Prognostic factors affecting long-term outcome after stem cell transplantation in Hodgkin s lymphoma autografted after a first relapse

Transcription

1 Original article Annals of Oncology 16: , 2005 doi: /annonc/mdi119 Published online 28 February 2005 Prognostic factors affecting long-term outcome after stem cell transplantation in Hodgkin s lymphoma autografted after a first relapse A. Sureda*, M. Constans, A. Iriondo, R. Arranz, M. D. Caballero, M. J. Vidal, J. Petit, A. López, J. J. Lahuerta, E. Carreras, J. García-Conde, J. García-Laraña, R. Cabrera, I. Jarque, D. Carrera, J. C. García-Ruiz, M. J. Pascual, J. Rifón, J. M. Moraleda, K. Pérez-Equiza, C. Albó, J. Díaz-Mediavilla, A. Torres, P. Torres, J. Besalduch, J. Marín, M. V. Mateos, J. M. Fernández-Rañada, J. Sierra & E. Conde for The Grupo Español de Linfomas/Trasplante Autólogo de Médula Osea (GEL/TAMO) Cooperative Group Received 28 July 2004; revised 8 October 2004; accepted 23 November 2004 Introduction Purpose: To analyse outcome and prognostic factors for overall survival (OS) and time to treatment failure (TTF) in 357 patients with Hodgkin s lymphoma (HL) undergoing an autologous stem cell transplantation (ASCT) after a first relapse and reported to the The Grupo Español de Linfomas/ Trasplante Autólogo de Médula Osea (GEL/TAMO) Cooperative Group. Methods: Two hundred and twenty males and 137 females with a median age of 29 years were autografted in second remission (n = 181), first sensitive relapse (n = 148) and first resistant relapse (n = 28). Results: Five-year actuarial TTF and OS were of 49% ± 3% and 57% ± 3%. Advanced stage at diagnosis, complementary radiotherapy before ASCT, a short first complete response (CR) and detectable disease at ASCT adversely influenced TTF. Year of transplant <_ 1995, bulky disease at diagnosis, a short first CR, detectable disease at ASCT and >_ 1 extranodal areas involved at ASCT were adverse factors for OS. Conclusions: ASCT constitutes a therapeutic option for HL patients after a first relapse. Promising results are observed in patients with low tumour burden at diagnosis, autografted after a long CR and without detectable disease at ASCT. Innovative approaches should be pursued for patients with risk factors at relapse. Key words: autologous stem cell transplantation, first relapse, Hodgkin s lymphoma Most of the patients suffering from Hodgkin s lymphoma (HL) can be successfully treated with radiotherapy (RT) and/or conventional-dose chemotherapy, with 70% of them being alive 10 years after diagnosis [1, 2]. Nevertheless, a substantial proportion of patients do not adequately respond to conventional therapy, and ultimately die of the disease. In this subset of patients, the use of conventional-dose, second-line chemotherapy induces low remission rates, with long-term disease-free survival (DFS) in no more than 10% of patients [3, 4]. High-dose therapy with autologous stem cell transplantation (ASCT) has been extensively tested in patients with relapsed *Correspondence to: Dr A. Sureda, Clinical Hematology Division, Hospital de la Santa Creu i Sant Pau, Antoni Maria i Claret, 167, Barcelona, Spain. Tel: ; Fax: ; asureda@hsp.santpau.es and refractory HL during recent years [5 14]. Of particular interest is the analysis of the results of high-dose therapy in patients with HL at the time of first relapse after chemotherapy [8, 11, 12, 15]. Single institution studies have shown better outcome after ASCT in this group of patients when compared to historical controls receiving conventional treatment [16, 17]. Indeed, two prospective randomised analyses, which include patients with HL in first relapse, also indicate a significant advantage of intensification versus conventional salvage regimens, in terms of long-term DFS [18, 19], especially in patients relapsing after a short first complete remission (CR) [19]. Retrospective analyses performed by several groups indicate that the extent of prior chemotherapy [8, 11], disease status at relapse [8, 15, 20 23], chemosensitivity to salvage chemotherapy [22, 23] and duration of initial CR [15, 20] emerge as significant prognostic factors for the long-term outcome of intensive therapy in this setting. The purpose of the present analysis was to determine independent prognostic factors correlated with the long-term outcome of q 2005 European Society for Medical Oncology

