Introduction

It has been demonstrated that high-dose therapy with autologous stem cell rescue can successfully salvage many patients with relapsed or refractory Hodgkin's lymphoma (HL), with two randomized studies demonstrating the superiority of such treatment over conventional dose salvage chemotherapy.^{1, 2}

In contrast, results with myeloablative allogeneic transplantation in HL have been disappointing, with high non-relapse mortality (NRM). No randomized studies comparing autologous transplantation and allogeneic transplantation exist, but a retrospective European Group for Blood and Marrow Transplantation registry study reported improved outcome post-autograft, and this has become the consolidation of choice in relapsed or refractory disease.³

Despite the success of autologous transplantation, there remains a cohort of patients whose disease progresses/relapses following transplant, and the outcome in this group is extremely poor, with a median survival post-relapse of less than 1 year.^{4, 5}

With the advent of reduced-intensity conditioning regimens, there has been renewed interest in allogeneic transplantation in poor-risk patients with HL.^{6, 7} We previously reported a series of reduced-intensity allogeneic transplant (RIT), where a proportion of patients had failed a previous autograft. This showed encouraging estimated overall survival (OS) of 56% at 4 years with 2-year NRM of 16% .⁸

The current study aims to compare the outcomes for those patients with HL who relapsed following autograft and received RIT, and a historical control group of patients with HL who relapsed following autograft and received standard treatment of chemotherapyradiotherapy.

Materials and methods

Patients

Data were collected on all patients regarding date of diagnosis, age, histological subtype, lines of therapy pre-autograft, date of autograft, time to relapse following autograft, treatment following relapse, time to and cause of death.

For the RIT group, data were also collected on donor type, stem cell source, disease status prior to allogeneic transplantation, occurrence of acute and chronic GVHD, chimerism analysis, time to relapse, dose of and indication for donor lymphocyte infusions (DLI), response to DLI, time to and cause of death.

Control group

Patients who had relapsed following an autograft for relapsed/refractory HL performed at University College London Hospital between 1990 and 1996 were identified from a database of all autografts performed. This centre performed 60% of the allogeneic transplants in the present series. All patients had received conditioning with BEAM, as previously described¹ before the return of bone marrow or peripheral blood stem cells. Case notes of relapsing patients were then retrospectively analysed.

Treatment following relapse was with conventional salvage therapy using chemotherapyradiotherapy, no patients receiving RIT. Patients relapsing following autograft were included only if they responded to salvage therapy and lived for at least 1 year following relapse. In this way, it is intended that only those patients who would have been eligible for RIT had this been available were included. By these methods, 34 suitable patients were identified, representing 44% of the total relapsing. Two patients relapsed at a time when the RIT programme had just commenced and declined the offer of this treatment.

RIT group

A total of 38 patients treated with RIT following failure of prior autologous transplant were entered into a prospective phase II trial conducted at seven centres in the United Kingdom between 1998 and 2004. All patients gave signed informed consent. Allogeneic transplantation was performed at a median of 11 months from the time of relapse post-autograft.

Median follow-up for surviving patients was 49 months (10–88), with 16 months of additional follow-up from the previous report of Peggs et al.⁸

Donors were HLA-matched related in 24 of 38 (63% ), and unrelated in 14 of 38 (37% ), of whom 3 were HLA-mismatched at up to two loci. A total of 34% patients (13 of 38) were in chemorefractory relapse at the time of RIT.

Conditioning was with fludarabine 150 mg/m², alemtuzumab 60–100 mg and melphalan 140 mg/m² before reinfusion of unmanipulated bone marrow (7 of 38; 18% ) or granulocyte colony-stimulating factor-mobilized peripheral blood stem cells (31 of 38; 82% ), as previously described.⁸ Prophylaxis for GVHD was with cyclosporin at 3 mg per kg per day from day 1, and was reduced from 3 months after transplant, in the absence of significant GVHD.

