- © 2003 by American Society of Clinical Oncology
Treatment of Indolent B-Cell Nonfollicular Lymphomas: Final Results of the LL01 Randomized Trial of the Gruppo Italiano per lo Studio dei Linfomi
- Luca Baldini,
- Maura Brugiatelli,
- Stefano Luminari,
- Marco Lombardo,
- Francesco Merli,
- Stefano Sacchi,
- Paolo Gobbi,
- Marina Liberati,
- Luigi Cavanna,
- Mariangela Colombi,
- Caterina Stelitano,
- Maria Goldaniga,
- Fortunato Morabito,
- Massimo Federico and
- Vittorio Silingardi
- From the Unità Operativa Ematologia 1, Dipartimento di Ematologia, Ospedale Maggiore, IRCCS, Milano; Divisione di Ematologia, Azienda Ospedaliera “Papardo,” Messina; Dipartimento di Oncologia ed Ematologia, Università di Modena e Reggio Emilia, Modena; Dipartimento di Ematologia ed Oncologia, Presidio Ospedaliero, Pescara; Servizio di Ematologia, Arcispedale S. Maria Nuova, Reggio Emilia; Medicina Interna e Oncologia Medica, Università di Pavia, Policlinico S. Matteo, IRCCS, Pavia; Dipartimento di Ematologia, Azienda Ospedaliera Bianchi-Melacrino-Morelli, Reggio Calabria; Istituto di Medicina Interna e Scienza Oncologiche, Policlinico Monteluce, Perugia; and Sezione di Ematologia, Divisione di Medicina I, Ospedale Civile, Piacenza, Italy.
- Address reprint requests to Luca Baldini, MD, Unità Operativa Ematologia 1, Centro “G. Marcora,” Ospedale Maggiore, IRCCS, Via F. Sforza 35, 20121 Milano, Italy; email: lubaldini{at}policlinico.mi.it.
Abstract
Purpose: To evaluate the effect of epirubicin on therapeutic response and survival in patients with indolent nonfollicular B-cell lymphomas (INFL) treated with pulsed high-dose chlorambucil.
Patients and Methods: A total of 170 untreated patients with advanced/active INFL were randomly assigned to receive either eight cycles of high-dose chlorambucil (15 mg/m2/d) plus prednisone (100 mg/d) for 5 days (HD-CHL-P; arm A) or eight cycles of HD-CHL-P plus epirubicin 60 mg/m2 intravenous on day 1 (arm B). The responding patients were randomly assigned to either maintenance therapy with interferon alfa (IFNα-2a; 3 MU, three times weekly) for 12 months or observation.
Results: There were 160 assessable patients (82 males, 78 females; median age, 63 years; range, 33 to 77 years); 77 patients were assigned to arm A, and 83 were assigned to arm B. Induction therapy led to 47 complete responses (CRs; 29.4%) and 68 partial responses (PRs; 42.5%), with no significant difference between the two arms (60 CR + PR in arm A [77.9%] and 55 CR + PR in arm B [66.3%]; P = .07). After a median follow-up of 38 months (range, 2 to 103 months), there was no between-group difference in overall survival (OS; P = .45), failure-free survival (P = .07), or progression-free survival (PFS; P = .5). Eighty-eight patients were randomly assigned to either IFNα-2a (n = 43) or observation (n = 45), without any difference in 3-year PFS (44% and 42%, respectively). Univariate analysis showed that OS was influenced by age, anemia, serum lactate dehydrogenase levels, and International Prognostic Index distribution; multivariate analysis identified age and anemia as having influence on OS.
Conclusion: HD-CHL-P treatment outcome in INFL patients was good (50% 3-year PFS, minimal toxicity, and low costs); epirubicin did not add any advantage. One-year IFNα maintenance treatment did not prolong response duration.
SINCE THEIR working formulation identification as a group of lymphomas characterized by an indolent course in the early 1980s,1 low-grade B-cell non-Hodgkin’s lymphomas (NHLs) have been a subject of debate among hemato-oncologists because of their biologic and clinical heterogeneity. According to more recent lymphoma classifications.2,3 This subject of NHLs include small lymphocytic lymphomas (SLLs), lymphoplasmacytic lymphomas (LPCs), marginal-zone lymphomas (MZLs), and grade 1 and 2 follicular lymphomas (FLs).