2 626 ASCT performed after a first relapse in a cohort of 357 HL patients reported to The Grupo Español de Linfomas/Trasplante Autólogo de Médula Osea (GEL/TAMO) Cooperative Group prospective observational registry during a period of 15 years. Patients and methods From January 1984 to June 2002, all patients with HL who received an ASCT for recurrent disease after a first relapse (n = 357) were selected from a study group of 755 patients with HL autografted and consecutively reported to the GEL/TAMO Cooperative Group. Reported data were centrally reviewed to detect inconsistencies. All reporting physicians were contacted to provide additional information on patients characteristics at presentation when required. Follow-up was updated on December 2003, when all living patients had been followed for at least 1 year after ASCT. Eligibility criteria Common eligibility criteria for ASCT in all institutions were age <_ 65 years, left ventricular ejection fraction >50%, forced expiratory volume in 1 s (FEV1) >50%, diffusion capacity of the lung for carbon monoxide (DLCO) >50% predicted, and absence of major organ dysfunction of cause different to HL. All patients gave written informed consent before undergoing ASCT. Study definitions Patients were required to have progression of HL proven by biopsy or unequivocal radiological progression after a CR induced by primary chemotherapy. Relapse was distinguished from primary progressive disease by a period of at least 3 months in remission after completion of first-line treatment. A sensitive relapse was defined as at least a 50% reduction in the bidimensional measurements of the disease with the use of conventional salvage chemotherapy. A resistant relapse was defined as <50% reduction in the size of the tumour with the use of conventional salvage chemotherapy. Early relapse was defined as CR after first-line therapy lasting 3 12 months. Late relapse was defined as CR lasting >12 months after completion of induction treatment. Patients were staged according to the Ann Arbor system [24]. Patients were clinically staged at the time of ASCT, on day +90 after ASCT, every 6 months for the first 2 years, then yearly or as clinically indicated. Patients who survived more than 90 days after ASCT without evidence of tumour, by clinical and radiological evaluation including computed tomography scan, were classified as CR. Patients with small residual radiographic abnormalities, which did not progress for 6 months after transplant, were also classified as being in CR. Partial remission (PR) was defined as a >_ 50% reduction of pre-transplant measurable disease, for at least 1 month. Patients achieving <50% tumour reduction after ASCT were considered to be non-responders (NR). Patients Main characteristics of the 357 patients at diagnosis and at ASCT are shown in Tables 1 and 2, respectively. Two hundred and twenty (62%) patients were male and 137 were female (38%). The median age at the time of ASCT was 29 years (range 8 65 years). At initial presentation, 54% of the patients had stage III IV disease. Other adverse features were B symptoms (52%) and bulky disease only in 27% of the whole series. Initial treatment had varied according to the initial HL stage, the date of diagnosis, and the protocol used in each hospital: most patients received either adriamycin-containing regimens (n = 250, 70%) or Table 1. Clinical characteristics of the 357 patients at diagnosis Characteristic n (%) Sex Male 220 (62) Female 137 (38) Age, median (range) in years 29 (8 66) Histology Nodular sclerosis 223 (62) Mixed cellularity 99 (28) Lymphocyte predominance 17 (5) Lymphocyte depletion 8 (2) Not specified 10 (3) Ann Arbor stage I II 162 (46) III IV 195 (54) B symptoms 186 (52) BM involvement 40 (11) Bulky disease (>10 cm) 96 (27) No. of extranodal areas involved (91) >_2 34 (9) First-line therapy MOPP-like regimens 107 (30) Adriamycin-containing regimens 250 (70) Complementary RT 139 (39) BM, bone marrow; MOPP, mechlorethamine, vincristine, procarbazine and prednisone; RT, radiotherapy. mechlorethamine, vincristine, procarbazine and prednisone (MOPP)-like regimens (n = 107, 30%). Complementary RT was used in 139 of the 357 patients, according to active protocols in participating institutions. The median duration of the first CR had been 13 months (range months); of these, 155 patients had relapsed within 12 months of achieving CR (43%) and 169 thereafter. Data were not available in 33 additional patients. The time interval between diagnosis and the date of ASCT ranged from 9 to 224 months, with a median of 31.5 months. Disease status at ASCT appears in Table 2. One hundred and eightyone patients achieved a second CR after salvage chemotherapy (51% of the series) and the remaining 176 patients were autografted with visible disease: 134 with active chemosensitive first relapse, 28 (8%) with resistant first relapse and 14 (4%) proceeded to ASCT without receiving conventional salvage treatment at the time of their relapse (untreated first relapse). Source of autologous stem cells Two hundred and thirty-two patients (65%) were autografted using peripheral blood progenitor cells (PBPCs) as the source of hematopoietic stem cells and the remaining 125 (35%) were autografted using bone marrow (BM). Bone marrow was harvested under general anaesthesia and cryopreserved following standard guidelines.

