A phase II multicenter trial of hyperCVAD MTX/Ara-C and rituximab in patients with previously untreated mantle cell lymphoma; SWOG 0213

Abstract

Background

Rituximab-hyper-CVAD alternating with rituximab-high-dose methotrexate and cytarabine is a commonly utilized regimen in the United States for mantle cell lymphoma (MCL) based on phase II single institutional data. To confirm the clinical efficacy of this regimen and determine its feasibility in a multicenter study that includes both academic and community-based practices, a phase II study of this regimen was conducted by SWOG.

Patients and methods

Forty-nine patients with advanced stage, previously untreated MCL were eligible. The median age was 57.4 years (35–69.8 years).

Results

Nineteen patients (39%) did not complete the full scheduled course of treatment due to toxicity. There was one treatment-related death and two cases of secondary myelodysplastic syndrome (MDS). There were 10 episodes of grade 3 febrile neutropenia, 19 episodes of grade 3 and 1 episode of grade 4 infection. With a median follow-up of 4.8 years, the median progression-free survival was 4.8 years (5.5 years for those ≤65 years) and the median overall survival (OS) was 6.8 years.

Conclusions

Although this regimen is toxic, it is active for patients ≤65 years of age and can be given both at academic centers and in experienced community centers.

introduction

Although recent studies have shown improved survival for patients suffering from mantle cell lymphoma (MCL), there is still a great debate as to how to optimally treat these patients. What is not debatable, however, is that standard dose of CHOP [cyclophosphamide (Bristol Meyers Squibb, Princeton, NJ), adriamycin (Pharmacia Inc., Kalamazzo, MU), vincristine (Eli Lilly, Indianapolis, Indiana) and prednisone] and rituximab (anti-CD20 monoclonal antibody) alone is inadequate treatment resulting in median progression-free survival (PFS) ranging only from 16 to 20 months [1, 2]. Therefore, there has been great interest in developing more effective treatment approaches for patients with this disease.

One approach, popularized by investigators at the MD Anderson Cancer Center (MDA), employs a hyper-fractionated schedule of the component drugs of CHOP (hyper-CVAD) alternating with the combination of high-dose methotrexate and cytarabine (MTX/AraC). Such a regimen resulted in an overall response rate of 97% with an 87% complete response (CR) rate in a study of 97 patients treated at MDA. With a median follow-up of 8 years, the median time-to-treatment failure (TTTF) was 4.6 years with the median OS not having been reached [3]. Although this clearly is an effective regimen, it is associated with significant toxicity as the treatment-related mortality in this study was 8% [3].

Despite the fact that the feasibility and efficacy of this regimen have not been shown in a US multi-institutional cooperative group setting that includes both community-based and academic centers, many US physicians have nonetheless adopted this regimen as their initial treatment approach for patients with newly diagnosed MCL based solely on the MDA data. To provide such critical information, we instituted SWOG 0213 in 2002, a phase II trial of R-HyperCVAD/MTX/AraC in patients with previously untreated MCL.

patients and methods

This study started in October 2002 and was completed in September 2006. The primary objective was to estimate PFS in patients with previously untreated MCL treated with R-HyperCVAD/MTX/AraC. Patients must have had newly diagnosed, previously untreated, biopsy proven Stage III/IV or bulky Stage II MCL. Patients must have had adequate renal, hepatic, hematologic and cardiac functions, have an ECOG performance status (PS) of 0–2 and be between 18 and 70 years of age. Patients with known acquired immunodeficiency syndrome (AIDS) syndrome or human immunodeficiency virus-associated complex were not eligible.

Representative hematoxylin and eosin-stained sections from the original pre-treatment diagnostic biopsies were requested for central pathology review. Diagnoses were established using the World Health Organization criteria [4]. The characteristic morphologic features of MCL along with supporting documentation of B-cell monoclonality, co-expression of CD5 and Cyclin D1 protein, and/or the t(11;14)(q13;q32) translocation demonstrated by cytogenetics or fluorescent in situ hybridization (FISH) were required for inclusion.

