Outcomes of hematopoietic cell transplantation for diffuse large B cell lymphoma transformed from follicular lymphoma

Abstract

There are limited data on the outcomes of autologous or allogeneic hematopoietic cell transplantation in diffuse large B-cell lymphoma transformed from follicular lymphoma. We analyzed transplant outcomes in 141 subjects with biopsy-proven diffuse large B-cell lymphoma transformed from follicular lymphoma reported to the Center for International Blood and Marrow Transplant Research from 1990–2009. Two groups were identified: autotransplant (N=108) and allotransplant (N=33). Fewer autotransplants were done for transformed follicular lymphoma in 2003–2009, with a shift favoring allotransplants. Autotransplant had 1 year non-relapse mortality of 8% (95% confidence intervals [CI] 4–14), 5 year progression free survival of 35% (95% CI 26–45), and 5 year overall survival of 50% (95% CI 40–59). Allotransplant had 1 year non-relapse mortality of 41% (95% CI 23–58), 5 year progression free survival of 18% (95% CI 6–35), and 5 year overall survival of 22% (95% CI 8–41). Autotransplant for transformed follicular lymphoma achieves sustained remission in a high proportion of subjects. The high non-relapse mortality of allotransplant obscured any benefit that might be associated with this transplant modality.

Introduction

Follicular lymphoma (FL) is the second most common non-Hodgkin lymphoma (NHL) in the western hemisphere.^1,2 The histologic transformation of FL to diffuse large B-cell lymphoma (DLBCL) occurs in up to 30% of patients at 10 years.^3–7 The rate of transformed FL (tFL) varies according to the definition of transformation (whether the definition includes only DLBCL, or also Burkitt lymphoma, grade 3B FL, and composite or discordant lymphomas), method of diagnosis (biopsy, cytology or clinical suspicion), duration of follow up, and inclusion of autopsy data.⁸

Compared with FL, in which the median survival is historically in the 10 year range without a plateau, histologic transformation of FL is usually associated with chemotherapy resistance and shorter survival after chemotherapy.^9–13 There is no standard of care for tFL. Therapy for DLBCL transformed from FL is mainly based on guidelines for de novo advanced DLBCL. Because patients with tFL are typically excluded from FL and DLBCL clinical trials, there are limited data on the role of autologous or allogeneic hematopoietic cell transplantation. Most reports are small retrospective studies with brief follow up.^14–20 Outcomes vary greatly because of different inclusion criteria.^14–20 In addition, most of the studies were done before the availability of rituximab, an agent that has improved the outcomes of FL and DLBCL.^14–20 We analyzed the outcomes of transplants for biopsy-proven transformation of FL to DLBCL in a larger patient cohort reported to the Center for International Blood and Marrow Transplant Research (CIBMTR).

Methods

Data Source

The CIBMTR comprises a voluntary network of more than 500 transplantation centers globally that submit comprehensive data on consecutive autotransplants and allotransplants to a centralized statistical center. The CIBMTR is a combined research program of the Medical College of Wisconsin and the National Marrow Donor Program. Protected health information during the performance of this observational research is collected and maintained in the CIBMTR's capacity as a Public Health Authority under the Health Insurance Portability and Accountability Act (HIPAA). The observational research is conducted with a waiver of informed consent and in compliance with all applicable federal regulations regarding the protection of human research participants as assessed by the Institutional Review Board and the Privacy Officer at the Medical College of Wisconsin. Further details of the data source are described by Horowitz.²¹

Patient Population

Patients 18 years of age or older with FL at diagnosis by the World Health Organization classification, with subsequent biopsy-proven histologic transformation to DLBCL were included in this study.²² Pathology reports from the centers were reviewed in each case at the CIBMTR to confirm transformation to DLBCL. Histologic transformation to DLBCL was defined as large centroblasts diffusely infiltrating the lymph node and effacing the follicular architecture. No cases of composite or discordant lymphoma at diagnosis were included. Histologic transformation of other low-grade lymphomas, such as marginal zone lymphoma, chronic lymphocytic leukemia, was excluded. In addition, cases with transformation to histology other than DLBCL, such as Burkitt lymphoma, lymphoblastic lymphoma or Hodgkin lymphoma, were also excluded.