3 627 Table 2. Characteristics of the patients at autologous transplantation Characteristic n (%) Diagnosis: ASCT, (median/range) in months 31.5 (9 224) Year of ASCT < (59) > (41) Status Second complete response 181 (51) Untreated relapse 14 (4) Sensitive relapse 134 (37) Refractory relapse 28 (8) Ann Arbor stage a I II 94 (52) III IV 84 (48) B symptoms a 15 (8) Bulky disease a 12 (7) No. of extranodal areas involved a (68) >_ 1 57 (32) [median 13 (3 209) (range)] in months ECOG performance status 0, (86) >_ 2 11 (3) Not reported 39 (11) Stem cell source BM 124 (35) PB 232 (65) Conditioning regimen Cy + TBI 26 (7) Chemotherapy 331 (93) CBV 162 (45) BEAM/BEAC 110/43 (31/12) Others 16 (5) a For patients not in complete response at ASCT, n = 176. ASCT, Autologous stem cell transplantation; ECOG, Eastern Cooperative Oncology Group classification; Cy, cyclophosphamide; TBI, total body irradiation; BM, bone marrow; PB, peripheral blood. CBV, cyclophosphamide; BCNU, etoposide; BEAM, BCNU, etoposide, cytarabine, melphalan; BEAC, BCNU, etoposide, cytarabine, cyclophosphamide. High-dose therapy and transplantation procedures Details of high-dose therapy are given in Table 2. Regimens that included total body irradiation (TBI) were used in 7% of the patients (n = 26). The remaining 93% received chemotherapy-only high-dose regimens, the most frequent being CBV (n = 162, 45%) which consisted of cyclophosphamide ( g/m 2 i.v. for 4 days), etoposide ( mg/m 2 twice daily i.v. for 3 days) and BCNU ( mg/m 2 i.v. for 1 day). The BEAM protocol was administered to 110 patients (31%) [BCNU ( mg/m 2 i.v. for 1 day), etoposide ( mg/m 2 i.v. for 4 days), cytarabine ( mg/m 2 twice daily i.v. for 4 days) and melphalan (140 mg/m 2 i.v. for 1 day)] and the BEAC combination to 43 patients (12%), this regimen being the association of BCNU ( mg/m 2 i.v. for 1 day), etoposide ( mg/m 2 i.v. for 4 days), cytarabine (200 mg/m 2 twice daily i.v. for 4 days) and cyclophosphamide ( g/m 2 i.v. for 3 days) and 16 patients (5% of the series) received other combination chemotherapy protocols. A total of 186 patients (52%) received granulocyte colony-stimulating factor (G-CSF) or granulocyte-macrophage colony-stimulating factor (GM-CSF) until the absolute neutrophil count exceeded /l per 3 consecutive days. Statistical analysis Actuarial curves were performed according to the Kaplan Meier method [25]. Overall survival (OS) was calculated in months from the date of autologous stem cell reinfusion to the date of death from any cause. Time to treatment failure (TTF) was measured in months from the date of transplantation to the time of failure or death from any cause following previously described criteria for non-hodgkin s lymphomas (NHL) [26]. Overall non-relapse mortality (NRM) was defined as death from any cause other than HL. Comparison of the survival curves in univariate analysis was performed using the log rank test [27]. Analysis of prognostic factors influencing both CR and NRM rates was performed by Fisher s exact test and logistic regression analysis. Comparison of continuous variables was performed by Mann Whitney s U test and linear regression analysis. Multivariate analysis was performed using a forward stepwise Cox proportional-hazards model. The prognostic factors analysed for both TTF and OS were: age at transplantation, year of transplant according to the median (<_ 1995 and >1995), complementary RT, first-line therapy (MOPP-like regimens versus other protocols), number of treatment lines (one versus two or more), duration of first CR (<_12 months versus >12 months), conditioning regimen (TBI versus chemotherapy alone), source of stem cells (BM versus PB), Ann Arbor stage (early versus advanced), B symptoms, extranodal involvement, BM involvement, bulky disease and Eastern Cooperative Oncology Group (ECOG) performance status (0 1 versus >_ 2). The last six characteristics were evaluated at diagnosis and at ASCT. All P values reported are two-sided and statistical significance is defined as a P <0.05. The statistical analyses were computed by means of the SPSS statistical software. Results Hematological recovery and transplant-related complications Hematological recovery. All 357 patients became cytopenic after conditioning therapy. One patient died on day +13 with no evidence of engraftment. For the remaining 356, the median time to recover a neutrophil count > /l was 12 days (range 4 57 days) and to achieve a self-sustained platelet count > /l was 14 days (range days). Hematological recovery was significantly faster in patients autografted with PB progenitors cells compared to BM for both granulocytes and platelets [11 (4 32) days versus 18 (5 57) days, P <0.0001, and 13 (5 198) days versus 21 (9 91) days, P = 0.006, respectively]. NRM. Thirty-seven patients (10%) died after ASCT due to procedure-related complications: 26 patients during the first