The treatment regimen used was the MDA R-HyperCVAD/MTX/AraC regimen as published by Romaguera et al. in the manuscript entitled ‘High Rate of Durable Remissions After Treatment of Newly Diagnosed Aggressive Mantle-Cell Lymphoma With Rituximab Plus Hyper-CVAD Alternating With Rituximab Plus High-Dose Methotrexate and Cytarabine [5].’

The primary end point was PFS. We planned to accrue 50 eligible patients, which was sufficient to estimate the 1-year PFS rate (given complete follow-up) to within 14% (95% CI). Given historical data at the time the study were initiated, we considered a 1-year PFS estimate of ≥68% to warrant further investigation of this therapy. Fifty patients were also sufficient to estimate the best response rate, 1-year survival rate and toxicity rates to within ±14%. Any toxicity occurring with at least 5% probability was likely to be seen at least once (92% chance). Toxic effects were coded using the NCI's CTCAE, v3.0 and PFS was defined as the time from registration to the first observation of progressive disease or death due to any cause. Survival was estimated according to the method of Kaplan and Meier [6]. Analyses of survival differences by prognostic factors were carried out using Cox regression [7]. This report presents results with follow-up of ≥4 years. For the Ki67 analyses, the reviewing pathologist (WRB) and technologist were blinded to all outcome data. Ki67 index was determined by immunohistochemistry using the MIB-1 clone (DAKO; 1/100; flex polymer system).

results

characteristics of the patients

A total of 56 patients were registered. Seven patients were ineligible; two patients had no measurable disease at baseline, one patient was older then 70 years of age and four patients had insufficient pre-study information. The median follow-up among patients still alive was 4.8 years (maximum, 7.5 years).

The clinical and pathological characteristics of the 49 eligible patients are shown in Table 1. The median age was 57.4 years with a range of 35–69.8 years; 7 patients were >65 years of age. As expected, the majority of patients (78%) were male. All patients had either stage III or IV disease. Thirty-seven percent of the patients reported ‘B’ symptoms. Eight percent of the patients met criteria for having bulky disease. Forty-one percent of the patients had an ECOG performance status (PS) of 1–2 with 59% having a PS of 0. According to their International Prognostic Index (IPI), 33%, 35%, 22% and 10% of patients were classified as having low, low-intermediate, high-intermediate and high-risk disease, respectively [8]. According to the Mantle Cell Lymphoma International Prognostic index (MIPI), 55%, 31% and 14% of patients were classified as low, intermediate and high risk, respectively [9]. The most common histological pattern of disease was mantle zone, seen in 57% of the cases. Twenty seven percent and 6% of the cases were diffuse and nodular, respectively. The blastoid variant was seen in 8% of the cases. The Ki67 score was centrally assessed in 17 patients by immunohistochemistry on a tissue microarray. The median Ki67 was 28 (range 5–78).

Table 1.

Patient characteristics

		Number of patients	Percent of patients
Total patients		49
Age, years
Median (range) (years)	57.4 (35.0–69.8)
≤65		42	86
>65		7	14
Sex, male		38	78
ECOG PS 1 or 2		20	41
Stage III or IV		49	100
B symptoms		18	37
Blastoid variant		4	8
IPI
Low		16	33
Low-intermediate		17	35
High-intermediate		11	22
High		5	10
MIPI
Low		27	55
Intermediate		15	31
High		7	14
Ki-67a
Median (range)	28 (5–78)

^aAssessed only in 17 patients for which the blocks were available to make tissue microarray.