Subjects who had an initial single autotransplant or allotransplant for tFL were included. Subjects who had a prior transplant for FL before transformation or after transformation were excluded.

Study End Points and Definitions

The main objective of this study was to describe the outcomes of autotransplant and allotransplant for patients with DLBCL transformed from FL. The primary end point was overall survival (OS) and other end points of interest were the progression free survival (PFS), relapse/progression, non-relapse mortality (NRM), and the incidence of acute graft-versus-host disease (aGVHD) and chronic GVHD (cGVHD). OS was defined as time to death after transplant. Death from any cause was considered an event. Surviving patients were censored at time of last follow up. PFS was defined as survival without disease relapse or progression after transplant. Relapse or progression of disease and death were events. Relapse/progression was defined as any new lesion after complete remission or increase in size of previously involved sites after transplant with NRM as a competing risk.²³ NRM was defined as any death within the first 28 days of transplant or any death occurring after day 28 in the absence of disease relapse/progression. Relapse was a competing risk. Those who survived without relapse or progression were censored at last contact for PFS, progression/relapse and NRM. Acute GVHD was diagnosed by established criteria.²⁴ Chronic GVHD was defined by standard criteria.²⁵ The intensity of the conditioning regimen was defined by CIBMTR criteria.²⁶ Related and unrelated donor (URD) transplant recipients were classified based on available HLA typing as previously described by Weisdorf et al.²⁷

Statistical Methods

Univariate probabilities of OS and PFS for the autotransplant and allotransplant cohorts were calculated using the Kaplan-Meier estimator, with the variance estimated by Greenwood's formula.²⁸ Variables considered for univariate analysis were age, Karnofsky performance status, presence of extranodal disease at transplant, prior use of rituximab, chemoresistance, interval from diagnosis of FL to transformation, and total body irradiation (TBI)-based conditioning. For allotransplant, additional variables tested were use of reduced intensity/non-myeloablative conditioning (RIC/NMAC), antithymocyte globulin (ATG), alemtuzumab, and donor source. Relapse/progression, NRM and the incidence of acute and chronic GVHD were estimated using cumulative incidence estimates to accommodate for competing risk.

Cox proportional hazards regression model was used to identify the risk factors significantly associated with treatment failure (1 – PFS) and overall mortality (1 – OS) for autotransplant. No multivariate analysis was considered for allotransplant due to small size of the study cohort. The variables considered in the multivariate models are listed in Table 2a. The assumption of proportional hazards for each factor in the Cox model was tested by adding a time-dependent covariate. The proportionality assumption was satisfied for each factor. A forward and backward stepwise model selection approaches were used to identify all significant risk factors.

Table 2a.

Univariate survival analysis for autologous hematopoietic cell transplantation

		Treatment Failure (1-PFS)		Overall Mortality (1-OS)
Covariates	N	Hazard ratio (95% CI)	P-value	Hazard ratio (95% CI)	P-value
TBI based conditioning
No	84	Reference		Reference
Yes	24	1.27 (0.75–2.14)	0.38	1.41 (0.82–2.44)	0.22
Chemo sensitivity			0.38a		0.32a
Sensitive	90	Reference		Reference
Resistant	10	1.89 (0.90–3.97)	0.09	1.98 (0.89–4.40)	0.09
Time from diagnosis to transformation
< 1 year	17	Reference		Reference
≥1 year	91	0.69 (0.37–1.29)	0.25	0.68 (0.35–1.31)	0.25
Age (years)
> 60	39	Reference		Reference
≤ 60	69	1.06 (0.66–1.72)	0.80	1.20 (0.71–2.02)	0.50
Karnofsky score			0.50a		0.71a
<90%	35	Reference		Reference
90–100%	68	0.76 (0.46–1.24)	0.27	0.81 (0.48–1.37)	0.43
Extranodal disease prior to HCT			0.25a		0.66a
no	75	Reference		Reference
yes	28	0.95 (0.56–1.60)	0.85	0.84 (0.48–1.49)	0.55
Rituximab exposure prior to HCT
no	78	Reference		Reference
yes	30	0.99 (0.59–1.67)	0.98	0.65 (0.35–1.20)	0.17

^amulti degree-of-freedom overall test.

