 (90)	19 (83)
Salvage chemo prior to autoHCT
ICE	17	16
ESHAP	2	10
Others	1	2
AutoHCT	19	19
Salvage chemo after autoHCT
ICE	4	3
ESHAP	0	5
Gemcitabine based	14	17
Bendamustine	3	0
Others	4	9
Radiotherapy	10	17
Consolidation	9	12
Treatment for relapse/refractory	1	5
Median time from diagnosis to HCT, months (range)	60.6 (13.8, 258.3)	36.4 (13.6, 214.7)
Intermittent Therapy between BV and alloHCT	7
Gemcitabine based	5	N/A
ICE	1
Bendamustine	3
XRT	1
Median cycles of BV	7 (2–16)	0
Best Response to BV
CR	6 (29)	NA
PR	14 (67)
SD/PD	1 (5)
Disease Status at end of BV
CR	4 (19)	NA
PR	7 (33)
SD/PD	10 (48)
Disease Status at HCT (%)
CR	6 (29)	1 (4)
PR	9 (42)	9 (39)
SD/PD	6 (29)	13 (57)
Stem Cell Source (%)
Bone Marrow	1 (5)	3 (13)
Peripheral Blood	20 (95)	20 (87)
Median HCT-CI Score (range)	0 (0, 3)	2 (0, 4)
Type of Donor (%)
MRD	10 (48)	12 (52)
MUD	11 (52)	11 (48)
GVHD prophylaxis (%)
Tacrolimus/Sirolimus-based	19 (90)	16 (70)
Cyclosporine A/MMF-based	2 (10)	6 (26)
Tacrolimus/Methotrexate	0 (0)	1 (4)

BV, brentuximab vedotin; ABVD, adriamycin, bleomycin, vinblastine, dacarbazine; ICE, ifosfamide, carboplatin, etoposide; ESHAP, etoposide, cisplatin, Ara-C, methylprednisolone; MOPP, nitrogen mustard, vincristine, procarbazine, predisone; HCT, hematopoietic cell transplantation; ABVD, SD, stable disease; PD, progressive disease; HCT-CI, hematopoietic cell transplantation comorbidity index; MRD, matched related donor; MUD, matched unrelated donor; GVHD, graft versus host disease; MMF, mycophenolate mofetil; XRT, radiation therapy

Post-transplantation evaluation of disease status with imaging studies, bone marrow biopsies and engraftment analysis occurred at 30 days, 100 days, 1 year post-transplantation and yearly thereafter, or as clinically indicated. HL disease response was scored using standard criteria [8]. Overall survival (OS) and progression-free survival (PFS) probabilities were calculated using Kaplan-Meier [9] (differences assessed by Log-rank test) and cumulative incidence of relapse/progression and non-relapse mortality (NRM) were calculated as competing risks [10] (differences assessed using the Gray method).

Results

There were no significant baseline differences between the two groups in terms of age, disease stage at diagnosis, response to induction, number of prior therapies, donor type, stem cell source, and time from diagnosis to RIC-alloHCT. The patients in this study represent a heavily pretreated population in which the majority of patients had undergone high-dose chemotherapy and autoHCT; the median number of prior regimens was 4. The ratio of MRD to MUD in each group was roughly half. Graft versus host disease (GVHD) prophylaxis differed slightly between the two groups due to an institutional shift to tacrolimus/sirolimus in 2005. Although the median number of prior regimens was the same, the no-BV group received more combination chemotherapy and radiotherapy (Table 1). The two groups also differed in terms of disease status at the time of RIC-alloHCT and the HCT-CI score. The median HCT-CI score was significantly better in the BV group (0 versus 2, p<0.01), and patients in this group were also more likely to be in complete remission pre-RIC-alloHCT (28.6% vs. 4.3%, p=0.04).

There were no significant differences between groups in terms of engraftment or acute/chronic GVHD incidence. All patients engrafted, with median time to absolute neutrophil count ≥ 500 cells/µl of 14 days (range: 11–21) in the BV group and 15 days (range: 10–20) in the no-BV group. The median time to platelet count >20,000 cells/µl without transfusion support was 13 days (range: 11–26) in the BV group, and 13 days (range: 8–35) in the no-BV group. All patients in both groups, achieved >99% donor chimerism by day +30. Acute GVHD occurred in 7/21 patients (33.3%) in the BV group and 13/23 patients (56.5%) in the no-BV group. The cumulative incidence of acute GVHD grades II–IV was 23.8% (95%CI: 8.4–43.6) in the BV group versus 47.8% (95%CI: 26.2–66.7) in the no-BV group (p=0.06). For chronic GVHD, the 2-year cumulative incidence was 70.0% (95%CI: 43.3–85.9) for the BV group and 65.2% (95%CI: 40.0–81.9) for the no-BV group (p=0.56).