The higher incidence and typical clinicobiologic features of FLs have made them the object of many investigations, and a large body of biologic, prognostic, and therapeutic knowledge has been accumulated.4,5 Conversely, indolent nonfollicular B-cell lymphomas (INFLs) have never been extensively investigated, and only retrospective studies or prospective trials involving limited series have so far been published.6–,9
Although much information has been acquired over the last few years about the clinical course and therapeutic approach in MZL,10–,12 INFLs have long been considered a single group of diseases in therapeutic terms. As in the case of chronic lymphocytic leukemia (CLL), chemotherapy is not indicated in early and stable indolent disease, but it is commonly used in advanced or progressive disease with the aim of controlling its clinical course. The conventional approaches to this subset of NHLs include single-agent chemotherapy (chlorambucil or cyclophosphamide) or combination chemotherapy with or without doxorubicin (cyclophosphamide, vincristine, and prednisone or cyclophosphamide, doxorubicin, vincristine, and prednisone [CHOP]).6–8,13 More recently, purine analogues (fludarabine or 2-chorodeoxy adenosine) have been found to be highly effective.14–,16
The overall response (OR) rates to conventional chemotherapy range from 63% to 84%, but low complete response (CR) rates have been reported (7% to 39%); the median time to treatment failure is 24 to 48 months without any evidence of a plateau phase, and median overall survival (OS) ranges from 60 to 118 months. However, none of the studies have shown a significant cure rate.9,14,15,17 Conversely, few prospective trials have been specifically designed for patients with INFL, and the above results relate to small series frequently considered together with other types of low-grade B-cell NHLs.
In the early 1990s, the promising results of modified anthracycline-based regimens in advanced CLL patients reported by different groups18–,20 induced a number of investigators to test the addition of an anthracycline in the treatment of indolent lymphomas, but no published randomized trial has yet investigated this option in INFL. During the 1980s, it was found that recombinant interferon (IFN) has moderate single-agent activity in relapsed patients with indolent lymphoma21; however, in the early 1990s, no randomized trials investigating the role of IFN in combination with chemotherapy in the treatment of low-grade lymphomas had been published, and the question of the role of biologic response modifiers in the treatment of INFL was still open.
On the basis of these considerations, the Gruppo Italiano per lo Studio dei Linfomi (GISL) decided to evaluate whether the addition of epirubicin to pulsed high-dose chlorambucil could improve outcome in a randomized trial involving a selected series of patients with active-phase INFLs. The secondary objective of the study was to verify the hypothesis that 1-year maintenance therapy with IFNα-2a after induction may improve the duration of response.
PATIENTS AND METHODS
Patients
Between July 1993 and December 2001, 170 patients with newly diagnosed INFLs were enrolled onto the LL01 trial. The patients were eligible if they had a histologic diagnosis of SLL, lymphoplasmacytoid lymphoma/immunocytoma (IC), LPC, or monocytoid lymphoma on nodal or extranodal biopsy according to the revised Kiel classification.22 After the publication of the Revised European-American Lymphoma (REAL) classification,2 the inclusion criteria were extended to include patients with a diagnosis of MZL. In the absence of nodal or extranodal diagnostic material, an immunocytomorphologic diagnosis of CD5- and CD10-negative, mature, B-cell leukemia/lymphoma on the basis of peripheral or bone marrow blood analysis was accepted. Patients with primary mucosa-associated lymphoid tissue (MALT) lymphoma, mantle-cell lymphoma, and classic CLL (lymphocytes > 5,000/μL, with CD5 and CD23 expression and dim surface immunoglobulins) were excluded. Other inclusion criteria included no previous treatment; Ann Arbor stage II (with more than three involved sites), III, or IV; and the presence of at least one of the following GISL criteria for the definition of active disease23: systemic symptoms; hemoglobin less than 10g/dL; platelets less than 100,000/μL; diffuse bone marrow infiltrate; lymphocyte doubling time less than 12 months; or bulky disease or a doubling of the maximum diameter of at least three measurable sites in less than 12 months. The patients had to have normal liver and renal function, no human immunodeficiency virus infection, and no cardiac abnormality contraindicating anthracycline therapy. The protocol was first approved by the ethics committee of the coordinating center and then by the institutional review board of each participating center. All of the patients were required to give their informed consent before entering the study.
To improve the quality of study reporting, we retrospectively followed the recommendations of the revised CONSORT (consolidated standards of reporting trials) statement for reporting randomized trials.24 All of the data are reported on an intent-to-treat basis.