ECOG, Eastern Cooperative Group; IPI, International Prognostic Index; MIPI, Mantle Cell Lymphoma International Prognostic Index.

response

All 49 eligible patients were assessable for response. Seven patients had inadequate response assessment and are assumed to be non-responders. The overall response rate was 86% (95% CI, 73% to 94%). Twenty-three patients had a CR (47%) with four patients having an unconfirmed CR (Cru 8%) resulting in a CR/Cru rate of 55%. Fifteen patients had a partial response (PR, 31%). The response rate was 86% for both younger (≤65 years) and older patients (66–70 years). The median time until the achievement of first response (PR or CR) was 3 months, or ∼4.5 cycles under the planned treatment schedule (i.e. 21-day cycles).

progression-free survival (PFS)

All 49 eligible patients were assessable for PFS. Twenty-four of the 49 eligible patients have relapsed or died. The 1-year PFS estimate is 90% (95% CI, 81% to 98%). The lower 95% confidence bound for 1-year PFS (81%) exceeded the protocol-specified alternative rate of 68%. The median PFS was 4.8 years (95% CI, 3 years-infinity). The estimated 3- and 5-year PFS was 66% (95% CI, 52% to 80%) and 49% (95% CI, 34% to 65%), respectively (Table 2, Figure 1A).

Table 2.

Progression-free survival (PFS) and overall survival (OS)

	N	Median		3 year		5 year
		Estimate (years)	95% CI	Estimate (%)	95% CI	Estimate (%)	95% CI
Progression-free survival (PFS)
Overall	49	4.8	3.0-inf	66	52–80	49	34–65
Age (years)
≤65	42	5.5	3.8-Inf	73	59–87	53	36–70
>65	7	1.6	0.7-Inf	29	0.1–62	29	0.1–62
IPI
Low/low-inter	33	NR	3.7-Inf	71	55–87	55	37–72
High-inter/high	16	3.9	1.2–5.5	56	32–81	33	2–65
MIPI
Low	27	NR	3.8-Inf	81	66–96	58	39–77
Intermediate	15	4.8	1.2–5.5	53	28–79	36	2–69
High	7	1.6	0.5–1.6	36	0.1–74	36	0.1–74
Overall survival (OS)
Overall	49	6.8	3.7-inf	81	70–92	63	48–78
age (years)
≤65	42	6.8	6.8-Inf	85	74–96	69	53–85
>65	7	3.1	0.9-Inf	57	20–94	29	0.1–62
IPI
Low/low-inter	33	NR	4.9-Inf	84	71–97	68	50–86
High-inter/high	16	6.8	2.8-Inf	75	54–96	56	32–81
MIPI
Low	27	NR	ND	88	76–100	78	60–96
Intermediate	15	4.7	1.8–6.8	80	60–100	48	19–77
High	7	3.1	0.7-Inf	50	10–90	33	0.1–71

ND, not defined; NR, not reached; Inf, infinity; IPI, International Prognostic Index; MIPI, mantle cell lymphoma International Prognostic Index.

Figure 1.

Kaplan–Meier estimates of progression-free survival (PFS) from diagnosis; (A) of the 49 patients evaluable for response; (B) by age (P = 0.08); (C) by MIPI score (P = 0.07).

Median, 3-year and 5-year Kaplan–Meier estimates of PFS according to the subgroups of age, IPI and MIPI are shown in Table 2 and illustrated in Figure 1B and C. In univariate models, non-statistically significant trends toward improved PFS were observed for younger patients (P = 0.08) and patients with low/low intermediate IPI (P = 0.20). The MIPI level (low, intermediate and high risk) was marginally predictive of PFS (P = 0.07) as an ordered categorical variable. Finally, there was no difference in the PFS of the patients treated in the community setting (through the Community Clinical Oncology Program; CCOP) and the patients treated through the academic and affiliate institutions of SWOG (P = 0.61). Given that the Ki67 analysis was only carried out on a subset of patients from whom a tissue microarray could be generated, the predictive value of the Ki67 was not analyzed.

overall survival (OS)

All 49 eligible patients were assessable for OS. Seventeen patients have died. The 1-year overall survival estimate was 92% (95% CI, 84% to 99%). The median OS was 6.8 years (95% CI, 3.7 years-infinity). The 3- and 5-year OS estimates are 81% (95% CI, 70% to 92%) and 63% (95% CI, 48% to 78%), respectively (Table 2, Figure 2A).

Figure 2.

Kaplan–Meier estimates of overall survival (OS) from diagnosis (A) of the 49 patients assessable for response, (B) by age (P = 0.01), (C) by MIPI score (P = 0.01).