Abbreviations for tables 1,2: OS overall survival, PFS progression free survival, N number of patient evaluated, # number, dx diagnosis, 95% CI confidence interval, Prob probability, autoHCT autologous hematopoietic cell transplantation, alloHCT allogeneic hematopoietic cell transplantation without prior autoHCT, chemoRx chemotherapy, TBI total body irradiation, MAC myeloablative conditioning, RIC/NMAC reduced intensity conditioning/ non-myeloablative conditioning, FL follicular lymphoma, tFL transformed follicular lymphoma, BM bone marrow, PB peripheral blood, GVHD graft-versus-host disease, FK tacrolimus, CSA cyclosporine, TCD T cell depletion, ATG antithymocyte globulin, CAMPATH alemtuzumab, MRD matched related donor, URD unrelated donor, CR complete remission, PR partial remission, REL relapsed, PIF primary induction failure, NE non evaluable, KPS Karnofsky Performance status.

Results

Subject, Disease and Transplant Related Variables

Two groups reported to the CIBMTR from 1990–2009 were identified: 108 subjects underwent an autotransplant and 33 underwent an allotransplant without a prior autotransplant (Table 1). Median follow-ups were 85 and 64 months. Overall completeness index of the follow-up for the population at 5 years was 89%.²⁹ Most autotransplants were from 1990–2002 compared with 2003–2009 for most allotransplants. Median interval from diagnosis of FL to tFL was 47 and 48 months. Disease variables at the time of diagnosis of tFL were unavailable. Median intervals from diagnosis of FL to transplant for auto- and allotransplant cohorts were similar (54 vs. 55 months) as was the interval from the diagnosis of tFL to transplant (6 vs. 8 months).

Table 1.

Characteristics of patients with transformed diffuse large B-cell lymphoma from follicular lymphoma who received Hematopoietic Cell Transplant (HCT) from 1990–2009

	Auto HCT	Allo HCT
Variable	N (%)	N (%)
Number of patients	108	33
Age, median (range)	56 (19–74)	49 (31–66)
Gender
Male	65 (60)	20 (61)
Female	43 (40)	13 (39)
Karnofsky score
< 90%	35 (32)	8 (24)
90–100%	68 (63)	25 (76)
Missing	5 (5)	0
Stage at diagnosis
I–II	32 (30)	9 (27)
III–IV	72 (67)	23 (70)
Missing	4 (4)	1 (3)
Disease status prior to HCT
CR1	9 (8)	2 (6)
CR2	23 (21)	7 (21)
Primary induction failure sensitive	13 (12)	7 (21)
Primary induction failure resistant	3 (3)	3 (10)
Relapsed sensitive	39 (36)	7 (21)
Relapsed resistant	5 (5)	5 (15)
Relapse untreated	5 (5)	1 (3)
Missing*	11 (10)	1 (3)
Chemosensitivity prior to HCT
Sensitive	90 (83)	23 (70)
Resistant	10 (10)	8 (24)
Untreated/unknown	8 (7)	2 (6)
Number of lines of chemo prior to HCT
1–2	33(31)	7(21)
≥3	66(61)	26(79)
Missing	9(8)	0
Rituximab given between diagnosis and HCT
No	78(72)	11(33)
Yes	30(28)	22(67)
Known extranodal disease immediately prior to HCT
No	75 (69)	24 (73)
Yes	28 (26)	9 (27)
Missing	5 (5)	0
Size of involved lymph nodes at HCT
Smaller than 5 cm	18 (17)	4 (12)
Equal to or greater than 5 cm	14 (13)	2 (6)
No lymphadenopathy at HCT	35 (32)	19 (58)
Missing	41 (38)	8 (24)
Interval between diagnosis and HCT (months)	54 (6–347)	55 (8–203)
Interval between diagnosis and transformation (months)	47 (1–281)	48 (1–173)
Interval between HCT and transformation (months)	6 (2–76)	8 (1–31)
Conditioning regimen
Myeloablative	108 (100)	20 (61)
Reduced intensity		11 (33)
Missing		2 (6)
Total body irradiation based conditioning
No	84 (78)	20 (61)
Yes	24 (22)	12 (36)
Missing	0	1 (3)
Graft type
Bone marrow	16 (15)	10 (30)
Peripheral blood	92 (85)	23 (70)
Use of antithymocyte globulin (ATG)/Alemtuzumab
ATG/Alemtuzumab	0	10 (30)
No ATG/Alemtuzumab	0	23 (70)
Auto	108 (100)	0
Type of donor
Matched related donor	0	15(45)
Matched unrelated donor	0	9(27)
Other	0	9(27)
Autologous	108 (100)	0
Year of transplant
1990–1994	32 (30)	1 (3)
1995–2002	51 (47)	9 (27)
2003–2009	25 (23)	23 (70)
Median follow-up of survivors (range), months	85 (3–233)	64 (3–97)

