Prognostic Factors for the Clinical Outcome of Patients with Follicular Lymphoma Following High-Dose Therapy and Peripheral Blood Stem Cell Transplantation (PBSCT)

Bone Marrow Transplantation (2000) 25, 957–964  2000 Macmillan Publishers Ltd All rights reserved 0268–3369/00 $15.00 www.nature.com/bmt Prognostic factors for the clinical outcome of patients with follicular lymphoma following high-dose therapy and peripheral blood stem cell transplantation (PBSCT)

MT Voso1, S Martin2, S Hohaus2, A Abdallah2, RF Schlenk2,ADHo2 and R Haas1,2

1German Cancer Research Center, Heidelberg; and 2Department of Internal Medicine V, University of Heidelberg, Germany

Summary: modalities, particularly for patients with advanced stage disease. This is a report on 111 patients with advanced stage The role of dose-intensive therapy with hematopoietic follicular lymphoma who where autografted using stem cell support in aggressive lymphomas has been shown. PBSC. Seventy patients were enrolled in first remission, In a prospective randomized study in patients with high- whereas 41 were treated in second or higher remission. grade non-Hodgkin’s lymphoma (NHL), Philip et al5 have High-dose therapy consisted of total body irradiation analyzed the long-term outcome of patients with a chemo- plus cyclophosphamide in 103 patients, while eight sensitive relapse following conventional chemotherapy. patients received BEAM (carmustine, etoposide, cyto- The response rate after high-dose therapy with autologous sine-arabinoside, melphalan). Autografts contained bone marrow transplantation (ABMT) was 84%, compared ,؋ 106 CD34+ cells/kg body weight. At a with 44% after conventional therapy alone. At 5 years 0.6 ؎ 8.1 median follow-up of 44.2 months from PBSCT (range patients treated with high-dose chemotherapy and ABMT 4.9–77.4 months), 93 patients are alive, with a prob- showed a significantly better event-free and overall survival ability of overall and relapse-free survival (RFS) of 83% (OS) in comparison with patients receiving conventional and 64%, respectively. A significantly higher prob- dose chemotherapy.5 The role of high-dose therapy with ability of relapse was associated with male gender, hematopoietic stem cell support for patients with low-grade involvement of more than eight lymph node areas, NHL is however still controversial.6–10 extra-nodal manifestations other than bone marrow and In this report we have analyzed the outcome of 111 PBSCT performed in second or higher remission. In the patients with follicular lymphoma who were treated with latter group of patients, previous radiotherapy was high-dose therapy with PBSC in our institution between associated with poor prognosis. The relevance of chemo- 1991 and 1998. Particular emphasis was put on the identi- sensitivity as a prognostic factor was reflected by a bet- fication of prognostic factors which have impact on the ter RFS in patients who had achieved complete long-term outcome. remission at the time of PBSC mobilization. In a multivariate analysis, involvement of eight or more lymph nodes and high-dose therapy performed in second or Patients and methods higher remission were independent prognostic factors. Bone Marrow Transplantation (2000) 25, 957–964. Patients and treatment Keywords: follicular lymphoma; peripheral blood stem cell transplantation Patients were eligible for this study if they were age 60 or less and had an advanced-stage follicular lymphoma (centrocytic-centroblastic NHL according to the Kiel classification, or grade II follicular center lymphoma according Follicular lymphoma (FL) accounts for approximately 15% to the REAL classification). Advanced stage was defined to 30% of newly diagnosed lymphomas. Despite the initial as stage III–IV or stage II with bulky disease at the time indolent clinical course and the high response rates to con- of first diagnosis, or as relapse following conventional ther- ventional therapy, most patients will ultimately develop apy. Additional criteria were a Karnofsky score above 80% refractory disease or present with transformation into an and normal heart, lung, kidney and liver function. Informed 1–4 aggressive large cell lymphoma. In a group of 513 consent was obtained from each patient before therapy. The patients with FL treated at Stanford the failure-free survival study was conducted according to the guidelines of the at 10 years was 25% but no plateau in the survival curve Joint Ethical Committee of the University of Heidelberg. 