Stratified outcome curves are shown in Figure 1. For the BV-treated group, with a median follow-up of 29.9 months (range: 12.4–48.5) in surviving patients, the 2-year PFS was 59.3% (95%CI: 33.9–77.7), 2-year OS was 71.1% (95%CI: 43.2–87), 100-day non-relapse mortality (NRM) was 0%, 1-year NRM was 9.5% (95%CI: 2.5–35.6), and the 2-year relapse/progression incidence was 23.8% (95%CI: 11.1–51.2). Surviving patients in the no-BV group had longer follow-up (median 85.3 months, range: 51.5–103.3) due to the time-period difference. In the no-BV group the 2-year PFS was 26.1% (95%CI: 10.6–44.7), 2-year OS was 56.5% (95%CI: 34.3–73.8), 100-day NRM was 4.3% (95%CI: 0.6–29.6), and 1-year NRM was 17.4% (95%CI: 7.1–42.4). The cumulative incidence of relapse/progression at two years was 56.5% (CI 39.5–80.9). The BV group showed improvement in 2-year PFS (59.3% vs. 26.1%, p=0.04) and a reduction in 2-year relapse/progression incidence (23.8% vs. 56.5%, p=0.03) compared to the no-BV group.

Figure 1. Outcomes.

N= 44 patients: Surviving patients in the BV group (N=21) had median follow-up of 29.9 months, No-BV group (n=23) had median follow-up of 85.3 months. (A) Kaplan Meier survival probabilities for progression-free survival (PFS) for BV group (solid line) and No-BV group (dashed line). PFS was defined as time from stem cell infusion to recurrence, progression or death from any cause, whichever occurred first. (B) Overall survival (OS) for BV group (solid line) and No-BV group (dashed line). OS was measured from stem cell infusion to death from any cause. (C) Cumulative incidence of non-relapse mortality (NRM - solid line) for BV group (solid line) and No-BV group (dashed line). NRM was measured from transplant to death from any cause other than disease relapse or progression. (D) The cumulative incidence of relapse/progression for BV group (solid line) and No-BV group (dashed line). The cumulative incidence of relapse/progression was defined as time from stem cell infusion to recurrence or progression. Relapse/progression and NRM were calculated as competing risks.

We evaluated regimen-related toxicities in both groups using the Bearman toxicity scale through day +100. There were no grade III–IV events in the BV group, there were 7 grade III events among 4 patients in the no-BV group: bladder (n=1), gastrointestinal (n=1), pulmonary/renal (n=1), and pulmonary/renal/stomatitis (n=1). In the no-BV group, there were also an increased number of co-morbid conditions: median HCT-CI of 2. In the BV group, fewer patients received multi-agent salvage chemotherapies, possibly due to substitution of BV for a combination regimen. Indeed we found that patients in the BV group had a lower median HCT-CI score (median=0), higher percentages of patients in CR (28.9%), and fewer peri-transplant toxicities (n=0 grade III–IV events). Of the baseline patient, disease and treatment characteristics evaluated by Cox univariate analysis, only prior BV exposure (yes/no) (HR:2.27, 95%CI: 1.04–4.97; p=0.04) and HCT-CI score (HR:1.40, 95%CI: 1.03–1.90; p=0.03 -modeled as a continuous variable) were predictive of PFS.

Discussion

The median overall survival of patients who relapse after autoHCT is only 2.4 years [11]. AlloHCT has been the only option that offers the possibility of long-term remission. Unfortunately, this approach is limited by age, performance status, and comorbidities of patients who have been previously exposed to many rounds of combination chemotherapy. Historical data also show a relatively high relapse rate. Some studies report a low 2-year PFS (23% to 32%) for HL patients undergoing alloHCT [4, 12, 13] while others report 4-year PFS in a similar range (24%-39%) for HL patients undergoing alloHCT [14, 15]. This low PFS could result from lack of disease control prior to alloHCT, as many patients were not in CR or had chemoresistant disease. Brentuximab vedotin, an antibody drug conjugate, was granted accelerated approval by the FDA in 2011 for the treatment of relapsed/refractory HL after failure of auto-HCT [1]. In the pivotal phase II trial that lead to its approval, patients who achieved responses on BV were allowed to come off trial to proceed to alloHCT. We had previously reported on the successful use of BV as a bridge to alloHCT [5]. This report serves to 1) update that experience with longer follow up 2) provide a more homogenous patient population, and 3) compare this group with an historical cohort. Patients in the current report had one additional year of follow-up and thus we are able to provide 2-year PFS data. All the patients in this report received fludarabine and melphalan as conditioning regimen and had matched related or unrelated donor stem cell sources (no cord or haploidentical).