^*Chemosensitivity known for 8 cases

Abbreviations for tables 1,2: OS overall survival, PFS progression free survival, N number of patient evaluated, # number, dx diagnosis, 95% CI confidence interval, Prob probability, autoHCT autologous hematopoietic cell transplantation, alloHCT allogeneic hematopoietic cell transplantation without prior autoHCT, chemoRx chemotherapy, TBI total body irradiation, MAC myeloablative conditioning, RIC/NMAC reduced intensity conditioning/ non-myeloablative conditioning, FL follicular lymphoma, tFL transformed follicular lymphoma, BM bone marrow, PB peripheral blood, GVHD graft-versus-host disease, FK tacrolimus, CSA cyclosporine, TCD T cell depletion, ATG antithymocyte globulin, CAMPATH alemtuzumab, MRD matched related donor, URD unrelated donor, CR complete remission, PR partial remission, REL relapsed, PIF primary induction failure, NE non evaluable, KPS Karnofsky Performance status.

Rituximab was given pretransplant in 28% and 61% of the cohorts (Table 1). Radioimmunotherapy with ⁹⁰Y-ibritumomab tiuxetan or ¹³¹I-tositumomab was given to 5 subjects (1 autotransplant [0.9%] and 4 [12%] allotransplant recipients). Most of the subjects were responsive to chemotherapy before autotransplant (83%) and allotransplant (70%). However patients with poor risk features at transplant were also included in this study. Chemoresistance at transplant was present in 10% of autotransplant and 24% of allotransplant patients. Median numbers of lines of chemotherapy were 3 and 4 indicating a heavily pretreated group. Further, bulky lymphadenopathy of 5 cm or more was present at transplant in 13% of autotransplant and 6% of allotransplant patients. In addition, extranodal disease at transplant was present in 26% of the autotransplant group and 27% of the allotransplant group indicating a sizeable fraction of patients with poor risk features at transplant. Cytogenetic data were not available.

Carmustine, etoposide, cytarabine, melphalan (BEAM) was used in 27% and cyclophosphamide, carmustine, etoposide (CBV) in 29% of autotransplant recipients. Most allotransplant recipients (61%) received myeloablative conditioning (MAC) with cyclophosphamide, total body irradiation (Cy/TBI). The remaining subjects had RIC/NMAC, primarily fludarabine-2 Gy TBI (Flu-TBI), fludarabine-cyclophosphamide (Flu-Cy), or fludarabine-melphalan (Flu-Mel).

Transplantation Outcomes

The autotransplant cohort had a 1 year NRM of 8% (95% CI 4–14), 5 year probability of relapse/progression of 54% (95% CI 44–63), 5 year PFS of 35% (95% CI 26–45), and 5 year OS of 50% (95% CI 40–59). There was a plateau after 40 months at 39% (95% CI 30–48) for PFS and 53% (95% CI 43–63) for OS (Figure 1). Causes of death included relapse/progression (41%) and second cancers (4%).

Figure 1.