1 was observed. This indicates the need for better treatment The cut-off date of this report is 31 December 1998. Between September 1991 and September 1998, 111 patients were included in this study. Fifty-eight patients Correspondence: Dr MT Voso, Ist Semeiotica Medica, Universita’ Cattol- were females and 53 were males, with a median age of 45 ica S Cuore, Largo A Gemelli, 1, 00168 Rome, Italy years (range 21–59) (Table 1). Out of 70 patients enrolled Received 26 August 1999; accepted 16 January 2000 in first remission, two had stage II disease with abdominal Prognostic factors in follicular lymphoma following PBSCT MT Voso et al 958 Table 1 Patients’ characteristics therapy. Two patients received radiotherapy alone before PBSC mobilization (Table 2). n 111 PBSCs were collected during granulocyte colony-stimul- Age (median, range) 45 ating factor (G-CSF)-enhanced marrow recovery (21–59) ␮ Sex (M/F) 53/58 (Neupogen; Amgen, Thousands Oaks, CA, 300 g s.c./day) PBSCT following chemotherapy. In one patient, PBSC were mobil- 1st remission 70 ized using G-CSF alone. Mobilization regimens used were Stage II 2a HAM (cytosine-arabinoside, 2 g/m2 every 12 h on days 1 Stage III 14 and 2, and mitoxantrone, 10 mg/m2, days 2 and 3) in 106 Stabe IV 54 у2nd remission 41 patients, DexaBEAM (dexamethasone, 24 mg/day, days 1– B symptoms 22 7, melphalan, 30 mg/m2, day 2, carmustine, 60 mg/m2, day Bulky disease 3, etoposide, 75 mg/m2 and cytosine-arabinoside, 200 у5cm 30 mg/m2, days 4–7) (n = 2 patients), and two subsequent у10cm 19 Abdominal 28 chemotherapy cycles in three patients (HAM/DexaBEAM Splenomegaly 32 and HAM/CHOP (cyclophosphamide, doxorubicin, vincris- + Bone marrow infiltration 65 tine and prednisone) (one patient)). CD34 cells were selec- Extranodal involvement ted in 15 patients using the Baxter Isolex 300 Magnetic GI tract 6 Cell Separation Systems (Baxter Immunotherapy, Irvine, Liver 8 11 Pleura 6 CA, USA). Otherb 10 The conditioning regimens consisted of total body Leukemic presentation 3 irradiation (TBI) (14.4 Gy, hyperfractionated into eight fractions over 4 days) followed by cyclophosphamide (200 aWith abdominal bulky disease. mg/kg) (n = 103 patients) or of BEAM (carmustine, 300 b Involvement of kidney (2 patients), bones (1 patient), soft-tissues (3 mg/m2, etoposide, 1.2 g/m2, cytosine-arabinoside, 800 patients), pancreas (2 patients) and central nervous system (1 patient). mg/m2, melphalan, 140 mg/m2) in eight patients who had previously received radiotherapy. No hematopoietic growth factors were administered following PBSCT. Hematological reconstitution was defined as the median bulk, while 14 and 54 patients had stage III and IV disease, time to reach an absolute neutrophil count (ANC) of respectively. Forty-one patients were enrolled in second or 0.5 ϫ 109/l and an unsubstituted platelet count of higher remission (range 2–5), following conventional 20 ϫ 109/l. therapy (Table 2). All patients autografted in first remission received induc- Response evaluation tion therapy, consisting of a median of four cycles (range 2–21 cycles) of anthracycline-containing regimens. One Follow-up restaging, including physical examination, blood patient additionally received involved-field radiotherapy as counts and chemistry, bone marrow biopsy, chest X-ray and part of the first-line therapy. Thirty-nine of 41 patients ultrasound of the abdomen, was performed at 3- to 6-month enrolled in second or higher remission received a median intervals following PBSCT. Computer-tomography of the of seven cycles of conventional cytotoxic chemotherapy chest and abdomen was performed in patients with previous (range 0–20), including 14 patients, who also had radio- bulky disease or with a suspicious chest X-ray or ultrasound of abdomen, respectively. Response was defined according to standard World Health Organization (WHO) criteria. Table 2 Disease history Overall survival was defined as the time between PBSCT and death or until last follow-up. Relapse-free survival PBSCT 1st remission у2nd remission (RFS) was defined as the time between PBSCT and recur- n = 70 n = 41 rence of disease. Patients dying of treatment-related causes were excluded from the latter analysis. Patients with an Previous treatment adverse event following PBSCT were censored at the time Radiation therapy 1 16 of its occurrence. Cycles of chemotherapy 4 7 (2–21) (0–20) Disease status at Statistical analysis PBSC mobilization MR/NC 3 Survival curves were estimated using the Kaplan–Meier Relapse 21 product limit method. Univariate and multivariate analyses PR 53 14 were performed to identify risk factors associated with CR 14 6 overall and relapse-free survival. Differences between the PBSC transplantation survival curves were compared using the log-rank test and Relapse 1 PR 30 17 the Wilcoxon test. Multivariate analysis was performed CR 40 23 using the Cox regression model with stepwise analysis (P values for entry = 0.15 and for removal = 0.05). MR = minimal response; NC = no change. The following risk factors were first examined in a univa-

