Predictors of response to reinduction chemotherapy for patients with acute myeloid leukemia who do not achieve complete remission with frontline induction chemotherapy

Eighty-one patients with acute myeloid leukemia who had persistent leukemia following standard induction therapy with cytarabine plus daunorubicin (7+3 regimen) underwent reinduction therapy with a combination of mitoxantrone, etoposide, and high-dose cytarabine (HiDAC). Patients achieving complete remission (CR) then received consolidation therapy with HiDAC plus mitoxantrone. Patients with matched sibling donors were referred for allogeneic bone marrow transplantation (BMT) in CR-1. The overall response rate to reinduction was 53%. The major adverse predictors of CR on multivariate analysis were poor risk cytogenetics, a higher % bone marrow blasts prior to reinduction therapy and increased age. The median relapse-free survival (RFS) was 9 months and the estimated 2-year RFS was 30%. No significant predictors of RFS or overall survival (OS) were found among the patients achieving CR. Patients undergoing allogeneic BMT in CR-1 after double induction had a 50% 2-year OS. Patients relapsing after achieving CR with double induction had a poor outcome with a 4% 1-year OS. The results indicate that patients with poor risk cytogenetics or marrow blast percentage >or= 60% following 7+3 induction have a low probability of achieving CR with reinduction and should be considered for novel approaches to improve CR rates. Patients achieving CR are at high risk of relapse and should be considered for allogeneic BMT or novel strategies to attempt to reduce relapse rates.     

Outcome of 93 patients with relapse or progression following allogeneic hematopoietic cell transplantation

Relapse/progression after allogeneic hematopoietic cell transplantation (allo‐HCT) remains the major cause of treatment failure. In this study, the subsequent clinical outcome was overviewed in 292 patients with leukemia/myelodysplastic syndrome who received allo‐HCT. Among them, 93 (32%) showed relapse/progression. Cohort 1 was chosen to receive no interventions with curative intent (n = 25). Cohort 2 received reinduction chemotherapy and/or donor lymphocyte infusion (n = 48), and Cohort 3 underwent a second allo‐HCT (n = 20). Sixty‐three patients received reinduction chemotherapy, and 27 (43%) achieved subsequent complete remission (CR). The incidence of nonrelapse mortality (NRM) was similar among the three cohorts (4, 15, and 5%). The 1‐year overall survival (OS) after relapse was significantly better in patients with a second HCT (58%) than in others (14%, Cohorts 1 and 2; P <.001). However, the 2‐year OS did not differ between the two groups, which suggests that it is difficult to maintain CR after the second HCT. Multivariate analysis showed that reinduction chemotherapy, CR after intervention, second HCT, and longer time to post‐transplant relapse were associated with improved survival. In conclusion, for patients with relapse after allo‐HCT, successful reinduction chemotherapy and a second HCT may be effective for prolonging survival without excessive NRM. However, effective measures to prevent disease progression after a second HCT clearly need to be developed. Am. J. Hematol. 2009. © 2009 Wiley‐Liss, Inc.     

Duration of second complete remission in patients with acute myeloid leukemia treated with chemotherapy: a retrospective single-center study

 A total of 168 patients with de novo AML were retreated with chemotherapy at relapse following first CR; 66 patients (39%) achieved a second complete remission (CR). The probability of achieving a second CR was highly dependent on the duration of the first remission. Patients who received no or conventional postremission chemotherapy after second CR had a median remission duration of 7.5 months, and the probability of remaining in remission at 3 years was 24%. Patients with a first CR of more than 12 months had a median second remission duration of 18 months. The probability of a second CCR was 35% at 3 years and 24% at 5 years, whereas none of the patients with a first CR of less than 12 months was in remission at 3 years. Only a poor correlation (p=0.31) was found when the durations of the first and second CR were compared in patients with a second relapse. Patients with long-lasting remissions and long-term survivors after second CR are characterized by a first CR duration of >12 months and favorable or normal cytogenetics. The type of salvage treatment seems to be less important for achievment of long-term remission, but it is probably important to administer consolidation chemotherapy after second CR. Other so-far ill-defined factors may be responsible for the supression of the leukemic clone in patients with long-lasting remissions following chemotherapy for relapse after second CR. Received: 28 December 1995 / Accepted: 4 January 1996     

Combination chemotherapy for haematological relapse in adult acute lymphoblastic leukaemia (ALL).

Twenty-three adult patients (median age 22 years) with haematological relapse of acute lymphoblastic leukaemia were given reinduction therapy with vincristine, adriamycin, prednisolone, and L-asparaginase. Complete remission (CR) was achieved in 16 patients (69%), and only one patient was completely unresponsive to therapy. The duration of second remission was short except in three patients who had experienced long first remissions (> 2.5 years). The median duration of survival from first relapse was seven months.     