Kaplan-Meier curves for progression free survival and overall survival for autologous hematopoietic cell transplantation

The allotransplant cohort had a 1 year NRM of 41% (95% CI 23–58), 5 year probability of relapse/progression of 33% (95% CI 17–50), 5 year PFS of 18% (95% CI 6–35) and 5 year OS of 22% (95% CI 8–41). The cumulative incidences of grades II–IV and III–IV aGVHD at day 100 were 42% (95% CI 26–59) and 27% (95% CI 14–42), respectively. The cumulative incidence of cGVHD at 1 year was 26% (95% CI 13–43). The major causes of death were organ failure (24%), relapse/progression (18%) and GVHD (12%). For MAC allotransplant, 3 year PFS was 11% (95% CI 1–30) and 3 year OS was 11% (95% CI 1–29). Three year PFS was 48% (95% CI 18–79) and 3 year OS was 67% (95% CI 35–93) for NMAC/RIC allotransplant. One year NRM for MAC allotransplant was 57% (95% CI 31–77). Patients who underwent RIC/NMAC allotransplant did not experience NRM within 5 years but 5 of the 13 (38%) patients died of relapse/progression.

Univariate Analysis

For autotransplant, age, Karnofsky performance status, presence of extranodal disease at HCT, rituximab use pre-transplant, time from diagnosis of FL to tFL of ≥1 year vs. < 1 year, chemosensitivity, or TBI conditioning did not have a statistically significant impact on PFS/OS (Table 2a). Notably, subjects with chemotherapy resistant disease achieved a 3 year OS of 27% (95% CI 7–59) after autotransplant.

For allotransplant, age, Karnofsky performance status, presence of extranodal disease at HCT, pre-transplant rituximab use, time from diagnosis of FL to tFL of ≥1 year vs. < 1 year, chemosensitivity, use of ATG or alemtuzumab or donor source did not have a statistically significant impact on PFS/OS (Table 2b). Subjects with chemotherapy resistant disease achieved a 3 year OS of 21% (95% CI 0–62) after allotransplant. The 3 year PFS was 11% (95% CI 1–30) for MAC and 48% (95% CI 18–79) for RIC/NMAC, p=0.001. The 3 year OS was 11% (95% CI 1–29) for MAC and was 67% (95% CI 35–93) for RIC/NMAC, p<0.001. RIC/NMAC was associated with significantly higher PFS and OS compared to MAC (Table 2b and Figure 2a and 2b).

Table 2b.

Univariate survival analysis for allogeneic hematopoietic cell transplantation

		PFS		OS
Covariates	N	Hazard ratio (95% CI)	P-value	Hazard ratio (95% CI)	P-value
TBI based conditioning
No	20	Reference		Reference
Yes	12	1.07 (0.46–2.49)	0.88	0.93 (0.40–2.18)	0.87
Chemo sensitivity
Sensitive	23	Reference		Reference
Resistant	8	1.36 (0.52–3.52)	0.53	1.30 (0.50–3.37)	0.59
Time from diagnosis to transformation
Less than a year	9	Reference		Reference
A year or greater	24	0.43 (0.18–1.03)	0.06	0.53 (0.23–1.25)	0.15
Age (years)
> 50	15	Reference		Reference
≤ 50	18	0.66 (0.29–1.50)	0.32	0.64 (0.28–1.45)	0.28
Karnofsky score
<90 %	8	Reference		Reference
90–100%	25	0.88 (0.34–2.24)	0.79	0.80 (0.31–2.05)	0.64
Extranodal disease pre HCT
no	24	Reference		Reference
yes	9	2.21 (0.91–5.37)	0.08	2.06 (0.86–4.96)	0.11
Rituximab exposure pre HCT
no	11	Reference		Reference
yes	22	1.28 (0.52–3.11)	0.59	1.40 (0.57–3.40)	0.46
ATG/Alemtuzumab
ATG and/or Alemtuzumab	10	Reference		Reference
No ATG and/or Alemtuzumab	23	1.51 (0.59–3.84)	0.39	1.66 (0.65–4.23)	0.29
Conditioning regimen intensity			0.01a		0.007a
Myeloablative	20	Reference		Reference
RIC/NMAC	11	0.16 (0.04–0.54)	0.004	0.14 (0.04–0.49)	0.002
Disease status at transplant			0.24a		0.31a
CR	8	Reference		Reference
PIF/REL	21	1.07 (0.38–2.98)	0.90	0.90 (0.32–2.52)	0.84
Donor type			0.53a		0.29a
HLA identical sibling	12	Reference		Reference
URD	18	0.95 (0.39–2.32)	0.92	0.82 (0.34–2.01)	0.67

^amulti degree-of-freedom overall test.