Median, 3-year and 5-year Kaplan–Meier estimates of overall survival according to the subgroups of age, IPI and MIPI are shown in Table 2 and illustrated in Figure 1B and C. In univariate models, younger patients were observed to have improved survival compared with older patients (P = 0.01), while a non-statistically significant trend towards improved survival was observed for patients with low/low intermediate IPI (P = 0.22). The MIPI level (low, intermediate or high risk) was a significant predictor of OS (P = 0.01). Similar to that seen with PFS, there was no observed difference in the OS of patients treated at CCOP institutions compared with that of patients treated at member or affiliate institutions (P = 0.46).

toxicity

All 49 eligible patients were evaluated for toxicity. A total of eight cycles of therapy was planned according to the clinical protocol. Nineteen (39%) patients were removed early from the protocol due to toxicity [43% (3 of 7) among patients >65 years of age]. The percentage of patients who actually received a given number of treatment cycles was 53% (eight cycles); 8% (six cycles); 12% (five cycles); 14% (four cycles); 2% (three cycles); 8% (two cycles); 2% (one cycle).

There was one treatment-related death. This was a man of 62 years of age whom was admitted with neutropenic fever after the fifth cycle of therapy and developed C difficile colitis, toxic megacolon and sepsis.

Two cases of myelodysplastic syndrome (MDS) were reported. The first was in a 66-year-old patient who after completion of eight cycles of therapy subsequently received rituximab, ifosfamide, carboplatin and etoposide. Approximately 13 months after the last protocol therapy, the patient had a bone marrow which showed mild to moderate dysgranulopoiesis with cytogenetic analysis revealing a new der(6;17) clone raising the possibility of an evolving MDS, although the morphological findings were thought to be insufficient to be diagnostic. The second patient was a 59-year-old who received three cycles of protocol therapy and came off study due to prolonged thrombocytopenia. The patient subsequently received CHOP chemotherapy and then had an autologous stem cell transplant with a busulfan, cyclophosphamide and etoposide preparative regimen. About 2 years after transplant, the patient developed MDS, which subsequently evolved into acute myeloid leukemia.

Table 3 shows toxic effects for which more than 10% of patients experienced the broadly categorized toxicity. Forty-four patients (90%) had grade 4 hematologic toxicity; grade 4 anemia, leukopenia, neutropenia and thrombocytopenia seen in 6, 42, 38 and 35 patients, respectively. Table 4 also shows selected grade 3 and 4 non-hematological toxic effects. Grade 4 non-hematological toxic effects included ataxia (1 patient), hyperglycemia (1 patient) and infection with grade 3–4 neutropenia (1 patient). Grade 3 febrile neutropenia was seen in 10 patients. Grade 3 infection with and without grade 3–4 neutropenia was seen in 15 and 4 patients, respectively. Five patients had grade 3 respiratory infections (three with and two without neutropenia) and one patient had a grade 3 catheter-related infection. Other grade 3 non-hematological toxic effects seen in at least 10 patients included flu-like symptoms (11 patients; primarily fatigue) and neurological symptoms (11 patients).

Table 3.

Toxicity (N = 49)

	Percentage
	Grade 3	Grade 4
Hematologic Toxicity	0	90
Neutropenia	10	78
Thrombocytopenia	16	71
Anemia	61	12
Leukopenia	2	86
Platelet transfusions	16	0
PRBC transfusions	24	0
Infection and fever	55	2
Febrile neutropenia	20	0
Infection with Grade 3-4 neutropenia	31	2
Infection without Grade 3-4 neutropenia	8	0
Respiratory tract infection with neutropenia	6	0
Respiratory tract infection with no neutropenia	4	0
Urinary tract infection with neutropenia	2	0
Catheter-related infection	2	0
Cardiovascular	10	0
Flu-like illness	22	0
Fatigue/malaise/lethargy	22	0
Gastrointestinal	12	0
Colitis (C. dificile)	0	0
Hemorrhage	14	0
Lung	12	0
Metabolic	18	2
Hyperglycemia	14	2
Neurologic	22	2
Ataxia	6	2
Pain	8	2
Myalgia/arthralgia NOS	2	2
Secondary malignancy	0	2

NOS, not otherwise specified.