Bone Marrow Transplantation Prognostic factors in follicular lymphoma following PBSCT MT Voso et al 959 riate analysis using the log rank test: age, sex, high-dose the other case. One patient died of invasive pulmonary therapy in first remission vs second or higher remission, aspergillosis. presence of bulky disease (tumor mass у5cminone Patients remained in hospital for a median time of 18 diameter), involvement of eight or more lymph nodes, days (range 9–124 days) following PBSCT and required a presence of extra-nodal involvement, BM involvement, median of six platelet and four erythrocyte transfusions splenomegaly, serum lactic dehydrogenase (LDH) levels, B (range 2–37 and 0–48, respectively). They had a median of symptoms at the time of start of conventional therapy. Dis- 4 days of fever over 38.5°C (range 0–43 days) and required ease status was analyzed at the time of PBSC mobilization intravenous antibiotic treatment for a median of 13 days and transplantation. The following lymph node areas were (range 0–52 days). Thirty-seven patients received ampho- scored: right and left cervical, supraclavicular, axillar, tericin-B for a median of 5.5 days (range 2–21 days). iliacal and inguinal, mediastinal and retroperitoneal. A pre- Following PBSCT, three patients received adjuvant treat- vious analysis performed by grouping patients according to ment with ␣-interferon and four patients with irradiation to the number of involved lymph nodes, indicated eight or sites of bulky disease. more areas as cut-off. The stage of the disease was examined as a prognostic parameter in patients autografted in Therapeutic outcome first remission, while previous radiotherapy and duration of first remission was taken into consideration in the group of Excluding the five treatment-related deaths, 106 patients patients autografted in second or higher remission. As a were evaluable for relapse-free survival. The 11 relapses in second step, variables which were significant in the univari- 65 patients autografted while in first remission were ate analysis were included in a multivariate analysis as observed after a median follow-up of 13.8 months (range stated above. Statistical computations were performed 0.3–42.5 months), while 17 of 41 patients autografted in using the statistical software package SAS, version 6.11 second or higher remission, relapsed at a median follow- (SAS Institute Inc, Heidelberg, Germany, 1995). up of 12.7 months (range 4.6–44.8 months). Independent of Results are expressed as mean Ϯ standard error of the the remission status, relapses occurred at sites of previous mean (s.e.m.) or as median and range. disease in 86% of patients, whereas exclusively new sites were noted in 14% of the patients (Table 3). Eight of 10 patients with BM relapse had a history of BM involvement Results during the course of the disease.