Abbreviations for tables 1,2: OS overall survival, PFS progression free survival, N number of patient evaluated, # number, dx diagnosis, 95% CI confidence interval, Prob probability, autoHCT autologous hematopoietic cell transplantation, alloHCT allogeneic hematopoietic cell transplantation without prior autoHCT, chemoRx chemotherapy, TBI total body irradiation, MAC myeloablative conditioning, RIC/NMAC reduced intensity conditioning/ non-myeloablative conditioning, FL follicular lymphoma, tFL transformed follicular lymphoma, BM bone marrow, PB peripheral blood, GVHD graft-versus-host disease, FK tacrolimus, CSA cyclosporine, TCD T cell depletion, ATG antithymocyte globulin, CAMPATH alemtuzumab, MRD matched related donor, URD unrelated donor, CR complete remission, PR partial remission, REL relapsed, PIF primary induction failure, NE non evaluable, KPS Karnofsky Performance status.

Figure 2.

Kaplan-Meier curve for (a.) progression free survival and (b.) overall survival for allogeneic hematopoietic cell transplantation by conditioning regimen intensity.

Multivariate analysis

Cox proportional hazards regression model was used to identify risk factors significantly associated with the treatment failure (1 – PFS) and overall mortality (1 – OS) for autotransplant. With a 5% significance level, no risk factors were significant. No multivariate analysis was considered for allotransplant due to the small size of the study cohort.

Discussion

In this study we determined that autotransplant for tFL results in durable remissions in a high proportion of subjects. There was an unexpected high NRM in the allotransplant recipients that may have obscured any benefit seen with allotransplant. It is interesting that clinicians seemingly favored the use of allotransplants in the 2003–2009 era. As this is a retrospective study it is impossible to know why clinicians chose autotransplant or allotransplant. The pre-transplant variables were comparable in terms of risk factors and markers of poor prognosis. Availability of a good donor is not the explanation as more than half the patients did not have a matched sibling. Perhaps this reflects a perception that autotransplant would be less effective at producing long-term remission for tFL in the rituximab era. There are limited data on the impact of prior rituximab on outcomes of autotransplant for tFL.^16,30 We found a 5 year PFS of 35% (95% CI 26–45) and 5 year OS of 50% (95% CI 40–59) after autotransplant for tFL, with a seemingly similar benefit in those with prior rituximab therapy. However, only 28% received rituximab prior to autotransplant, so firm conclusions regarding the impact of prior rituximab awaits larger studies. There was a plateau for PFS and OS, suggesting that a subset may be cured with autotransplant, although longer follow up in needed to determine if these patients remain disease free.

Our results are similar to those of the European Bone Marrow Transplant (EBMT) registry report of 50 tFL subjects from the pre-rituximab era (Table 3).¹⁷ There is only one phase 2 study in tFL, also pre-rituximab, reporting a 5 year 0S of 47% (95% CI 29–65) with autotransplant, which compared favorably to those not transplant eligible.³¹ More recently the Canadian Blood and Marrow Transplant Group (CBMTG) reported a 5 year OS of 61% (standard error 7%) with rituximab-chemotherapy alone without transplant. However, a third of this subset had received no therapy for FL before histological transformation and a quarter had limited stage tFL, a highly favorable subset likely contributing to these results.³² Even in the pre-rituximab era, watchful waiting before transformation and limited stage of tFL were significant predictors of long-term survival (median survival 81 months), albeit with a continuous risk of relapse and without a plateau on the survival curve.^3,5,33 In contrast, we demonstrated a plateau on the survival curve with autotransplant in a heavily pretreated cohort. A key limitation to analyses such as this is that without a prospective comparison, it is not certain that autotransplantation is a better approach than nontransplant therapies. However, the durability of the response is an important advantage of autotransplant for tFL. Likewise, in the largest study of radioimmunotherapy for tFL, the response rates were high but their durability was poor with 5 year PFS of only 17%.³⁴

Table 3.