Supportive care was with anti-fungal prophylaxis and treatment according to institutional guidelines, and all patients also received prophylaxis against Pneumocystis carinii pneumonia. Patients were screened regularly post transplant for the presence of CMV infection by PCR detection of viral DNA in peripheral blood, or by the detection of viral pp65 antigenemia. Pre-emptive anti-viral therapy was given according to institutional guidelines.

Chimerism analysis was performed by PCR assay of informative minisatellite regions from 3 months post transplant, at three monthly intervals for at least the first year. Disease re-staging using computed tomography (CT), positron-emission tomography (PET) or combined modality imaging (CT-PET) was also performed at regular intervals from 3 months, the modality chosen varying according to local policy. Although regular assays continued, the interval of re-staging and repeat chimerism analysis increased beyond the first year, and were partially dictated by the findings of previous tests and the monitoring of immune intervention with DLI. Patients were eligible for DLI for residual, relapsed or progressive disease, or mixed chimerism from 6 months post transplant, in the absence of active GVHD, and DLI were administered in escalating doses every 3 months if no response was achieved and no GVHD induced. The starting dose varied according to donor type and indication, but a minimum of 1 10⁶ per kg recipient weight CD3⁺ cells were infused as a first dose.

Statistics

Patient characteristics were compared using the unpaired Student's t-test. Actuarial curves were estimated for OS and current progression-free survival (cPFS) according to the Kaplan–Meier method, and univariate analyses were performed on subgroups using the logrank test. In the allograft group, patients with relapsed or progressive disease who responded with a partial or complete remission to DLI without subsequent progression were censored at the last follow-up date in the analysis of cPFS. NRM was defined as death post transplant from all causes other than relapse and was estimated according to the cumulative incidence method, with relapse defined as a competing risk.

Results

Patient characteristics

There were no statistically significant differences in age, disease subtype and lines of prior therapy pre-autograft between the two groups (Table 1). Median time from diagnosis to autograft was 18 months (7–139) for the RIT group and 20 months (4–185) for the controls (P=0.44), and from autograft to relapse was slightly longer for the RIT group at 13 months (2–56) compared to 10 months (3–40) (P=0.04).

Table 1 - Patient characteristics of the two groups.

Full table

Survival

Overall survival from both diagnosis and autograft was significantly better in the RIT group, with actuarial survival at 10 years from diagnosis of 48% compared to 15% (P=0.001), and survival at 5 years from autograft of 65% compared to 15% (P0.0001) (Figure 1). Of the group treated with chemotherapyRT alone, only 1 of 34 patients is currently alive and in remission, with 1 further patient alive with progressive disease. Median survival from relapse for this group was 23 months.

Figure 1.

Overall survival from autograft for the reduced-intensity transplantation (RIT) group (n=38; solid line) and the control group (n=34; dotted line). Estimated overall survival (OS) for the RIT group at 5 years is 65% and the control group 15% , P0.0001.

Full figure and legend (12K)

For the group receiving RIT, actuarial survival from the time of allograft was 68% at 1 year and 51% at 5 years (Figure 2a), with 12 patients dying of relapsed lymphoma, 5 of whom had chemorefractory disease at the time of allograft. Current PFS for the whole RIT cohort was 58% at 1 year and 34% at 5 years (Figure 2b).

Figure 2.

(a) Overall survival (OS) from the time of allogeneic transplant for the reduced-intensity transplantation (RIT) cohort (n=38). Estimated OS at 1 year is 68% and at 5 years 51% . (b) Current progression-free survival (cPFS) from the time of allogeneic transplant for the RIT group (n=38). cPFS is 58% at 1 year, 34% at 5 years.

Full figure and legend (18K)

Outcome in those patients who were at least partially responsive to chemotherapy (25 of 38; 66% ) was particularly encouraging, with OS at 5 years of 58% compared to 36% in the chemorefractory group (13 of 38; 34% ) (P=0.108) and cPFS at 5 years of 42% (Figure 3) compared to 18% in the chemorefractory group (P=0.075).