When compared with our own historical cohort (consecutive case series) who received the same conditioning regimens and stem cell sources, we were able to show that BV prior to alloHCT improves HCT-CI, peri-transplant toxicities, and disease status at alloHCT. We believe that these two groups of patients were essentially matched with the exception of BV exposure. This is evident by their stage, age, response to induction, prior auto-HCT, and median number of prior treatments. All the patients had relapsed/refractory disease prior to receiving BV in the BV group or prior to salvage combination chemotherapy in the no-BV group.

In the pivotal phase II trial, BV had a high overall response rate and was well tolerated with a low toxicity profile. The no-BV group had a higher percentage of patients who received ESHAP, which is typically a second salvage chemotherapy regimen. For these patients who were relapsed/refractory to multiple chemotherapy regimens, another round of salvage chemo could cause greater harm than it adds benefit. Indeed, patients in the no-BV arm had worse HCT-CI scores and more grade III–IV Bearman scale peri-transplant toxicities. It is also not surprising that more patients in the BV group were in CR at the time of alloHCT (statistically significant). There were some subtle differences between the groups that were not statistically significant. The no-BV group had received more radiation therapy which could be due to the use of radiation to achieve disease control prior to alloHCT. The BV group also had longer time to alloHCT. This could be explained by the fact that BV can be given for multiple cycles due to its relative low toxicity profile, whereas multi-agent salvage chemotherapies are only given for a maximum of 2–3 cycles prior to alloHCT. Also some patients who had achieved CR/PR waited to undergo alloHCT by choice. Not every patient in the BV group proceeded to alloHCT immediately after BV. Six patients progressed while on BV treatment and some received additional chemotherapy. We did not exclude these patients from the analysis as they were still given BV and achieved response to BV initially.

It is not surprising that we found improved 2-year PFS and cumulative incidence of relapse/progression in the BV group, since multiple prior reports show that improved disease status at transplant and HCT-CI is associated with improved PFS after alloHCT [6, 12, 16–18]. We understand that this study is not a prospective trial and therefore suffers from biases inherent in retrospective analyses. However, this comparison was performed on two consecutive case series of patients from 2003–2008 (pre-BV era) and 2009–2012 (post BV area). Essentially this reflects actual practice patterns occurring at our institution for the past 10 years and shows that alloHCT outcomes have improved in the BV era. OS did not change significantly at the 2-year time point. Of the 7 deaths in the BV group, 2 were due to disease progression, 2 to GVHD, 1 due to infection, and 2 to cardiovascular events (presumed long-term complications of radiation therapy). Of the 17 deaths in the no-BV group, 10 were due to disease progression, 4 due to GVHD, 1 to infection, and 2 due to other causes. If the improvement in relapse rate continues after the 2-year time point, it is very likely we will see an improvement in OS as most of the deaths in the no-BV group were due to disease progression. We will continue to follow these patients and it is possible that we will be able to see improvement in OS at the 3- or 5-year time points.

While our 2-year PFS estimate of 59.3% is less striking than the 1-year PFS of 92.3% seen in our initial report on BV salvage prior to RIC-alloHCT [5], we felt it important to update our initial observations for increased accuracy resulting from longer follow-up, to evaluate outcomes in a more homogenous patient population, and to compare the outcomes of BV patients to a BV-naïve case-series. There have been few other reports on BV prior to alloHCT [19, 20], and our current study represents the largest series of patients with the longest follow-up successfully undergoing RIC-alloHCT post-BV therapy. The patients in our BV and no-BV groups are well matched at diagnosis and at relapse, and diverge in their disease characteristics only after BV treatment. We believe these data demonstrate that BV allows clinicians to improve disease status at transplant, median HCT-CI, and peri-transplant severe toxicities. These in turn are associated with an improvement in 2-year PFS, and cumulative incidence of relapse/progression. For patients with relapsed/refractory HL who are considered candidates for RIC-alloHCT, BV is a reasonable option as a bridge to RIC-alloHCT.