Studies on outcomes of autologous hematopoietic cell transplantation for follicular lymphoma with histologic transformation

STUDY Years of HCT	N	AGE Median (range) years	tFL Definition	PFS	OS	NRM	COMMENTS
PRE RITUXIMAB ERA
Williams17 EBMT Retrospective 1982–1995	50 auto HCT	45 (28–61)	FL→DLBCL or any high grade lymphoma	2y 40% 5y 30%	2y 60% 5y 51%	18%	Improved survival if chemosensitive at HCT. Same outcomes for FL, de novo relapsed DLBCL and transformed lymphoma in matched controlled analysis.
Friedberg18 Retrospective 1982–1997	27 auto HCT	44 (29–58)	FL→DLBCL CLL→DLBCL	5y 46%	5y 58%	0	All relapses after autoHCT were DLBCL. tFL within 18 months from diagnosis had better OS.
Eide31 Prospective Phase 2 1991–2007	30 auto HCT	55 (31–65)	FL or MZL →DLBCL or between DLBCL and BL, composite lymphoma	5y 32%	5y 47%	7%	Plateau on PFS after 40 months at 30%.
RITUXIMAB ERA
Hamadani16 Retrospective 1991–2007	24 auto HCT	56 (47–68)	FL→DLBCL	3y 40%	3y 52%	8%	Improved PFS in 62% of patients with rituximab in treatment course. Chemosensitivity at HCT had no effect on outcomes after HCT.
Ban-Hoefen30 Retrospective 1998–2010	18 auto HCT	58 (40–65)	FL or MZL→DLBCL	2y 59%	2y 82%	0	Improved outcomes compared to historical pre-rituximab cohorts.17
Villa32 CBMTG Retrospective 2001–2010	97 auto HCT	56 (32–66)	FL→DLBCL or BL	5y 55%	5y 57%	5%	AutoHCT had better PFS/OS than rituximab-chemoRx.
Current study Retrospective 1990–2009	108 auto HCT	56 (19–74)	FL→DLBCL	5y 36%	5y 50%	8%	No impact of chemosensitivity or prior rituximab use on outcomes.

Abbreviations for Tables 3 and 4: OS overall survival, PFS progression free survival, NRM non-relapse-mortality, N number of patients evaluated, Y year, autoHCT autologous hematopoietic cell transplantation, alloHCT allogeneic hematopoietic cell transplantation, chemoRx chemotherapy, TBI total body irradiation, ChemoR chemoresistance, MAC myeloablative conditioning, RIC reduced intensity conditioning, NMAC non-myeloablative conditioning, FL follicular lymphoma, tFL transformed follicular lymphoma, DLBCL diffuse large B cell lymphoma, BL Burkitt lymphoma, SLL small lymphocytic lymphoma, MZL marginal zone lymphoma, EBMT European Bone Marrow Transplant Registry, CBMTG Canadian Blood and Marrow Transplant Group, CIBMTR Center for International Blood and Marrow Transplant Research, MRD matched related donor, URD unrelated donor.

In our study, RIC/NMAC allotransplant provided improved PFS and OS compared with MAC (Table 2b and Figure 2). The high NRM following MAC allotransplant influenced these outcomes and so cannot be recommended. RIC/NMAC allotransplant maybe the best strategy to solve the problem of NRM in allotransplant since the low NRM with RIC/NMAC did not obscure the potential benefits of allotransplant. Firm conclusions about the role of RIC/NMAC allotransplant in tFL await prospective comparisons with autotransplant. Nonetheless, outcomes of allotransplant for tFL appear to be inferior to FL. In the study by Khoury et al, NMAC allotransplant for FL lead to a 5 year OS of 85% (95% CI 71–93) with a plateau on the survival curve, implying a curative potential and stronger graft versus lymphoma effect.³⁶ In a recent CIBMTR analysis, GVHD was associated with lower relapse in FL but not de novo DLBCL and this effect was more prominent with RIC than MAC.³⁷ Similar to de novo DLBCL, graft versus lymphoma effects seem to be less effective in transformed DLBCL than in FL.³⁸