Figure 3.

Current progression-free survival from the time of allogeneic transplant for recipients of reduced-intensity transplantation (RIT) who were chemosensitive prior to transplant (n=25; solid line) of 42% at 5 years, and for recipients of RIT who were chemorefractory prior to transplant (n=13; dotted line) of 18% at 5 years.

Full figure and legend (10K)

RIT: non-relapse mortality and GVHD

Eight patients died of non-relapse causes, giving estimated NRM for the RIT group of 13% at 1 year and 19% at 5 years from transplant (Figure 4). For the 24 recipients with HLA-matched sibling donors, 5-year NRM was only 9% . The causes of death were GVHD in two (one of whom had received DLI for treatment of mixed chimerism), and one case each of systemic adenoviral infection, CMV pneumonitis, bacterial pneumonia, multi-organ failure secondary to sepsis early post transplant, pneumonitis of unknown origin and secondary malignancy.

Figure 4.

Non-relapse mortality (NRM) following reduced-intensity transplantation (RIT, n=38). NRM is 13% at 1 year, 19% at 5 years.

Full figure and legend (9K)

Before the administration of DLI, 23 patients (60% ) developed no acute GVHD, with grade I in 9 patients (24% ). Six patients (16% ) developed grade II–IV (grade II in four and grade III in two). Five patients developed either extensive (three) or limited (two) chronic GVHD before the administration of DLI. None of the three patients with extensive chronic GVHD had had acute GVHD and the two with limited chronic GVHD had experienced grade I acute GVHD only.

RIT: DLI responses for relapse

Disease relapse or progression occurred in 21 patients following RIT, 15 of whom received DLI. Three of the six patients who did not receive DLI developed rapidly progressive recurrence very early post transplant (within 3 months for two, both of whom were chemorefractory prior to transplant), and the remaining three had already developed GVHD, so DLI were precluded.

For the other 15 patients, donor lymphocytes were given with a dose range of 1 10⁶ to 3 10⁸ per kg CD3⁺ cells, escalating threefold each time to disease response or limiting GHVD. Five of the patients had disease relapse that was chemorefractory prior to transplant. Responses occurred in eight of the fifteen patients given DLI, only one of whom had received prior debulking chemotherapy (Table 2). These responses were strongly associated with the development of significant GVHD, six of the eight having either at least grade III acute or extensive chronic GVHD. Of the seven non-responders, no extensive chronic GVHD was seen, and grade II (two patients) was the maximum acute GVHD.

Table 2 - Data on patients who responded to donor lymphocyte infusions.

Full table

Of the eight responding patients, one died of GVHD, two subsequently progressed (including the patient who had received prior chemotherapy), and the remaining five are in ongoing CR, at a median follow-up from DLI of 45 months (28–55).

RIT: DLI responses for chimerism

Donor lymphocytes were given to five patients for persistent mixed chimerism at a starting dose of 1 10⁶ per kg CD3⁺ cells. Three patients converted to full donor haemopoeisis in the absence of GVHD, two of whom required serial escalation to 3 10⁶ and 1 10⁷ per kg respectively. One patient died of grade IV acute GVHD following DLI at a dose of 1 10⁶ per kg, and one patient remains with mixed chimerism, despite the development of limited chronic cutaneous GVHD.

Discussion

These results provide further supportive evidence for the efficacy of RIT in HL. OS both from time of diagnosis and time of autograft were significantly improved following allogeneic transplant, when compared to a historical control group treated with chemotherapyRT.

While accepting the limitations of non-randomized retrospective studies of this nature, this is an area where no randomized prospective studies exist, and when attempting to assess the efficacy of RIT it is instructive to view the outcome in the context of a comparator group. The comparator group was selected as far as possible to reflect a cohort who would have been considered for and would have survived to such a procedure, had this been available.