Prognostic factors at relapse for adults with acute myeloid leukemia

To identify prognostic factors for patients with acute myeloid leukemia in first relapse we have analyzed the courses of 57 patients undergoing intensive reinduction therapy. The median duration of first remission was 9 months. A second complete remission was achieved by 31 patients (54%). The median duration of second remission was 5 months (range 1-142 + months). Two patients are still disease-free with 42 months and 142 months follow-up. Clinical, laboratory, and treatment variables at the time of original diagnosis and at relapse were examined for possible associations with the outcome of reinduction chemotherapy. Only duration of first remission consistently correlated with ability to achieve a second remission, second remission duration, and overall survival from reinduction. The second complete remission rate was 69% for those with a first remission longer than 9 months but only 39% for those with a first remission less than 9 months (P less than 0.05). The patients with longer first remissions also had longer second remissions (8 months median vs. 4 months, P = 0.02) and a longer median survival following reinduction chemotherapy (8 months vs. 5 months, P = 0.02). We conclude that the duration of a patient's first remission should be considered a useful prognostic factor both in treatment planning and in the interpretation of clinical trials of patients in relapse.     

Allogeneic bone marrow transplantation after high-dose busulfan and cyclophosphamide in patients with acute nonlymphocytic leukemia

Ninety-nine patients with acute nonlymphocytic leukemia (ANLL) received HLA-identical bone marrow transplants (BMTs) from sibling donors after preparation with high doses of busulfan and cyclophosphamide. Forty- nine patients were transplanted in first complete remission (CR), and 50 patients were transplanted in second and third CR and early relapse. Fifty-three received one of three regimens containing primarily low- dose cyclophosphamide (group I) for graft-v-host disease (GVHD) prophylaxis; since March 1983, 46 patients received intravenous (IV) cyclosporine (group II). After December 1983, only cytomegalovirus (CMV)-seronegative blood products were used in appropriate patients, and since April 1984 patients seropositive for herpes-simplex virus (HSV) and CMV received high-dose acyclovir prophylaxis. For patients transplanted in first CR, there was a significantly lower incidence of acute GVHD (P = .005) and deaths related to GVHD and interstitial pneumonitis (P = .001) in patients in group II. This was reflected in an improved Kaplan-Meier probability of disease-free survival (DFS) in the 22 patients transplanted in group II as compared with the 27 patients in group I (64% +/- 10% v 30% +/- 9%, P = .017). The probability of remaining in remission was slightly lower in group II (82% +/- 9% v 94% +/- 6%, P = .479). For patients transplanted in second and third CR and early relapse, the incidence of acute GVHD (P = .026) and deaths related to GVHD and interstitial pneumonitis was significantly lower in group II (P = .029); the probability of remaining in remission was also less (47% +/- 15% v 91% +/- 15%, P = .022). However, the probability of DFS was not significantly different between the two groups (26% +/- 10% v 35% +/- 18%, P = .957). We conclude that transplantation for patients in first CR who received IV cyclosporine therapy is effective treatment; patients with more refractory disease treated with the same cyclosporine regimen (group II) had a lower incidence of GVHD than those treated in group I, but survival did not improve because of an increase in the number of relapses and other nonleukemic complications.     

Survival after cord blood transplantation from unrelated donor as a second hematopoietic stem cell transplantation for recurrent pediatric acute myeloid leukemia

The Japan Cord Blood Bank Network (JCBBN) reports the treatment of 22 children with acute myeloid leukemia (AML) who received umbilical cord blood transplantation from unrelated donors (CBT) as their second hematopoietic stem cell transplantation (HSCT). Provided by the JCBBN, between February 1997 and September 2006, 22 patients had CBT as a second HSCT. In the initial HSCT, eight received autologous, seven received CBT, and the remaining had allogenic BMT. At the time of CBT as a second HSCT, seven were in the second complete remission (CR2), two in the third CR (CR3), the remaining were not in remission. Reduced intensity conditioning (RIC) conducted for 10 cases and myeloablative conditioning (MAC) for 12 cases. The overall survival rate was 31.3%, 5 years after CBT. Second complete remission at second transplantation was favorable prognosis (58.3 ± 18.6%, compared with 17.1 ± 10.8% for the non-CR group. Mortality after CBT as a second HSCT accounted for 15 cases, 8 from treatment-related mortality. In conclusion, CBT combined with RIC as second HSCT may be useful against a recurrence of AML in children after the initial HSCT.     