PBSC mobilization and transplantation Outcome of relapse following PBSCT In 87 patients, PBSC were mobilized when best response Twenty-two of 28 patients received a therapy with palli- was achieved following conventional anthracycline- ative intention at the time of relapse. This included ␣-inter- containing chemotherapy. At that time, 20 patients were in feron (11 patients), local radiotherapy (three patients), cyto- complete remission (CR) and 67 in partial remission (PR). toxic chemotherapy (one patient) and the administration of Twenty-four patients received mobilization therapy without the humanized CD20-antibody (Rituximab, Roche, Basel, previous induction. The mobilization therapy in 105 Switzerland) (one patient). In six patients a combination of patients consisted of high-dose cytosine arabinoside and chemotherapy and ␣-interferon or CD20-antibody was mitoxantrone, which resulted in a tumor response in 45 of the 91 patients (49.4%) with measurable disease (PR and relapse). As a consequence, 63 patients were in CR before Table 3 Clinical outcome the high-dose therapy was given, while 47 patients were in PR and one patient had no change (Table 2). n The patients received the PBSC-supported high-dose therapy at a median of 12.3 months following initial diag- Total 111 nosis (range 4.8–129 months). In 103 patients, high-dose Toxic deaths 5a therapy consisted of TBI plus cyclophosphamide, while Death in remission 4b BEAM was administered to eight patients. Following auto- Death due to progression 9 Alive 93 Ϯ ϫ 6 ϩ grafting with 8.1 0.6 10 CD34 cells/kg body weight, Continuous CR 74 106 patients engrafted with a median time to a neutrophil Relapse 19 count of 0.5 ϫ 109/l and a platelet count of 20 ϫ 109/l of Sites of relapse 13 and 12 days (range 8–34 and 3–51), respectively. CD34- Previous sites 18 Previous and new sites 6 selected autografts were used in 15 patients. New sites only 4 Five patients (4.5%) died at a median of 17 days follow- BM with previous BM+ 8 ing PBSCT (range 13–188 days). One patient transplanted BM new 2 using 5.6 ϫ 106/kg BW selected CD34+ cells and another Secondary malignancies 5c patient who received unseparated leukapheresis products, containing 4.9 ϫ 106 CD34+ cells/kg BW, had graft failure aTwo graft failures, three infectious complications. bTwo myelodysplastic syndromes, one multifocal encefalopathy and one and died of sepsis and pneumonia, respectively. Two pneumonia. patients died of septicemia: Pseudomonas aeruginosa was cTwo Hodgkin’s disease, two myelodysplastic syndromes and one isolated in one case, no bacteria could be isolated in gastric cancer.

Bone Marrow Transplantation Prognostic factors in follicular lymphoma following PBSCT MT Voso et al 960 given. As a result, 19 patients are alive at a median follow- 1.0 up of 28.9 months (range 3–57.2 months) after the relapse 0.9 was diagnosed. This group includes 11 patients autografted Females (n = 56) in second or higher remission. 0.8 Three patients transplanted in ﬁrst remission and six 0.7 patients transplanted in second or higher remission died of 0.6 tumor progression following a median time of 7.3 months (range 0–16.5 months) after they had relapsed. Two 0.5 0.4 Males (n = 50)

patients without evidence of lymphoma died of compli- Probability cations related to a myelodysplastic syndrome at 32 and 0.3 20.6 months following PBSCT, while two patients died of multifocal encephalopathy and pneumonia at 22.2 and 10.4 0.2 months from PBSCT, respectively (Table 3). 0.1 P = 0.048