There were differences in therapy given, in that more intensive multi-agent salvage (for example ESHAP) was prevalent in the RIT cohort, as the aim here was to get to a further definitive therapy, whereas, for the control cohort for whom RIT was not an option, treatment was less frequently with such therapy. While it might be argued that less intensive therapy could have meant the patients survived for a shorter period of time than might otherwise have been possible, it could also be argued that a better risk group are being selected here, in that only those surviving to >12 months and responding to relatively less intensive treatment were eligible for inclusion in the control group. Had we given more intensive therapy, it could be argued that an additional cohort of patients who were considered too poor risk to be included in the controls in the current analysis would have been eligible for inclusion.

The control group fared extremely badly, with 15% survival at 5 years from autograft and only one patient alive in remission at the time of reporting, despite the patients being selected only if they survived at least 12 months beyond the time of relapse. As the median survival following relapse post-autograft has been consistently reported as less than 12 months in HL,^{4, 5} this current group clearly represents a subset with a favourable prognosis relative to relapsed patients in general, and, even allowing for a degree of attrition preventing all of these patients proceeding to RIT, it is rational to argue that those who would have gone on to allogeneic transplant are contained within this group. As only one is currently alive and disease-free, it is not possible to construct a subgroup whose survival would have been comparable to the RIT cohort.

In contrast, OS of the allogeneic transplant group showed a substantial improvement, 51% at 5 years from RIT, and 65% at 5 years from autograft. It should also be noted that the RIT group contained a significant number of patients (34% ) who were refractory to the salvage therapy given for post-autograft relapse. These are particularly poor-risk patients and equivalent patients were excluded from the control group.

It is also relevant to note that patients relapsing in the pre-RIT era were treated without curative intent, thus excluding some poorer-risk patients from the control cohort, who would have reached RIT had aggressive systemic therapy been invoked with a view to further high-dose therapy, as is current practice.

Previous experience of allogeneic transplantation in this disease has been disappointing. Myeloablative transplantation has been performed in patients with high-risk disease, but event-free survival at up to 5 years has been only 15–22% .^{3, 9, 10, 11} Despite the use of HLA-matched sibling donors, NRM was substantial (48–61% ), which accounted in large part for the poor outcomes. It is also relevant that these results were obtained mainly in the absence of prior high-dose therapy, which would be expected to add to toxicity. Several groups have reported studies of myeloablative transplantation in the presence of prior autograft in haematological malignancy and the NRM has been formidable, at consistently >50% .^{12, 13}

The introduction of reduced-intensity conditioning has led, therefore, to renewed interest in allogeneic transplantation in this disease, and although the literature of RIT for HL remains limited, several groups have reported interesting results although with generally short follow-up of less than 18 months.^{6, 7, 14} In the current report, a cohort of 38 patients is presented with a median follow-up of 49 months. Estimated OS is 51% at 5 years and cPFS 34% , with cPFS of 42% if in chemosensitive relapse at the time of transplant, suggesting the early promising results might translate into a favourable long-term outcome. Current PFS in the chemorefractory group was 18% at 5 years, with OS of 36% . The role of RIT in chemorefractory disease is unclear, although there is some indication that a subset of even these poor-risk patients may benefit. The outstanding questions are whether benefit truly exists in any group, and if so, whether such benefit is confined to those patients chemorefractory but non-progressive, as opposed to those transplanted with progressive disease. There is also a need to identify how benefit from RIT alters with increasing numbers of lines of therapy to guide how aggressively to pursue response to salvage pre-transplant, and finally, the role of RIT in primary refractory patients remains to be established.