4 628 3 months after ASCT (early NRM) and the remaining 11 patients later (late NRM). Overall NRM was 8.0% [95% confidence interval (CI): ] at 1 year. Both the use of TBIcontaining protocols as conditioning regimen [relative risk (RR) 6.09, 95% CI , P = ] and year of ASCT <_1995 (RR 3.44, 95% CI , P = 0.006) were independent adverse prognostic factors for NRM in the multivariate analysis (Table 3). Patients autografted <_ 1995 had some differential characteristics with respect to those autografted after this year that can account for the significant difference in NRM [more frequent use of MOPP-like protocols (34% versus 16%, P = 0.001), of TBI-containing protocols as conditioning regimen (10% versus 3%, P = and of BM as the source of hematopoietic stem cells (57% versus 3%, P = )]. When considering only those patients autografted after 1995 with chemotherapy-containing protocols (n = 140) whose overall NRM was 3% at 12 months (95% CI ), the only significant adverse prognostic factor for NRM was refractoriness to salvage chemotherapy before ASCT (RR 15.15, 95% CI , P = 0.011). The causes of death were infectious episodes in 13 patients (bacterial in three, fungal in seven, and CMV pneumonitis in three patients), interstitial pneumonitis/adult respiratory distress syndrome (IP/ARDS) in nine patients, secondary neoplasias in six patients (one patient died 169 months after ASCT), multiorgan failure in four patients, hemorrhage in two patients, venooclusive disease in two patients and cardiac toxicity in Table 3. Multivariate analysis for non-relapse mortality and complete remission Prognostic factor n No. of deaths (%) Non-relapse mortality Year of ASCT RR CI (95%) P value < (15) 3.44 ( ) > (3) Conditioning regimen Chemotherapy (9) TBI-containing protocols 26 7 (27) 6.09 ( ) Prognostic factor n RR CI (95%) P value Complete remission rate Disease status at ASCT Sensitive disease 122 Refractory disease ( ) Extranodal areas involved at ASCT >_ ( ) >12 months 69 <_ 12 months ( ) RR, relative risk; CI, confidence interval; ASCT, autologous stem cell transplantation; TBI, total body irradiation. one patient. Three other additional patients developed and died from a secondary neoplasia after having relapsed from their underlying disease. Response to ASCT and survival Twenty-six patients (7%) died within the first 3 months after ASCT due to NRM and were not evaluable for response. At the same time point, 292 (82%) patients were considered as complete responders, 171 of them being in continuous CR. Nineteen patients (5%) were in PR and 20 (6%) did not respond or progressed after ASCT. Multivariate analysis indicated that refractory relapse (RR 3.52, 95% CI , P = 0.015), >_ 1 extranodal areas involvement at ASCT (RR 2.68, 95% CI , P = 0.017) and a short first CR (RR 2.43, 95% CI , P = 0.022) were significantly associated with a lower CR rate for patients not being in CR before ASCT (Table 3). The actuarial TTF at 5 years was 49% (95% CI 46 52%) for the entire group of patients with an OS of 57% (95% CI 54 60%) at the same time point (Figures 1 and 2). Median follow-up of the surviving patients was 39 months (range ). Prognostic factors influencing TTF in the univariate analysis are shown in Table 4. On multivariate analysis (Table 4), advanced stage at diagnosis (RR 1.6, 95% CI , P = 0.007), complementary RT (RR 1.9, 95% CI , P = ), first CR <_12 months (RR 1.6, 95% CI , P = 0.005) and refractoriness to salvage chemotherapy (RR 4.5, 95% CI , P = ) significantly shortened TTF. The combination of these four adverse prognostic factors was used to construct a prognostic factor model that distinguishes patients with different degrees of prognosis after ASCT ranging from good (0 1 adverse factor) to poor (>_ 2 factors) (Figure 3A). Chemosensitivity to prior CT was one of the most important prognostic factors for TTF (Figure 4) with figures ranging from 68% ± 4% at 5 years for patients autografted in second CR to 34% ± 5% and 11% ± 6% for patients autografted with sensitive and resistant disease, respectively. Prognostic factors influencing OS in the univariate analysis are shown in Table 5. On multivariate analysis (Table 5), ASCT performed after 1995, a first CR >12 months, absence of bulky disease at ASCT, <1 extranodal area involved at ASCT and sensitive disease at ASCT significantly improved survival after transplant. Again, the combination of these five prognostic factors allowed us to differentiate three groups of patients with significantly different 5-year OS according to the number of adverse features at relapse (Figure 3B). Chemosensitivity to the previous chemotherapy protocol administered was also one of the most important prognostic factors with an OS of 75% ± 4% at 5 years for patients autografted in second CR and of 43% ± 5% and 19% ± 7% at 5 years for those autografted with sensitive or resistant disease, respectively. Relapse after ASCT Of the 292 patients who were in CR 3 months after transplantation, 77 (26%) relapsed at a median time of 25 months