The corresponding Kaplan–Meier estimate for relapse- 0.0 free survival at a median of 44.2 months (range 4.9–77.4 0123456 months) following PBSCT was 64%, while the probability Years from PBSCT of overall survival was 83% (Figure 1). Figure 2 Male patients had a higher probability of relapse following PBSCT than females. Prognostic factors Our analysis showed that male gender was associated with a higher risk of relapse (RFS: 79% vs 45%, P = 0.048) (Figure 2). Similarly, the probability of relapse was higher trend towards better prognosis for patients who had achi- for patients with extensive disease, reflected by eight or eved CR before PBSC mobilization (57% vs 83%, P = more involved lymph node areas and extra-nodal involve- 0.054) (Figure 4a). On the other hand, the probability of ment or leukemic presentation (RFS: 40% vs 78%, P = RFS was not related to the number of previous chemo- 0.001, and 59% vs 67%, P = 0.045, respectively) (Table 4, therapy cycles or regimens administered. Confirming our Figure 3). Bulky disease with a lymphoma mass greater previous data on a smaller number of patients,7 a significant than 5 cm in one diameter was present in 30 patients, with difference in the probability of RFS was noted between abdominal localization in 28 of them. Nineteen of these patients transplanted in first remission and patients auto- patients presented with a retroperitoneal bulk, which was grafted in second or higher remission (78% vs 43%, P = larger than 10 cm. Nonetheless, independent of the size, 0.003) (Figure 4b). For the latter group of patients, previous the presence of bulky disease was not a prognostic factor. irradiation was associated with a significantly worse prog- The probability of relapse-free survival was also not related nosis (RFS: 21% vs 56%, P = 0.03) (Figure 5). to age, initial stage of the disease, presence of BM involve- None of the 14 evaluable patients autografted with ment and serum LDH levels (Table 4). CD34-selected PBSC relapsed at a median follow-up of The relevance of chemosensitivity was reflected by a 12.9 months (range 4.9–24.3 months). According to the multivariate analysis, the involvement of eight or more lymph node areas and high-dose therapy 1.0 performed in second or higher remission were independent ␹2 = 0.9 prognostic factors for disease recurrence ( -test: P 0.001 Overall survival (n = 111) and 0.009, respectively). 0.8 On the other hand, the presence of extra-nodal involve- 0.7 ment, other than bone marrow, was the only variable that significantly adversely affected overall survival (72% vs 0.6 Relapse-free survival (n = 106) 90%, P = 0.01) in the univariate analysis (Table 4). 0.5

0.4 Late effects of PBSCT Probability

0.3 At a median of 11.9 months (range 8.7–20.9 months) following PBSCT, 10 patients had leukocyte counts below 0.2 4 × 109/l, platelet counts below 150 ϫ 109/l and hemoglo- 0.1 bin below 14 g/dl. Of the ﬁve patients with platelet counts ϫ 9 0.0 below 100 10 /l over a median time of 11.7 months 0123456(range 8.9–17.4 months), one patient required platelet trans- Years from PBSCT fusions. Four patients had a hemoglobin below 10 g/dl, with two requiring erythrocyte transfusions. Five patients Figure 1 Overall (OS) and relapse-free survival (RFS) in 111 patients developed a secondary malignancy (two Hodgkin’s disease, with advanced stage follicular lymphoma following high-dose therapy and PBSCT. Patients dying of treatment-related causes were excluded from two myelodysplastic syndromes, and one gastric cancer) at the RFS analysis. Patients who had an adverse event were censored at the a median of 32.4 months following PBSCT (range 8.7–50.7 time of occurrence of the latter. months) (Table 3).

Bone Marrow Transplantation Prognostic factors in follicular lymphoma following PBSCT MT Voso et al 961 Table 4 Prognostic factors

Overall survival P Relapse-free survival P (%, n = 111) (%, n = 106)

Age, years р45 87 70 Ͼ45 76 0.3 62 0.7 Sex Male 76 45 Female 84 0.55 79 0.048a Lymph nodes р78578 у8 72 0.25 40 0.001a,b Extranodal involvement or leukemic presentation positive 72 59 negative 90 0.01a 67 0.045a Bone marrow negative 80 67 positive 83 0.7 61 0.36 Serum LDH levels р240 IU/l 85 70 Ͼ240 IU/l 77 0.33 55 0.38 Bulky disease negative 78 64 positive 84 0.26 64 Cycles of chemotherapy р48471 у5 80 1 60 0.57 High-dose therapy in 1st remission 86.5 78 in у2nd remission 72 0.27 43 0.003a,b Status at PBSC mobilization CR 88 83 measurable disease 80 0.37 57 0.054 Status at PBSCT CR 82 67 measurable disease 82 60 0.136

Overall and relapse-free survival were studied at a median of 44.2 months following PBSCT. aStatistic signiﬁcant by the log rank test. bIndependent prognostic factor in the multivariate analysis.