This study has the limitations of any retrospective study such as inherent patient selection bias as to the type of transplant performed. The comparison of autotransplant with allotransplant is biased since most autotransplants were performed a decade before allotransplants. Further, the number of allotransplant patients is low and the results of RIC/NMAC significantly better than with MAC making a direct comparison of allotransplant with autotransplant difficult. Nonetheless this multicenter study is the largest to our knowledge to describe transplant outcomes specifically in biopsy-proven DLBCL transformed from FL. Other studies have included heterogeneous low grade lymphomas transforming to a variety of high grade lymphomas, all of which can influence transplant outcomes and make the results difficult to interpret (Tables 3 and 4).^{17–20,30–32} Several conclusions can be made from our study. Firstly, allotransplant with RIC/NMAC had improved survival compared to MAC, but any potential benefit from MAC was obscured by the high NRM. The precise role of allotransplant awaits prospective comparison of RIC/NMAC allotransplant with autotransplant. Secondly, autotransplant provides durable survival for tFL irrespective of age, early or late histologic transformation, extranodal disease at transplant or prior rituximab use. The outcomes of autotransplant seem to be more durable than published data on non-transplant therapies and merits a prospective study to definitively answer this question.^{7,31,32,33,34}

Table 4.

Studies on outcomes of allogeneic HCT for follicular lymphoma with histologic transformation

Study N Years of HCT	Age Median (range) years	tFL definition	Median lines of chemoRx before HCT	Conditioning Donor	PFS	OS	NRM	Comments
Ramadan19 N=40 1989–2005	44 (28–57)	FL,SLL, MZL → intermediate or higher grade lymphoma, composite lymphoma	3 chemoR 20%	MAC MRD 25 URD 15	5y 23%	5y 23%	3y 36%	No impact on outcomes of composite vs. transformed lymphomas, URD vs. MRD. Performance of HCT within a year of diagnosis had better outcomes.
Rezvani20 N=62 of which 16 had TL 1998–2006	54 (33–66)	FL, SLL, MZL→aggressive NHL in 16 patients, remainder had low grade lymphoma.	6 ChemoR 23%	NMAC 2Gy TBI +/− FLU MRD 34 URD 28	3y 21% for TL 3y 43% for FL	3y 18% for TL 3y 52% for FL	3y 42%	Better outcomes for indolent lymphomas vs. transformed lymphoma, no impact of chemoresistance at HCT. 27 patients had prior autoHCT.
Hamadani35 N=8 1999–2007	56 (44–63)	FL→DLBCL	4 ChemoR 38%	MAC 6 RIC 2 MRD 5 URD 3	4y 56%	4y 66%	25%	No disease relapse after 1 year
Villa32 CBMTG N=22 2001–2010	48 (32–57)	FL→DLBCL, BL	3 ChemoR 18%	MAC >95% MRD14 URD 7 MMRD 1	5y 45%	5y 45%	5y 23%	2 had prior autoHCT for FL No difference in OS between alloHCT and autoHCT.
Current study N=33 1990–2009	49 (31–66)	FL→DLBCL	4 ChemoR 35%	MAC 20 RIC/NMAC 5/6 MRD 12 URD 21	5y 18%	5y 22%	iy 41% 5y 49%	No impact of chemoresistance, at HCT or URD on outcomes. Better 3y PFS/OS with RIC/NMAC (48%/67% vs. 11%) than with MAC.

Abbreviations for Tables 3 and 4: OS overall survival, PFS progression free survival, NRM non-relapse-mortality, N number of patients evaluated, Y year, autoHCT autologous hematopoietic cell transplantation, alloHCT allogeneic hematopoietic cell transplantation, chemoRx chemotherapy, TBI total body irradiation, ChemoR chemoresistance, MAC myeloablative conditioning, RIC reduced intensity conditioning, NMAC non-myeloablative conditioning, FL follicular lymphoma, tFL transformed follicular lymphoma, DLBCL diffuse large B cell lymphoma, BL Burkitt lymphoma, SLL small lymphocytic lymphoma, MZL marginal zone lymphoma, EBMT European Bone Marrow Transplant Registry, CBMTG Canadian Blood and Marrow Transplant Group, CIBMTR Center for International Blood and Marrow Transplant Research, MRD matched related donor, URD unrelated donor.