5 629 Figure 1. Time to treatment failure (TTF) of the whole series. ASCT, autologous stem cell transplantation. Figure 2. Overall survival (OS) of the whole series. ASCT, autologous stem cell transplantation. (range 4 174) after ASCT. Forty-two patients (66%) relapsed within the first year after ASCT. Of note, two patients relapsed more than 5 years after ASCT. Fifty-one of the relapsed patients died from progressive disease at a median time of 12 months (range 1 76) after relapse and the remaining 26 patients were alive at the time of the final analysis. Discussion At present, high dose chemo/radiotherapy followed by ASCT is the treatment of choice for patients with relapsed HL after first-line polychemotherapy. In several phase II studies, ASCT has been shown to produce between 30 and 65% long-term DFS in selected patients with refractory and relapsed HL [5 15]. Two randomised studies performed by the British Table 4. Univariate and multivariate analyses for time to treatment failure Prognostic factor n 5 year TTF P value Univariate analysis Stage at diagnosis I II % ± 4% III IV % ± 4% Complementary RT No % ± 4% 0.02 Yes % ± 5% Status at ASCT Second complete % ± 4% response Sensitive relapse % ± 5% Resistant relapse 28 11% ± 6% Stage at ASCT I II 93 39% ± 6% III IV 83 21% ± 5% Extranodal areas involved at ASCT % ± 3% >_ % ±6% ECOG at ASCT % ± 3% >_ % BM involvement at ASCT No % ± 3% 0.01 Yes 6 0% LDH at ASCT Normal % ± 3% High 27 35% ± 9% <_ 12 months % ± 4% >12 months % ± 4% n RR 95% CI P value Multivariate analysis Ann Arbor stage at diagnosis I II 162 III IV Complementary RT to first-line therapy No 218 Yes <_ 12 months >12 months 168 Disease status at ASCT Complete remission Sensitive relapse Resistant relapse TTF, time to treatment failure; RT, radiotherapy; ASCT, autologous stem cell transplantation; ECOG, Eastern Cooperative Oncology Group performance status; LDH, lactate dehydrogenase enzyme; RR, relative risk; CI, confidence interval.

6 630 Figure 3. Time to treatment failure (TTF) (A) and overall survival (OS) (B) according to the number of adverse prognostic factors analysed at relapse. ASCT, autologous stem cell transplantation. National Lymphoma Investigation (BNLI) [18] and the German Hodgkin s Study Group (GHSG)/European Bone Marrow Transplantation Group (EBMT) (HDR-1 protocol) [19] have shown improved outcome in patients with relapsed HL treated with high-dose chemotherapy, not only in patients in early relapse but also in patients in late relapse in terms of freedom from second failure [19]. Therefore, ASCT should be considered as standard treatment for patients relapsing after a first line chemotherapy, probably also including those with late relapse. Although all the above mentioned studies indicate the superiority of ASCT over conventional salvage chemotherapy, a proportion of patients will eventually develop recurrent disease after transplantation. In this sense, an effective assessment of prognostic factors, evaluated at the time of relapse, is required to guide the physician in the selection of the most appropriate therapeutic strategies and the development of Figure 4. Time to treatment failure (TTF) according to disease status at autologous stem cell transplantation (ASCT). experimental approaches in patients with a considered poor outcome after an ASCT. In the present analysis we have evaluated the prognostic factors influencing long-term outcome after an ASCT in a group of 357 patients autografted for HL relapsing after first line therapy with a median follow-up for surviving patients of 42 months (range ). Five-year OS and TTF were 57% (95% CI 54 60%) and 49% (95% CI 46 52%), respectively, for the entire group of patients, but these results significantly worsened in those patients presenting three or more risk factors analysed at relapse. In our study, two factors before relapse (advanced stage at diagnosis and the use of complementary RT to first line) and two factors at relapse (short first CR and the presence of visible disease after salvage chemotherapy) have become indicators of a poor outcome after transplantation. Patients with none or one risk factor had a 5-year TTF of 71%±4% after ASCT, whereas patients with three or more risk factors presented a 5-year TTF of only 18% ± 5%. The duration of the first CR, extranodal involvement at relapse and the present of visible disease at ASCT also significantly influenced long-term OS as well as the year of transplantation, an indirect indicator of the continuous improvement on transplant support in the last 10 years, and the presence of bulky disease at diagnosis. Patients with none or one of these risk factors had a 5-year OS of 82% ± 5% while patients with three or more risk factors had a 5-year OS of only 39% ± 5%. One of the most important prognostic factors for relapsed HL patients is the duration of the first CR. In 1992, The National Cancer Institute updated their experience with the long-term follow-up of patients who had relapsed after polychemotherapy [3]. Patients were divided into those whose initial CR had lasted for at least 12 months and those in whom the first CR had been shorter. The same applies for high-dose therapy and autologous rescue as has been previously indicated by the prospective randomised HDR-1