Discussion failure-free survival was the number of chemotherapy regimens given before autografting. In the present study, we report on the outcome of a group Although with a relatively short follow-up of approxi- of 111 patients with advanced-stage follicular NHL, treated mately 4 years, the 64% probability of relapse-free survival in our center between 1991 and 1998. The results obtained of our patients is better than that reported in studies using with our PBSC-supported high-dose therapy are in line with conventional cytotoxic chemotherapy.1,4,13,14 In a recent the results reported by Freedmann et al.9 At a median fol- report on a group of 165 patients with previously untreated low-up of 45 months, a RFS of 63% was observed in 77 stage III–IV FL, an EFS of 35% at 4 years was observed.4 patients with previously untreated stage III–IV FL.9 The The patients were randomized to receive cyclophospham- patients were in CR following CHOP chemotherapy, and ide, vincristine, prednisone or mitoxantrone and prednimus- received bone marrow purged with anti-B cell monoclonal tine, followed by a second randomization for maintainance antibodies and complement, following a TBI-containing therapy with or without interferon. conditioning regimen. The 44% 4-year estimated prob- By interpreting the data, one has to bear in mind that ability of failure-free survival reported by Bierman et al12 any transplantation study is hampered by a selection bias, in 100 patients with FL, probably results from the inclusion as for instance the patients’ median age is substantially of patients with refractory or recurrent follicular lymphoma. lower than that of patients undergoing conventional ther- In their study, the only prognostic factor for overall and apy. The inﬂuence of the patients’ age on treatment out-

Bone Marrow Transplantation Prognostic factors in follicular lymphoma following PBSCT MT Voso et al 962 a a 1.0 1.0 P = 0.0013 P = 0.054 0.9 0.9

0.8 0.8 CR at MOB (n = 19) 0.7 0.7 0–7 areas (n = 68) 0.6 0.6 0.5 PR or R at MOB (n = 87) 0.5 0.4 0.4 Probability у8 areas (n = 31) 0.3 0.3 0.2 0.2 0.1 0.1 0.0 na (n = 7) 0123456 0.0 0123456 b b 1.0 1.0 P = 0.0035 P = 0.0045 0.9 0.9 0.8 0.8 No extra-nodal (n = 76) 0.7 1st remission (n = 65) 0.7 0.6 0.6 0.5 0.5 Leukemic or extra-nodal (n = 30) 0.4 0.4 Probability2nd Probability remission (n = 41) Probability 0.3 0.3 0.2 0.2 0.1 0.1 0.0 0.0 0123456 0123456 Years from PBSCT Years from PBSCT Figure 4 (a) Patients achieving complete remission (CR) at PBSC mobi- Figure 3 a Extension of disease and relapse-free survival. ( ) Number lization (MOB); and (b) those autografted in ﬁrst remission had a signiﬁ- b of lymph-node areas involved and ( ) extra-nodel involvement and cantly better relapse-free survival than those with measurable disease and leukemic presentation. enrolled in second or higher remission, respectively (PR = partial remission; R = relapse).

come was particularly demonstrated by Dana et al14 in a group of 415 patients treated with doxorubicin-based 1.0 chemotherapy or chemo-immunotherapy: half of the P = 0.03 patients aged less than 40, and 26% of those aged more 0.9 than 60 were alive at a median follow-up of 10 years. 0.8 The International Prognostic Index, developed for 0.7 patients with high-grade NHL, has also been applied to patients with follicular lymphoma,15,16 but was not appli- 0.6 No radiotherapy (n = 25) cable to our analysis, since only young patients, with 0.5 advanced-stage disease and a good performance status were 0.4

13 Probability included in our study. As proposed by Romaguera et al, 0.3 patients’ characteristics, like male gender and high tumor burden, indicate a group of patients at high-risk for relapse. 0.2 Previous radiotherapy (n = 16) This also relates to patients with relapse or progressive dis- 0.1 ease. Accordingly, in our patients the extent of the disease 0.0 expressed as number of involved lymph node areas and 0123456 extra-nodal involvement were negative prognostic factors. Years from PBSCT 9 As reported by other groups, the relapse was predomi- Figure 5 Previous radiotherapy was associated with a signiﬁcantly nantly localized at sites of previous disease, while 80% of worse probability of relapse-free survival in patients autografted in second patients with BM relapse had former marrow inﬁltration. or higher remission.