Treatment Schedule
The patients fulfilling the inclusion criteria were randomly assigned to receive either high-dose oral chlorambucil 15 mg/m2 and prednisone (HD-CHL-P) 100 mg/d for 5 days (arm A) or HD-CHL-P plus epirubicin (HD-CHL-PE) 60 mg/m2 intravenous on day 1 (arm B). Both regimens were administered every 28 days. Randomization was stratified by institution and leukemic disease, defined as the presence of at least 5,000/μL circulating lymphocytes. After random assignment to treatment, the eligibility of the patients was verified, and any major violators of the inclusion criteria were excluded. The patients responding after three courses received an additional five courses, and those in CR or partial response (PR) at the end of induction therapy were offered subsequent random assignment to observational follow-up or maintenance treatment with IFNα-2a at a dose of 3 MU three times a week for 12 months.
The patients with stable or progressive disease were switched to fludarabine monophosphate monochemotherapy (25 mg/m2 intravenous days 1 to 5 every 28 days for four cycles) if they had been assigned to arm B, or to a combination of cyclophosphamide, epidoxorubicin, vincristine, and prednisone if they had been assigned to arm A.
Patient Evaluation and Response Criteria
The patients were observed until the progression of lymphoma or death. Standard response criteria were applied, and a response had to be confirmed at two consecutive visits separated by a 3-month interval. A CR was defined as the absence of constitutional symptoms and lymphadenopathy, splenomegaly, and hepatomegaly at physical and computed tomographic examination, a normal peripheral blood count and bone marrow biopsy, and the disappearance of a measurable M band when correlated with the diagnosis. A PR was defined as a reduction of at least 50% in the size of measurable lesions, as measured by the sum of products of perpendicular diameters of the greatest lesions, without the appearance of any new manifestations of disease, and a reduction of at least 50% in the serum M band and (for leukemic cases) in the number of peripheral blood lymphocytes versus the pretreatment value (in the presence of a peripheral neutrophil count of 1,500/μL, a platelet count of 100,000/μL, and a hemoglobin level of 11 g/dL).
Stable disease was defined as a regression of less than 50% in the size of assessable disease sites or an increase of less than 25% in the size of existing lesions, without the appearance of any new manifestations of disease. Progressive disease was defined as a size increase of at least 25% in a site involved before therapy or the appearance of any new disease manifestations. OS was measured from the time of random assignment to treatment to the time of death for any cause or the date of the last follow-up visit; progression-free survival (PFS) was measured from the end of induction therapy (when the response was evaluated) to the time of disease progression or death; and failure-free survival (FFS) was measured from the time of diagnosis to the time of treatment failure, which was defined as no response to induction therapy, disease progression or relapse, or death. Treatment toxicity was recorded and graded using a four-point scale according to the World Health Organization (WHO) recommendations for the grading of acute and subacute toxic effects.
Statistical Methods
The sample size was calculated on the basis of FFS, assuming a 20% 5-year FFS and a 15% lower 5-year FFS in the worst group. With a power of 80% and a significance level of 0.05 using a two-sided test, 82 patients per arm were required. Assuming a 10% rate of ineligible patients after randomization for any reason, a final accrual of 180 patients was planned
The statistical analyses were performed by the Statistics and Data Management Unit at the GISL Trial Office using version 10 of the Statistical Package for Social Sciences (SPSS Inc, Chicago, IL). Proportions were compared by means of Fisher’s exact test, and all of the time-to-event distributions were calculated using the Kaplan-Meyer method and compared using the log-rank test. For each group comparison, a P value of .05 (two-sided) was considered significant.
RESULTS
Flow of Participants
Between July 1993 and December 2001, 170 patients were registered in the study and randomly assigned to receive HD-CHL-P (82 patients; arm A) or HD-CHL-PE (88 patients; arm B). After registration, we could not collect any data for 10 patients who were excluded by local investigators for the following reasons: five patients on the basis of a histologic review performed by the local pathologist before treatment began (CLL, n = 2; mantle cell leukemia, n = 1; prolymphocytic leukemia, n = 1; hairy cell leukemia, n = 1), one because of unconfirmed active disease according to the defined criteria, one had received previous treatment, one refused treatment, and two missed additional treatment and follow-up. At the time of the final evaluation, data relating to 160 patients were available and are included in this analysis; 77 of these patients were assigned to HD-CHL-P, and 83 were assigned to HD-CHL-PE. The median age of the series as a whole was 62.8 years (range, 33 to 77 years). The clinical characteristics of the two groups, distributed according to treatment arm at diagnosis, are listed in Table 1⇓. An immunocytomorphologic diagnosis was made in 32 patients. The analysis was made according to the intent-to-treat principle. On the basis of the decision of their physician, three patients assigned to arm A were treated with the arm B schedule after random assignment to treatment, and three patients assigned to arm B were treated with chlorambucil alone.