7 protocol [19]. In our analysis, the duration of the first CR also was one of the most important predictive factors both for TTF and OS: 5-year TTF was of 57 ± 4% for patients with a long first CR while it was only of 43% ± 4% for patients that relapsed earlier. In the same way, response to second-line chemotherapy (chemosensitive disease) has been used as the major selection criterion to proceed to ASCT. In our series, over 95% of the patients were tested for chemosensitivity and treated with salvage therapy before ASCT. Those patients entering a second CR had the best long-term prognosis with a 5-year TTF and 5-year OS of 68% ± 4% and 75% ± 4%, respectively. By contrast, patients with refractory relapse had Table 5. Univariate and multivariate analyses for overall survival Prognostic factor n 5 year OS P value Univariate analysis Year of ASCT <_ % ± 3% 0.07 > % ± 9% 0.06 ECOG at diagnosis % ± 3% 0.03 >_ % ± 10% Complementary RT No % ± 4% 0.06 RT % ± 5% Status at ASCT Second complete % ± 4% response Sensitive relapse % ± 5% Resistant relapse 28 19% ± 7% Stage at ASCT I II 94 53% ± 6% III IV 83 24% ± 6% Extranodal areas involved at ASCT % ± 3% >_ % ± 7% ECOG at ASCT % ± 3% >_ % BM involvement at ASCT No % ± 3% Yes 6 0% LDH at ASCT Normal % ± 3% High 27 46% ± 10% <_ 12 months % ± 5% >12 months % ± 4% Conditioning regimen Chemotherapy % ± 3% 0.01 TBI-containing regimens 26 33% ± 10% Table 5. (Continued) Prognostic factor n RR 95% CI P value Multivariate analysis Date of ASCT <_ > Bulky disease at diagnosis No 261 Yes <_ 12 months Prognostic factor n 5 year OS P value >12 months 169 Extranodal involvement at ASCT 0 1 extranodal 346 areas involved >_ 2 extranodal areas involved Disease status at ASCT Complete remission Sensitive relapse Resistant relapse OS, overall survival; ASCT, autologous stem cell transplantation; ECOG, Eastern Cooperative Oncology Group performance status; RT, radiotherapy; BM, bone marrow; LDH, lactate dehydrogenase enzyme; TBI, total body irradiation; RR, relative risk; CI, confidence interval. 631 a significantly poorer outcome with a 5-year TTF of only 11% ± 6% and a 5-year OS of 19% ± 7% (Figure 4). Although, it has recently been indicated that chemorefractory HL patients had survival indexes not different from those presented by sensitive patients [28], our multicentre results suggest that these patients should be offered a different therapeutical approach. Our analysis, although retrospective in nature, suggests the benefit of achieving a CR before transplantation. Although in most studies it seems clear that chemosensitivity before transplantation is one of the major favourable prognostic factors for long-term outcome, the issue of number of cycles of salvage chemotherapy before transplantation and the need to achieve a CR before transplant has not been fully addressed. This issue should be prospectively analysed in future trials. Other prognostic factors may also predict for long-term survival in patients with relapsed and refractory HL and several reports describe prognostic factors identifiable before transplantation that can predict for a poor outcome with this approach; the extent of extranodal significantly influences both TTF and OS in our study, as previously indicated by Reece et al. [15] in a group of 58 patients autografted in a single institution, by the group of City of Hope and Stanford University [11, 21] as well as by the French group in their retrospective analysis involving 270 patients [20]. Year of transplantation has also been found to have an impact on OS in our analysis, as previously described by

8 632 others [22]. The impact of modern supportive care that includes hematopoietic growth factor support, peripheral blood instead of BM as the source of hematopoietic progenitor cells and newer antibiotic and antifungal agents have markedly decreased transplant-related toxicity in the last 10 years. The long time period over which patients included in this study have been autografted is a clear reflection of this fact. Finally, we have also found other variables, which measure tumour burden at diagnosis, to significantly modify long-term outcome after the intensive procedure. As in our analysis, prior RT has also been described to negatively influence the results of ASCT by City of Hope [11]. There have been other attempts to construct prognostic indexes at relapse with the objective to identify subgroups of patients with a poor outcome with a conventional ASCT. In a series of 128 patients with relapsed HL homogeneously treated with the CBV protocol, Bierman et al. [7] found that a poor performance status, the failure of two or more chemotherapy protocols, and the presence of mediastinal disease predicted for a poor outcome after ASCT, with a 4-year failure-free survival of only 10% in those patients failing with two or more protocols. Reece et al. [15] reported an analysis on 58 patients treated with ASCT in a single institution. Four prognostic groups were identified according to the presence of the