Table 4.

Comparison of the current study with other studies [3, 5, 10–13]

Variable	SWOG	MDA	GISL	Nordic	GELA
Median follow-up (years)	4.8	8	3.8	6.5	5.6
Number of patients	49	97	60	160	60
Age (years)					Median 57.5 years
≤65 (%)	86	67	≤60 years 55	≤60 years 74
>65 (%)	14	33	>60 years 35	>60 years 26
MIPI (%)
Low	55	51	60	51	55
Intermediate	31	29	31	26	32
High	14	16	9	23	13
Blastoid (%)	8	14	-	19	-
Median Ki-67(%)	28a	8.4	-	-	c
% of patients not completed full treatment	47	29	63	9.3	18
Treatment-related mortality (%) (including death from 20 MDS/AML)	2	9	5	4.4	1.6
Patients w/20 MDS/AML (%)	4	5	1.6	0.6	-
PFSb
Median (years)	4.8	4.6		EFS 7.4	7
≤65 years (5-year estimate)	53%	46%	3 years est. 6.1	-
>65 years (5-year estimate)	29%	33%	5 years est. 5.1	-
OS
Median (years)	6.8	NR		NR; >10 years	NR
≤65 years (5-year estimate)	63%	68%	3 years est 6.9
>65 years (5-year estimate)	29%	33%	5 years est 6.1

^aBased on n=17 patients with available samples.

^bDefined as time-to-treatment failure (TTF) for the MDA study.

^c0–9% Ki67 is 8.3%; 10–29% Ki67 is 50%; >29% Ki67 is 41.7%.

AML, acute myeloid leukemia; PFS, progression-free survival; OS, overall survival; NR, not reached; -, data not reported; MIPI, Mantle Cell Lymphoma International Prognostic Index.

discussion

The results of this US multi-institutional study confirm that of the single institutional study at MD Anderson which demonstrates that this regimen is indeed an effective treatment approach for patients ≤65 years of age. As seen in the MDA series, patients >65 years of age have a poor outcome to treatment and such patients should, therefore, not be treated with this regimen. When limiting the analyses to patients ≤65 years of age, the median PFS of 5.5 years with a 5-year PFS and OS of 53% and 69%, respectively, compare favorably with that seen in the MDA series. Indeed, despite differences in both the median time of follow-up and patient characteristics between the SWOG and MDA series, as shown in Table 4, the PFS (or time-to-treatment failure as reported in the MDA series) and OS were at least 50% at 5 years in both series for patients ≤65 years [3, 5]. That such favorable results were seen in the current study is significant, as in contrast to the MDA study, almost half of the patients in the current study were treated in community cancer centers that were CCOP affiliates of SWOG. Indeed, the outcome of patients treated at such CCOP institutions was no different than that of patients treated at the academic centers and their affiliates. Finally, the results of this study compare favorably with that of other series, which employ a dose intensive approach [10–13].

In a recent National Comprehensive Cancer Center Network (NCCN), retrospective analysis of 167 MCL patients <65 years of age treated at member NCI Comprehensive Cancer Centers and not treated on a clinical trial, 50% received R-HyperCVAD/MTX/AraC [14]. Of interest, the outcome of such patients appears to be somewhat inferior to that of patients in the current trial (which included CCOP community-based practices) with the 3-year PFS in the NCCN and SWOG study (for those patients ≤65 years of age) being 58% and 73%, respectively, with a similar 3-year OS of 85% in both the groups [14]. Such a difference is likely due in part to differences in the patient characteristics between these studies although based on MIPI criteria the patients were relatively similar in both the studies.