Bone Marrow Transplantation Prognostic factors in follicular lymphoma following PBSCT MT Voso et al 963 These data suggest that the origin of relapse is residual phomas: results of a prospective comparative analysis of the tumor cells in vivo, resistant to the high-dose treatment. German Low-Grade Lymphoma Study Group. J Clin Oncol In line with these results, we found that t(14;18) PCR- 1998; 16: 1922–1930. positivity in blood or bone marrow samples following 5 Philip T, Guglielmi C, Hagenbeek A et al. Autologous bone PBSCT, was associated with an 4.5 estimated hazard ratio marrow transplantation as compared with salvage chemotherapy in relapses of chemotherapy sensitive non Hodgkin’s for relapse in comparison with a PCR-negative finding lymphoma. New Engl J Med 1995; 333: 1540–1545. while continued PCR-negativity was associated with 6 Rohatiner AZ, Johnson PW, Price CG et al. Myeloablative 17 prolonged freedom from recurrence. therapy with autologous bone marrow transplantation as con- The significance of contaminating PCR-positive cells in solidation therapy for recurrent follicular lymphoma. J Clin the autografts of patients with low-grade NHL is not Oncol 1994; 12: 1177–1184. clear.4,7,17–19 Peripheral blood stem cells may contain fewer 7 Haas R, Moos M, Karcher A et al. Sequential high dose ther- tumor cells than bone marrow7,20,21 and patients autografted apy with peripheral blood progenitor cell support in low grade with a PCR-negative BM were reported to have signifi- non-Hodgkin’s lymphoma. J Clin Oncol 1994; 12: 1685– cantly longer time of freedom from recurrence than those 1692. whose BM was PCR-positive (88% vs 51%).9 This could 8 Bastion Y, Brice P, Haioun C et al. Intensive therapy with peripheral blood progenitor cell transplantation in 60 patients be simply an indicator for a lower tumor burden in vivo. with poor prognosis follicular lymphoma. Blood 1995; 86: The toxicity of our treatment was moderate, with a rate 3257–3262 of 4.5% treatment-related deaths. Secondary malignancies 9 Freedman AS, Gribben JG, Neuberg D et al. High dose ther- were rarely observed (4.5%) and involved the lympho- apy and autologous bone marrow transplantation in patients hematopoietic system in four patients, while one patient with follicular lymphoma during first remission. Blood 1996; developed gastric cancer. These data are compatible with 88: 2780–2786. previous reports.6,8,9,22 There was also a transient period of 10 Vose JM, Bierman PJ, Lynch JC et al. Effect of follicularity mild pancytopenia in 9% of patients, with two patients on autologous transplantation for large cell non-Hodgkin’s lymphoma. J Clin Oncol 1998; 16: 844–849. requiring blood transfusions. The delay of hematopoietic + recovery observed in these patients might reflect the impair- 11 Voso MT, Hohaus S, Moos M et al. Autografting with CD34 peripheral blood stem cells (PBSC): retained engraftment ment of the hematopoietic microenvironment or insufficient Br J Haematol 23,24 capability and reduced tumor cell content. homing following PBSCT. 1999; 104: 382–391. In our patients, the 78% probability of relapse-free sur- 12 Bierman PJ, Vose JM, Anderson JR et al. High dose therapy vival for patients autografted in first remission is encour- with autologous hematopoietic rescue for follicular low grade aging. To improve these results we are evaluating the incor- non Hodgkin’s lymphoma. J Clin Oncol 1997; 15: 445–450. poration of the humanized CD20-antibody in the PBSC 13 Romaguera JE, McLaughlin P, North L et al. Multivariate mobilization and conditioning regimens. In addition, ex analysis of prognostic factors in stage IV follicular low-grade vivo activated T cells loaded with a CD3xCD19 bi-specific lymphoma: a risk model. J Clin Oncol 1991; 9: 762–769. antibody or B cell-specific antibodies conjugated with 14 Dana BW, Dahlberg S, Nathwani BN et al. Long-term follow- 213Bismuth might be used as adjuvant treatment following up of patients with low-grade malignant lymphomas treated with doxorubicin-based chemotherapy or chemoimmunother- high-dose therapy. apy. J Clin Oncol 1993; 11: 644–651. 