Regarding procedural toxicity, NRM in the RIT reports to date is consistently lower than that ascribed to myeloablative conditioning regimens, despite the majority of these patients having undergone prior autologous transplant. Analyses of the potential negative impact of such high-dose therapy in RIT are scarce, although one reduced-intensity regimen has been reported as having markedly enhanced toxicity when performed in this setting in patients with NHL,¹⁵ and one other study has supported a significant increase in NRM post-RIT in the presence of a prior autograft, at least in patients >55 years of age.¹⁶ The NRM in the RIT group in this study compares relatively well with published data, at 19% at 5 years, with no events beyond the first 18 months. In addition, procedural toxicity in the cohort of 24 patients with HLA-matched sibling donors was particularly low, at 9% at 5 years.

Regarding a graft-versus-lymphoma effect, the existence of a clinically relevant and durable allogeneic response is a prerequisite if this approach is to prove curative. Early studies of myeloablative transplantation provided some evidence suggesting the existence of such an effect, and encouraging data are now emerging from RIT studies, with small numbers of patients receiving DLI for disease relapse following transplant.^{17, 18} The durability of the few responses seen remains unclear, with the largest of these series reporting responses in four of nine patients with median follow-up of only 7 months.¹⁸ We previously reported responses to DLI in eight of fifteen patients included in the current analysis (seven complete, one partial). The durability of such responses is an important issue, particularly in view of the association of remission with the development of clinically significant GVHD, treatment of which with steroids may provide a temporary anti-tumour effect. With extended follow-up, five have ongoing complete responses, none of whom had received chemotherapy prior to DLI, with a median time from DLI of 45 months (range 28–55). Given the extremely heavy pre-treatment of the RIT cohort, these are very encouraging results.

In conclusion, we report updated outcomes in patients receiving RIT following failure of autograft for HL, and put this in context with the outcome using non-transplant therapy in a comparator group. Given the provocative results with extended follow-up of over 4 years, we would propose that RIT should now be investigated earlier in therapy, for example in a prospective randomization against autologous transplantation in relapsed patients with poor-risk features who would be predicted to do poorly with autograft alone.