Similar to that of the MDA series, ∼50% of the patients in this study were at low risk by MIPI criteria (55% in SWOG series, 51% in MDA series). Indeed, MIPI was predictive of treatment outcome in both the studies, with the median PFS and OS for the low-risk MIPI patients not being reached in the present study, whereas the high-risk MIPI patients had a median PFS and OS of 1.6 and 3.1 years, respectively. The fact that about half of the patients in both the current study and that of MDA were at low risk by MIPI suggests that the overall favorable outcome seen in both the studies is driven in part by the favorable outcome of this low-risk group. This then leads to two questions. First, is the low-risk group the group of patients that truly benefits from dose intensive therapy? Alternatively would this low-risk group have done just as well with conventional dose chemotherapy and would thus have been spared the toxicity inherent with this and other dose intensive approaches? Such critical questions will only be answered in the context of randomized trials.

Although this is an effective regimen, it is also a toxic regimen, with 39% of the patients unable to complete the prescribed eight cycles of therapy predominately due to toxicity. Similarly, in the MDA experience, 29% could not complete eight cycles [5]. Given the toxic nature of this regimen, studies to determine the optimal number of treatment cycles may be warranted to assess whether a similar positive clinical outcome can be achieved with less toxicity using fewer than the eight prescribed cycles defined in this regimen. Furthermore, whereas there is recent evidence to suggest that high-dose cytarabine is an important component of therapy for MCL, the role of high-dose methotrexate, which almost certainly adds to the toxicity of this regimen, is questionable [12, 15, 16]. Finally, the potential risk of MDS/acute myeloid leukemia (AML) is concerning, with two cases seen in our study and now five cases of MDS seen in the MDA study with longer follow-up [3].

Recently, the Gruppo Italiano Studio Linfomi reported their findings of a multicenter trial of R-HyperCVAD/MTX/AraC in 60 patients ≤70 years of age [13]. There was significant activity of this regimen seen, with an estimated 5-year OS and PFS of 73% and 61%, respectively. Toxicity was again significant, with 3 treatment-related deaths, 13 patients with grade 3–4 non-hematological toxicity and 1 patient with a secondary AML reported in this series [13]. Taken together with our results and that of MDA, this regimen is associated with significant toxicity as are other dose intensive approaches [10–12]. For example, the Nordic regimen was associated with a 5% non-relapse mortality, and 17% and 12% of the cycles delivered resulted in hospitalization for grade 3 and grade 4 toxic effects, respectively. A determination of which regimen is optimal when taking clinical benefit and toxicity into account can only be achieved with a randomized trial [11, 12].

In summary, our data suggest that R-HyperCVAD/MTX/AraC is an effective yet toxic regimen for patients with MCL. The difficulty in comparing the outcomes from this phase II study with that of other phase II studies such as the Nordic group study is not only due to differences in the patient characteristics between the studies but also due to differences in the end points reported, i.e. EFS versus PFS. versus TTTF. The major question, however, given that there does not appear to be any evidence of a substantial cure rate with this, or any up-front regimen reported to date, is whether a dose intensive or non-dose intensive regimen should be the initial treatment of choice. This can only be answered in the context of a prospective, randomized trial however. The challenge is to determine what the optimal regimens are to compare in such a trial. In designing such regimens one must be cognizant of recent randomized trial data showing for example, the benefit of rituximab maintenance after R-CHOP in elderly patients and the benefit of high-dose cytarabine as part of the treatment regimen given to younger patients before high-dose therapy and autologous stem cell transplantation [17]. In addition, how such potentially effective newer novel agents such as lenolidamide [18], mTOR [19] and B-cell receptor signaling protein inhibitors [20] are to be incorporated into up-front therapy must also be determined. Finally, even if such a trial demonstrates the superiority of a dose intensive approach, whether such benefit is worth the cost regarding toxicity and quality-of-life issues, and whether the superiority is translated into a survival benefit must be determined. Until these questions are answered, however, our data support the notion that R-HyperCVAD/MTX/Ara-C is one viable regimen for patients ≤65 years of age which can be given both at academic centers and in experienced community centers.

funding

This work was supported by NIH CA32102.

disclosure

The authors have declared no conflicts of interest.