Patient Characteristics: Selected Variables
Response to Treatment and Survival
At the end of the planned therapy, 26 patients in arm A (33.8%; 95% confidence interval [CI], 23.2% to 44.3%) and 21 patients in arm B (25.3%; 95% CI, 15.9% to 34.7%) achieved a CR, with no difference between the two groups (P = .29). The OR rates were 77.9% in arm A (95% CI, 68.7% to 87.2%) and 66.3% in arm B (95% CI, 56.1% to 76.4%; P = .07; Table 2⇓). After a median follow-up of 38 months (range, 2 to 103 months), there were 45 induction therapy failures, 55 relapses, and eight deaths, for a total of 108 events. The 5-year FFS was 31% in arm A (95% CI, 18% to 43%) and 16% in arm B (95% CI, 7% to 26%; P = .07; Fig 1⇓). The 3-year PFS among the 115 patients who responded to induction therapy was 50% in arm A (95% CI, 34% to 66%) and 36% in arm B (95% CI, 20% to 51%; P = .37; Fig 2⇓). Fifty-five patients have died so far. Lymphoma was the cause of death in 16 (66.7%) of 24 patients in arm A (95% CI, 47.8% to 85.5%) and in 22 (70.9%) of 31 patients in arm B (95% CI, 55% to 86.9%); six patients (10.9%; 95% CI, 2.7% to 19.1%), four patients in arm A and two patients in arm B, died of a second malignancy; finally, four patients in arm A (5.2%; ]95% CI, 0.2% to 10.2%) and seven patients in arm B (8.4%; 95% CI, 2.5% to 14.4%) died without any evidence of disease progression during the follow-up. The 5-year OS rate was 66% (95% CI, 54% to 79%) among patients in arm A and 60% (95% CI, 48% to 72%) among patients in arm B (P = .45; Fig 3⇓).
Clinical Results in 160 Patients
Failure-free survival in the high-dose chlorambucil plus prednisone group (HD-CHL-P, 77 patients) versus the high-dose chlorambucil plus prednisone and epirubicin group (HD-CHL-PE, 83 patients).
Progression-free survival in the high-dose chlorambucil plus prednisone group (HD-CHL-P, 60 patients) versus the high-dose chlorambucil plus prednisone and epirubicin group (HD-CHL-PE, 55 patients).
Overall survival in the high-dose chlorambucil plus prednisone group (HD-CHL-P, 77 patients) versus the high-dose chlorambucil plus prednisone and epirubicin group (HD-CHL-PE, 83 patients).
Toxicity
Information about toxicity was available for 150 patients. Most of the side effects in both groups were grade 1 or 2, but there were significantly more cases of grade 3 or 4 hematologic toxicity among the patients treated with HD-CHL-PE (Table 3⇓). On the basis of the available data, three toxic deaths were recorded among the 160 patients (1.8%; 95% CI, 0% to 4%); two patients in arm A died of acute toxic hepatitis and bacterial infection, respectively, and one patient in arm B died of cardiac failure.
Toxicity (N = 150)
IFNα Maintenance Treatment
Of the 115 responders to induction therapy, 88 (40 CRs and 48 PRs) were randomly assigned to maintenance treatment; 43 patients were assigned to receive IFNα (20 patients from arm A and 23 patients from arm B), and 45 patients were assigned to observation (27 patients from arm A and 18 patients from arm B), without any difference in terms of the type of response to induction therapy. The 3-year PFS rates were 44% in the IFNα group (95% CI, 26% to 62%) and 42% in the observation group (95% CI, 27% to 58%; P = .90; Fig 4⇓). Neither the clinical characteristics of the patients nor the randomized induction treatment showed any correlation with the duration of response in the two groups (data not shown).
Progression-free survival in the interferon alfa group (43 patients) versus the observation group (45 patients).