15 Shipp MA, Harrington DP for the International Non-Hodg- kin’s Lymphoma Prognostic Factors Project. A predictive Acknowledgements model for aggressive non-Hodgkin’s lymphoma. New Engl J Med 1993; 329: 987–994. We wish to thank all the colleagues who during these years 16 Lopez-Guillermo A, Montserrat E, Bosch F et al. Applica- referred patients to our centre, the staff of the Transplantation bility of the international index for aggressive lymphomas to Unit and of the Department of Radiology for the outstanding care patients with low-grade lymphoma. J Clin Oncol 1994; 12: of patients. The excellent technical assistance of Renate Alexi, 1343–1348. Kirsten Flentje, Monika Hess, Evi Holdermann, Margit Pfo¨rsich, 17 Moos M, Schulz R, Martin S et al. The remission status before Petra Schmidt, Lena Volk and Mirjam Weis and the continuous and the PCR status after high dose therapy with peripheral secretarial help of Ursula Scheidler are also acknowledged. blood stem cell support are prognostic factors for relapse free survival in patients with follicular non Hodgkin’s lymphoma. Leukemia 1998; 12: 1971–1976. 18 Zwicky CS, Maddocks AB, Andersen N, Gribben JG. Eradi- References cation of polymerase chain reaction detectable immunoglob- ulin gene rearrangement in non-Hodgkin’s lymphoma is asso- 1 Rosenberg SA. Karnofsky memorial lecture. The low-grade ciated with decreased relapse after autologous bone marrow non-Hodgkin’s lymphomas: challenges and opportunities. J transplantation. Blood 1996; 88: 3314–3322. Clin Oncol 1985; 3: 299–310. 19 Widmer L, Pichert G, Jost LM, Stahel RA. Fate of contaminat- 2 Horning SJ. Treatment approaches to the low-grade lym- ing t(14;18)+ lymphoma cells during ex vivo expansion of phomas. Blood 1994; 834: 881–884. CD34 selected hematopoietic progenitor cells. Blood 1996; 3 Johnson PW, Rohatiner AZ, Whelan JS et al. Patterns of sur- 88: 3166–3175. vival in patients with recurrent follicular lymphoma: a 20 year 20 Vescio RA, Han EJ, Schiller GJ et al. Quantitative comparison study from a single center. J Clin Oncol 1995; 13: 140–147. of multiple myeloma tumor contamination in bone marrow 4 Hiddemann W, Unterhalt M, Herrmann R et al. Mantle cell harvest and leukapheresis autografts. Bone Marrow Trans- lymphomas have more widespread disease and a slower plant 1996; 18: 103–110. response to chemotherapy compared with follicle center lym- 21 Leonard BM, Hetu F, Busque L et al. Lymphoma cell burden

Bone Marrow Transplantation Prognostic factors in follicular lymphoma following PBSCT MT Voso et al 964 in progenitor cell grafts measured by competitive polymerase bone marrow transplantation: relationship with rate of chain reaction: less than one log difference between bone mar- engraftment and evidence for long lasting reduction in stem row and peripheral blood sources. Blood 1998; 91: 331–339. cell numbers. Blood 1996; 87: 3963–3969. 22 Rohatiner A. Myelodysplasia and acute myelogenous leuke- 24 Voso MT, Murea S, Goldschmidt H et al. High dose therapy mia after myeloablative therapy with autologous stem cell with peripheral blood stem cell transplantation results in a sig- transplantation. J Clin Oncol 1994; 12: 2521–2523. niﬁcant reduction of the haematopoietic progenitor cell com- 23 Kirkland MA, Spencer A, Davidson RJ et al. Quantitation of partment. Br J Haematol 1996; 94: 759–766. mafosfamide resistant pre-colony forming units in allogeneic

Bone Marrow Transplantation