Prognostic factors in a multicenter study for treatment of acute lymphoblastic leukemia in adults

In a prospective multicenter study, 368 acute lymphoblastic leukemia (ALL) patients aged 15 to 65 years were treated with an intensified induction and reinduction regimen; 272 (73.9%) achieved complete remission (CR). The median remission duration (MRD) is 24.3 months, and the probability of being in continuous CR (CCR) at greater than 5 years is .37. The median survival for all 368 patients is 27.5 months, and the probability of being alive at 5 years is .39. For the 272 patients in remission the median survival is 58.4 months, and the probability of being alive at 5 years is .49. A lower CR rate was seen for patients with bleeding at diagnosis or with splenomegaly/hepatosplenomegaly. The prognostic factors unfavorable for remission duration were time to CR greater than 4 weeks v less than 4 weeks (P = .0002), age greater than 35 years v less than 35 years (P = .0008), leukocyte count greater than 30,000/microL v less than 30,000/microL (P = .0112), and null ALL v common ALL (c-ALL)/T cell ALL (T-ALL) (P = .05). The remission duration correlated strongly (P = .0001) with the number of these independent prognostic factors. In patients with none of these adverse factors the MRD has not yet been reached, with one adverse factor the MRD is 21.9 months, and with two or three adverse factors the MRD is only 9.6 months. For the immunologic subtype T-ALL, the probability of being in CCR at greater than 5 years is .55; for c-ALL, .34; and for null ALL, .24. According to these results, patients were stratified into a low- risk group with a CCR rate of .62 and a high-risk group with a CCR rate of .28, with the latter now allocated to either further chemotherapy or bone marrow transplantation in first remission.     

Autologous bone marrow transplantation for childhood acute lymphoblastic leukaemia in second remission - long-term follow-up

From 1984 to 1996, 31 consecutive children without sibling donors, aged 5-19 years (median 8) with acute lymphoblastic leukaemia (ALL) in second complete remission (CR), received unpurged autologous bone marrow transplantation (ABMT) after melphalan and single fraction total body irradiation (TBI). ABMT was performed using fresh unmanipulated marrow harvested after standard reinduction and consolidation therapy 2-11 months (median 5) after relapse. With a median survival of 2.9 years the probability of survival for all patients in continuing second CR was 45.1% (95% CI, 24%-62%) after 5 years. Regimen-related and non-leukaemia mortality was 7% (95% CI, 2%-26%). The longest time to second relapse from ABMT was 3.1 years. Pituitary and gonadal dysfunction requiring hormonal replacement therapy occurred in the majority of long-term survivors. Twelve patients developed cataracts. ABMT with melphalan/single fraction TBI has proved an effective anti-leukaemia treatment with low regimen-related mortality but significant long-term morbidity. The current approach of allogeneic BMT from an unrelated donor when no sibling donor is available, following conditioning with cyclophosphamide/ fractionated TBI has resulted in a reduced relapse rate and improved short-term overall survival in the treatment of relapsed childhood ALL. However, long-term results are awaited.     

Outcome of a multicenter treatment program including autologous or allogeneic bone marrow transplantation for de novo acute myeloid leukemia

Abstract: The results of an intensive treatment program for patients 16–60 yr of age with de novo acute myeloid leukemia are presented. The patients were given conventional induction treatment with daunorubicin and cytarabine. Patients not entering complete remission (CR) after 1 course of daunorubicin/cytarabine were given 1 course of amsacrine/etoposide/cytarabine. Those entering complete remission received 3 consolidation courses using mitoxantrone, etoposide, amsacrine and cytarabine. One hundred and eighteen patients were enrolled. Complete remission was attained after 1–2 courses in 90 patients (76%). Another 6 patients reached CR after 3–4 induction courses for a total CR rate of 81%. If feasible, patients were offered either allogeneic or unpurged autologous bone marrow transplantation. Twenty-four patients underwent allogeneic bone marrow transplantation; 15 in first remission, 8 in second remission, 1 in early relapse. Thirty patients below 56 yr of age underwent autologous bone marrow transplantation in first remission. The overall probability of survival at 4 yr was 34%, and for patients below 40 yr of age 50%. Leukemia-free survival was 35% for the whole cohort of patients; 52% for patients below 40 yr of age. Patients undergoing allogeneic or autologous bone marrow transplantation in first remission had an overall survival of 86% and 47%, respectively, while the probability of leukemia-free survival in these groups was 87% vs. 40% at 4 yr. The CR rate and long-term results of this intensive treatment program compare favorably with other recent studies using intensive consolidation with allogeneic or autologous bone marrow transplantation or high dose cytarabine.     