Prognostic Factors
We tried to assess whether any clinicobiologic parameter could influence disease outcome by making a univariate analysis of all of the variables listed in Table 1⇑, with the addition of β2-microglobulin levels, bone marrow infiltration patterns, splenomegaly, and the presence of a serum monoclonal component. CR and OR rates, and OS, FFS, and PFS were considered as end points. None of the evaluated parameters showed any correlation with response rates, but an age of more than 60 years (P = .014), anemia (P = .002), and high serum lactate dehydrogenase levels (P = .024) significantly correlated with OS (data not shown). Furthermore, the International Prognostic Index showed a correlation with survival when low/low-intermediate–risk patients were compared with high-intermediate/high–risk patients (P = .02). However, multivariate analysis showed that only age (P = .001) and anemia (P = .01) significantly correlated with OS (Fig 5⇓).
Overall survival by age and the presence of anemia (hemoglobin < 11g/dL).
Finally, to validate the trial results after considering the current lymphoma classifications that were not available when the study was started, we centrally reviewed the pathologic diagnoses of the 111 patients for whom diagnostic material was available with the aim of classifying the disease according to the recent WHO/REAL classification.25 We could also verify the diagnostic criteria in the group of 32 patients registered on the basis of immunocytomorphologic criteria. An additional analysis of the data relating to the 123 patients with a confirmed diagnosis of indolent nonfollicular NHL (91 with confirmed histology according to the WHO/REAL classification plus 32 with an immunocytomorphologic diagnosis of CD5-negative mature B-cell leukemia) did not change the study results in terms of the defined end points. In particular, no differences between the two treatment arms or between the different histotypes emerged (data not shown).
DISCUSSION
Our findings show that oral chlorambucil as a single agent can still be considered a very active drug in the initial treatment of INFL and that the addition of epirubicin to a pulsed high-dose chlorambucil schedule does not add any therapeutic advantage and leads to more severe toxicity. The CR (34% v 25%, respectively) and OR rates (78% v 66%, respectively), as well as PFS, OS (5-year OS, 66% v 60%, respectively), and FFS (5-year FFS, 31% v 16%, respectively) were similar in arms A and B.
Many published studies have demonstrated the efficacy of intensified doses of chlorambucil in comparison with low-dose regimens in the treatment of patients with CLL and low-grade FL. Many intensified schedules have been used for CLL patients, including continuous chlorambucil (15 mg/d until response or toxicity)26 and pulsed chlorambucil,27 and have led to high response rates.26,28 In CLL, high-dose continuous chlorambucil has been prospectively compared with Binet’s modified CHOP regimen, leading to a higher OR rate (89.5% v 75%, respectively; P < .001) and longer OS time (median OS, 68 v 47 months, respectively; P < .005).29
Only few data have been published about INFLs. In 1993, Dana et al13 published a retrospective analysis of 415 patients involved in three different Southwest Oncology Group trials of a full-dose CHOP regimen. Although the analysis included many patients with FLs, a subset analysis of a group of 83 patients with SLLs showed a median OS time of 5.1 years, thus confirming the uselessness of a more toxic regimen.13 In their recent article, Summerfield et al30 described the efficacy of high-dose oral chlorambucil (30 mg/d for 4 days for eight courses) in patients with CLL and in a small group of eight patients with INFL.
This study is the first and largest randomized trial showing that high-dose chlorambucil plus prednisone can be considered the standard front-line treatment for INFLs and could, therefore, represent the control arm in future randomized studies. We have compared our findings with the interesting and promising data emerging since the advent of purine analogues. Foran et al14 recently reported the results of a phase II trial of fludarabine as a single agent for the treatment of patients with INFLs and mantle-cell lymphomas. The study involved 57 patients with a diagnosis of LPC (37 patients) or Waldenström’s macroglobulinemia (20 patients) and showed OR rates of 63% and 79%, respectively. Median OS time was not reached in the survival analysis and was longer than 2.9 and 1.9 years for patients with LPC and Waldenström’s macroglobulinemia, respectively.14 On the basis of this study of a relatively small number of patients, the results of fludarabine as a single agent are comparable with those obtained with HD-CHL-P in our trial. A number of authors have recently reported promising results with fludarabine alone or in combination with other agents in the treatment of indolent lymphomas. In particular, in the phase III trial by Zinzani et al15 analyzing the efficacy of fludarabine alone or fludarabine plus idarubicin in the treatment of low-grade lymphomas, a group of patients with INFL were included (38 patients with SLL and 29 with IC). The authors concluded that the combination regimen might be superior to fludarabine alone in terms of CR rate in the subgroup of SLL and IC patients. Similarly, promising therapeutic results recently have been reported with the combination of fludarabine and cyclophosphamide in CLL patients and in a small series of INFL patients.16,31 Therefore, a possible role of the association of fludarabine and alkylators has been suggested in INFL, but the combination also has a relevant toxicity profile.28 Thus, it is highly recommended that treatment modalities in phase III trials should be carefully compared by evaluating not only response rates, but also the tolerability and toxicity of the treatments.