following parameters at relapse: B symptoms, extranodal disease and a short first CR. Patients with no risk factor had a 3-year progression-free survival of 100% compared with 81% in patients with one risk factor, 40% in those with two risk factors and finally, 0% in patients with three risk factors. In patients autografted with the CBV protocol or the combination of TBI cyclophosphamide and etoposide, the group of City of Hope showed that more than two prior chemotherapy protocols, prior RT and extranodal disease at ASCT predicted for a poor outcome after the procedure [11]. Similarly, and in a group that included 119 relapsed or refractory HL patients autografted with the CBV regimen or the TBI cyclophosphamide and etoposide protocol, the combination of B symptoms at relapse, BM or pulmonary involvement at ASCT and the presence of lymph nodes >_2 cm were able to separate different prognostic groups with 4-year event-free survivals of 85% in patients with no adverse prognostic factors compared with 41% in patients with one bad-prognosis factor [21]. Up to now, the largest report dealing with prognostic factors at relapse after a first CR had been reported by the French cooperative group (GELA) in 280 patients undergoing an ASCT [20]. They developed a two-factor model incorporating a short first CR and the presence of extranodal disease at relapse as adverse prognostic factors. With this model, patients with zero, one or two risk factors presented progression-free survival rates of 93%, 59% and 43%, respectively. Finally, two other groups of investigators have analysed the usefulness of risk factors at relapse to predict long-term outcome. The GHSG [29] retrospectively analysed the prognostic risk factors at relapse in a group of 422 patients from an initial cohort of 4754 patients registered in the GHSG database between 1988 and In multivariate analysis, independent risk factors were time to relapse, clinical stage at relapse and anemia at relapse. The constructed prognostic score was independent of the type of therapy used to treat the relapse (e.g. conventional chemotherapy or high-dose therapy with ASCT). More recently, the Memorial Sloan Kettering Cancer Center investigators [30] have developed a prognostic model of risk factors at relapse (B symptoms, extranodal disease and CR duration of <1 year) in a group of 65 patients (22 primary refractory and 43 relapsed) treated with two biweekly cycles of ifosfamide, carboplatin and etoposide. The presence of zero to one risk factor was associated with an event-free survival rate of 83% while it decreased to 10% in patients presenting with the three risk factors. This prognostic model has been used to develop tailored therapeutic strategies [31]. The results and usefulness of a prognostic factor analysis are limited not only by the selection of the factors as potential candidates for the analysis but also by the population of patients included into it. One of the potential pitfalls of our analysis is the inevitable selection bias related to a study only involving patients reported to a transplant database, but the number of patients involved in the analysis is significantly higher to what has been previously reported and the follow-up is significantly longer (>3 years for the surviving patients). In summary, the results of the present study indicate that it is possible to construct a prognostic factor score using easily collectable clinical variables that can identify cohorts of patients with a significantly different long-term outcome after high dose therapy and ASCT. The prognostic factors identified may be useful to develop tailored therapeutic approaches, with special clinical relevance in patients with poor-risk relapse who should be treated with innovative strategies. Acknowledgements This work has been partially supported by grant number FIS 94/ and FIS 97/0626 awarded by the Fondo de Investigaciones Sanitarias del Ministerio de Sanidad y Consumo, Spain, grant from Fundació La Caixa and grant from Red de Centros de Cáncer (FIS) number CO3/010. References 1. Longo DL, Young RC, Wesley M et al. Twenty years of MOPP therapy for Hodgkin s Disease. J Clin Oncol 1986; 4: Prosnitz LR, Farber LR, Kapp DS et al. Combined modality therapy for advanced Hodgkin s disease: 15 year follow-up data. J Clin Oncol 1988; 6: Longo DL, Duffey PL, Young RC et al. Conventional dose salvage combination chemotherapy in patients relapsing with Hodgkin s disease after combination chemotherapy: the low probability for cure. J Clin Oncol 1992; 10: Bonfante V, Santoro A, Viviani S et al. Outcome of patients with Hodgkin s disease failing after primary MOPP-ABVD. J Clin Oncol 1997; 15: Phillips GL, Wolff SN, Herzig RH et al. Treatment of progressive Hodgkin s disease with intensive chemoradiotherapy and autologous bone marrow transplantation. Blood 1989; 73:

9 Reece DE, Barnet MJ, Connors JM et al. Intensive chemotherapy with cyclophosphamide, carmustine, and etoposide followed by autologous bone marrow transplantation for relapsed Hodgkin s disease. J Clin Oncol 1991; 9: Bierman PJ, Bagin RG, Jagannath S et al. High dose chemotherapy followed by autologous hematopoietic rescue in Hodgkin s disease: long term follow-up in 128 patients. Ann Oncol 1993; 4: Chopra R, McMillan AK, Linch DC et al. The place of high dose BEAM therapy and autologous bone marrow transplantation in poorrisk Hodgkin s disease. A single center 8-year study of 155 patients. Blood 1993; 81: Crump M, Smith AM, Brandwein J et al. High-dose etoposide and melphalan, and autologous bone marrow transplantation for patients with advanced Hodgkin s disease: importance of disease status at transplant. J Clin Oncol 1993; 11: Rapoport AP, Rowe JM, Kouides PA et al. One hundred autotransplants for relapsed or refractory Hodgkin s disease and lymphoma: value of pretransplant disease for predicting outcome. J Clin Oncol 1993; 11: Nademanee A, O Donell MR, Snyder DS et al. High-dose chemotherapy with or without total body irradiation followed by autologous bone marrow and/or peripheral blood stem cell transplantation for patients with relapsed and refractory Hodgkin s disease: results in 85 patients with analysis of prognostic factors. Blood 1995; 85: Bierman PJ, Anderson JR, Freeman MB et al. High-dose chemotherapy followed by autologous hematopoietic rescue for Hodgkin s disease patients following first relapse after chemotherapy. Ann Oncol 1996; 7: Caballero MD, Rubio V, Rifón J et al. BEAM chemotherapy followed by autologous stem cell support in lymphoma patients: analysis of efficacy, toxicity and prognostic factors. Bone Marrow Transplant 1997; 20: Sureda A, Arranz R, Iriondo A et al. Autologous stem cell transplantation for Hodgkin s disease: Results and prognostic factors in 494 patients from the GEL/TAMO Spanish Cooperative Group. J Clin Oncol 2001; 19: Reece DE, Connors JM, Spinelli JJ et al. Intensive therapy with cyclophosphamide, carmustine, etoposide & cisplatin, and autologous bone marrow transplantation for Hodgkin s disease in first relapse after combination chemotherapy. Blood 1995; 83: Lohri A, Barnett M, Fairey RN et al. Outcome of treatment of first relapse of Hodgkin s after primary chemotherapy: identification of risks factors from the British Columbia experience 1970 to Blood 1991; 77: Yuen AR, Rosenberg SA, Hoppe RT et al. Comparison between conventional salvage therapy and high-dose therapy with autografting for recurrent or refractory Hodgkin s disease. Blood 1997; 89: Linch DC, Winfield D, Goldstone AH 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: Schmitz N, Pfistner B, Sextro M et al. Aggressive conventional chemotherapy compared with high-dose chemotherapy requiring autologous haemopoietic stem cell transplantation for relapsed chemosensitive Hodgkin s disease: a randomised trial. Lancet 2002; 359: Brice P, Bouabdallah R, Moreau P et al. Prognostic factors for survival after HDT and ASCT for patients with relapsed HD: Analysis of 280 patients from the French Registry. Bone Marrow Transplant 1997; 20: Horning SJ, Chao NJ, Negrin RS et al. High-dose therapy and hematopoietic progenitor cell transplantation for recurrent or refractory Hodgkin s disease: Analysis of the Stanford University results and prognostic indices. Blood 1997; 89: Lancet JE, Rapoport AP, Brasacchio R et al. Autotransplantation for relapsed or refractory Hodgkin s disease: long-term follow-up and analysis of prognostic factors. Bone Marrow Transplant 1998; 22: Fermé C, Mounier N, Divine M et al. Intensive therapy with highdose chemotherapy for patients with advanced Hodgkin s disease in relapse or failure after initial chemotherapy: Results of the Groupe d Études des Lymphomes de l Adulte H89 Trial. J Clin Oncol 2002; 20: Carbone PP, Kaplan HD, Musshogg K et al. Report of the Committee of Hodgkin s Disease staging. Cancer Res 1971; 31: Kaplan EL, Meier P. Nonparametric estimation from incomplete estimations. J Am Stat Assoc 1958; 53: Cheson BD, Horning SJ, Coiffier B et al. Report of an international workshop to standardize response criteria for non-hodgkin s lymphomas. J Clin Oncol 1999; 17: Mantel N, Haenzel W. Statistical aspects of the analysis of data from retrospective studies of disease. J Natl Cancer Inst 1958; 22: Fung HC, Nademanee A, Krishnan A et al. Autologous hematopoietic cell transplantation (ASCT) for patients with chemo-refractory Hodgkin s lymphoma (HL). Blood 2003; 102: 119a (Abstr 404). 29. Josting A, Franklin J, May M et al. New prognostic score based on treatment outcome of patients with relapsed Hodgkin s lymphoma registered in the Database of the German Hodgkin s Study group. J Clin Oncol 2002; 20: Moskowitz CH, Nimer SD, Zelenetz AD et al. A 2-step comprehensive high-dose chemotherapy second-line program for relapsed and refractory Hodgkin disease: analysis by intent to treat and development of a prognostic model. Blood 2001; 97: Moskowitz CH, Kewalramani T, Nimer SD et al. Risk-adapted high dose chemoradiotherapy and ASCT for patients with relapsed or refractory Hodgkin s disease: An intent to treat analysis. Blood 2003; 102: 118a (Abstr 403).