References

1. Linch DC, Winfield D, Goldstone AH, Moir D, Hancock B, McMillan A et al. Dose intensification with autologous bone-marrow transplantation in relapsed and resistant Hodgkin's disease: results of a BNLI randomised trial. Lancet 1993; 341: 1051–1054. | Article | PubMed | ISI | ChemPort |
2. Schmitz N, Pfistner B, Sextro M, Sieber M, Carella AM, Haenel M et al. Aggressive conventional chemotherapy compared with high-dose chemotherapy with autologous haemopoietic stem-cell transplantation for relapsed chemosensitive Hodgkin's disease: a randomised trial. Lancet 2002; 359: 2065–2071. | Article | PubMed | ISI | ChemPort |
3. Milpied N, Fielding AK, Pearce RM, Ernst P, Goldstone AH. Allogeneic bone marrow transplant is not better than autologous transplant for patients with relapsed Hodgkin's disease. European Group for Blood and Bone Marrow Transplantation. J Clin Oncol 1996; 14: 1291–1296. | PubMed | ISI | ChemPort |
4. Vose JM, Bierman PJ, Anderson JR, Kessinger A, Pierson J, Nelson J et al. Progressive disease after high-dose therapy and autologous transplantation for lymphoid malignancy: clinical course and patient follow-up. Blood 1992; 80: 2142–2148. | PubMed | ISI | ChemPort |
5. Varterasian M, Ratanatharathorn V, Uberti JP, Karanes C, Abella E, Momin F et al. Clinical course and outcome of patients with Hodgkin's disease who progress after autologous transplantation. Leuk Lymphoma 1995; 20: 59–65. | Article | PubMed | ChemPort |
6. Carella AM, Beltrami G, Carella Jr M, Corsetti MT, Scalzulli RP, Greco M. Immunosuppressive non-myeloablative allografting as salvage therapy in advanced Hodgkin's disease. Haematologica 2001; 86: 1121–1123. | PubMed | ISI | ChemPort |
7. Anderlini P, Saliba R, Acholonu S, Okoroji GJ, Donato M, Giralt S et al. Reduced-intensity allogeneic stem cell transplantation in relapsed and refractory Hodgkin's disease: low transplant-related mortality and impact of intensity of conditioning regimen. Bone Marrow Transplant 2005; 35: 943–951. | Article | PubMed | ISI | ChemPort |
8. Peggs KS, Hunter A, Chopra R, Parker A, Mahendra P, Milligan D et al. Clinical evidence of a graft-versus-Hodgkin's-lymphoma effect after reduced-intensity allogeneic transplantation. Lancet 2005; 365: 1934–1941. | Article | PubMed | ISI |
9. Gajewski JL, Phillips GL, Sobocinski KA, Armitage JO, Gale RP, Champlin RE et al. Bone marrow transplants from HLA-identical siblings in advanced Hodgkin's disease. J Clin Oncol 1996; 14: 572–578. | PubMed | ISI | ChemPort |
10. Akpek G, Ambinder RF, Piantadosi S, Abrams RA, Brodsky RA, Vogelsang GB et al. Long-term results of blood and marrow transplantation for Hodgkin's lymphoma. J Clin Oncol 2001; 19: 4314–4321. | PubMed | ISI | ChemPort |
11. Anderson JE, Litzow MR, Appelbaum FR, Schoch G, Fisher LD, Buckner CD et al. Allogeneic, syngeneic, and autologous marrow transplantation for Hodgkin's disease: the 21-year Seattle experience. J Clin Oncol 1993; 11: 2342–2350. | PubMed | ISI | ChemPort |
12. Tsai T, Goodman S, Saez R, Schiller G, Adkins D, Callander N et al. Allogeneic bone marrow transplantation in patients who relapse after autologous transplantation. Bone Marrow Transplant 1997; 20: 859–863. | Article | PubMed | ISI | ChemPort |
13. di Grazia C, Raiola AM, Van Lint MT, Lamparelli T, Gualandi F, Berisso G et al. Conventional hematopoietic stem cell transplants from identical or alternative donors are feasible in recipients relapsing after an autograft. Haematologica 2001; 86: 646–651. | PubMed | ChemPort |
14. Robinson SP, Goldstone AH, Mackinnon S, Carella A, Russell N, de Elvira CR et al. Chemoresistant or aggressive lymphoma predicts for a poor outcome following reduced-intensity allogeneic progenitor cell transplantation: an analysis from the Lymphoma Working Party of the European Group for Blood and Bone Marrow Transplantation. Blood 2002; 100: 4310–4316. | Article | PubMed | ISI | ChemPort |
15. Faulkner RD, Craddock C, Byrne JL, Mahendra P, Haynes AP, Prentice HG et al. BEAM-alemtuzumab reduced-intensity allogeneic stem cell transplantation for lymphoproliferative diseases: GVHD, toxicity, and survival in 65 patients. Blood 2004; 103: 428–434. | Article | PubMed | ISI | ChemPort |
16. Corradini P, Zallio F, Mariotti J, Farina L, Bregni M, Valagussa P et al. Effect of age and previous autologous transplantation on nonrelapse mortality and survival in patients treated with reduced-intensity conditioning and allografting for advanced hematologic malignancies. J Clin Oncol 2005; 23: 6690–6698. | Article | PubMed | ISI | ChemPort |
17. Porter DL, Stadtmauer EA, Lazarus HM. 'GVHD': graft-versus-host disease or graft-versus-Hodgkin's disease? An old acronym with new meaning. Bone Marrow Transplant 2003; 31: 739–746. | Article | PubMed | ChemPort |
18. Anderlini P, Acholonu SA, Okoroji GJ, Andersson BS, Couriel DR, De Lima MJ et al. Donor leukocyte infusions in relapsed Hodgkin's lymphoma following allogeneic stem cell transplantation: CD3+ cell dose, GVHD and disease response. Bone Marrow Transplant 2004; 34: 511–514. | Article | PubMed | ChemPort |