Our trial included a second randomization designed to test the hypothesis that 1-year IFNα maintenance treatment may prolong disease-free survival. Given the small number of patients, the statistical power of our tests does not allow any definite conclusions to be drawn, but the results seem to confirm that 1-year treatment with IFNα at the dose used in our trial does not improve the outcome of INFL patients responding to induction therapy. Similar conclusions also recently have been reached in a large Southwest Oncology Group phase III study.32
Finally, the analysis of prognostic factors showed that only age and anemia were independent parameters significantly influencing survival. The prognostic role of age can be explained by the higher lymphoma-independent mortality rate of elderly patients. Furthermore, the prognostic power of disease-related anemia confirms the biologic similarity of INFL and classic CLL, and thus indicates that a modified IPI should be designed for this subset of indolent NHLs. Interestingly, 44 of our patients had a lymphocyte count of more than 5,000/μL, 32 of whom were diagnosed immunocytomorphologically (CD5 negativity excluded CLL or MCL). Although the number of patients is limited, there was no difference in response and survival between these patients and those with prevalently nodal disease, thus confirming that clinical behavior is not affected by leukemic expression (data not shown). This prospectively obtained finding in a series of patients with the spectrum of INFLs is in line with those findings recently published relating to a large retrospective study of patients with lymphomas of marginal-zone origin.10
Our results are also confirmed by the subset analysis of different histotypes, but because the trial was not designed and powered for this end point, no firm conclusions can be drawn. When the trial was planned, there were initial data indicating clinicobiologic differences among the INFLs (particularly the forms of MZL), and reliable differential diagnostic criteria were beginning to be found. However, we decided to group all of these diseases together for the following reasons: first, because the most important point in terms of inclusion criteria was to exclude FLs from the group of low-grade lymphomas as a whole; second, the primary objective of the study was to evaluate the role of the addition of epirubicin in the outcome of INFLs; and third, to avoid dealing with small subgroups of patients.
In conclusion, the results of the present study definitively confirm the hypothesis that epirubicin does not add any advantage in the treatment of INFL and that HD-CHL-P can be considered the standard conventional treatment for advanced INFL in terms of efficacy, toxicity, and treatment costs. Given the promising results obtained using purine analogues in combination regimens in INFL, a phase III trial is warranted to compare their efficacy with that of the old but inexpensive chlorambucil.
APPENDIX
The following institutions participated in this study: F. Angrilli, Ematologia Pescara; V. Pitini, Istituto di Oncologia, Policlinico Universitario, Messina; E. Naglieri, Istituto Oncologico, Bari; G. Giglio, Azienda OspedaleCardarelli, Campobasso; S. Molica, Divisione di Oncologia, Ospedale “A. P. Ciaccio,” Catanzaro; S. Morandi, Istituti Ospitalieri, Cremona; L. Dezza, Azienda Ospedale di Circolo, Melegnano; E. Iannitto, Divisione di Ematologia, Palermo; M. Lopez, Ospedale Regina Elena, Roma; G. Torelli, Divisione di Ematologia, Università di Modena e Reggio Emilia; G. Partesotti, Divisione di Medicina, Ospedale Sassuolo, Modena; L. Ginaldi, Università di l’Aquila-Teramo, Teramo; A. Riezzo, Divisione di Ematologia, Ospedale Civile “S. Nicola Pellegrino,” Trani; and G. Attardo Parrinello, Divisione di Medicina, Ospedale Civile, Vigevano, Italy.
Acknowledgments
We thank Dr A. Sirotti (Dipartimento di Oncologia ed Ematologia, Università di Modena e Reggio Emilia, Modena, Italy) for her support in data collection and management.
Footnotes
-
Supported by a grant (Ricerca Corrente) from the Italian Ministry of Health to Ospedale Maggiore Istituto Ricovero Cura Carattere Scientifico, Milan, Italy, and a grant from Associazione Angela Serra, Modena, Italy.
- Received August 22, 2002.
- Accepted January 7, 2003.