A prediction model for complete remission upon reinduction for patients with acute myeloid leukemia after failure of anthracycline and cytarabine standard chemotherapy

The aim of this study was to investigate clinical parameters that might predict complete remission (CR) following reinduction therapy in patients with acute myeloid leukemia (AML). We retrospectively analyzed outcomes in 142 patients who failed to achieve CR with standard anthracycline-plus-cytarabine induction chemotherapy (IND1) and who received the same regimen as a reduction therapy (IND2). The CR rate after reinduction was 54%. Multivariate analysis showed that the presence of peripheral blood (PB) blasts on the day of commencement of IND2 and ≥20% blasts in an interim BM sample taken during IND1 (C1-INT-BM) were independent risk factors for failure of remission. Our scoring system for prediction of CR after reinduction (CRAR score) included a favorable chromosome risk group, absence of PB blasts on the day of commencement of IND2, and ≤21% blasts in the C1-INT-BM. This scoring system predicted that the rates of CR in the four subgroups would be 92.1%, 68.9%, 29.5%, and 7.3%, respectively, in good agreement with observed CR rates. The CRAR scoring model might be associated with the possibility of achieving CR after reinduction, and may be helpful in choosing alternative strategy for AML patients refractory to standard induction chemotherapy, although further prospective validation is required.     

Reinduction therapy in 297 children with acute lymphoblastic leukemia in first bone marrow relapse: a Pediatric Oncology Group Study

Many children with acute lymphoblastic leukemia (ALL) develop a marrow relapse during or shortly following initial continuation chemotherapy. Achievement of a second complete remission is the initial step in a successful retreatment effort. Reinduction results using two or three drugs have been unsatisfactory, and previous reports of four-drug reinduction programs have included relatively small numbers of patients. Pediatric Oncology Group protocol 8303 was designed for patients with ALL in first marrow relapse during or within 6 months after cessation of chemotherapy. The results of reinduction therapy in 297 study patients are described here. Four-drug reinduction therapy consisted of daily oral prednisone, weekly vincristine and daunorubicin, and asparaginase three times weekly for 4 weeks (PVDA). CNS retreatment consisted of two doses of triple intrathecal chemotherapy. Of the 297 patients receiving reinduction, 245, or 82%, entered second complete remission, six died of infection or progressive disease, and 46 others still had M2 or M3 bone marrow status. Forty of these latter patients received four doses (during a 2-week period) of teniposide and cytarabine, after which 13 (32%) achieved complete remission status. Thus, the overall second complete remission rate with PVDA with or without teniposide/cytarabine was 258 of 297, or 87%. The treatment program was generally well tolerated. Among the numerous factors analyzed by using logistic regression, only female sex (P = .035), the presence of blasts on the blood smear at the time of relapse (P = .0002), and a length of initial complete remission less than 12 months (P = .021) were independent predictors of failure to enter second remission. We conclude that the intensive reinduction program described here is a highly effective first step in the delivery of salvage therapy to patients with ALL in first marrow relapse. The current challenge is to develop improved continuation treatment for these children.     

Prolonged disease-free survival in Hodgkin's disease with MOPP reinduction after first relapse.

Thirty-two patients with Hodgkin's disease who relapsed after a first complete remission induced by nitrogen mustard, vincristine, procarbazine, and prednisone (MOPP) were retreated with MOPP chemotherapy. Nineteen patients achieved a second complete remission. Median duration of the second complete remission was 21 months. The likelihood of achieving a second complete response could be predicted by the duration of the first response. Fourteen of 15 patients whose first complete remission was longer than 12 months achieved a second complete response in contrast to five of 17 patients whose initial complete remission lasted less than 12 months (P less than 0.001). Median survival of all patients in this study who were re-treated with MOPP was longer than 4 years after their first relapse. We conclude that patients with Hodgkin's disease who relapse after a first complete remission induced by MOPP are not necessarily resistant to further MOPP therapy and can achieve long-term survival with MOPP reinduction.     

Acute lymphoblastic leukaemia in adults in Sweden 1977-84: a retrospective analysis. Swedish ALL-Group

The present study is a retrospective analysis of the outcome in 210 patients diagnosed and treated as having acute lymphoblastic leukaemia (ALL) in Sweden during 1977-84. 131 patients were morphologically rediagnosed as ALL. For the ALL-patients, nine different remission induction regimens were used. Remission frequency was 69%, without statistical difference according to induction treatment. However, the reasons for remission failure differed among therapy groups. The number of responders was significantly higher among patients who received a remission induction therapy with an anthracycline and/or L-asparaginase. Maintenance therapy consisted in most cases of 6-mercaptopurine and methotrexate with reinduction courses for 2-3 years. Median survival time was 13 months and median duration of first remission (MRD) 11 months. For a subgroup of patients (n = 29) treated with the most intense remission induction regimens, including at least 4 cytostatic drugs with both an anthracyclilne and L-asparaginase, the MRD is not yet reached, the shortest follow up time is 43 + months, and the probability of remaining in complete remission is 66%. We conclude that aggressive cytostatic therapy, with induction regimens including both an anthracycline and L-asparaginase, may cure a considerable number of adult ALL-patients.     

The role of intensive remission induction and consolidation therapy in patients with acute myeloid leukaemia

Sixty-one patients with AML, 59 adults and two children, were treated with intensive remission induction and consolidation therapy. The median age was 36 years. Forty-four (72%) patients entered complete remission (CR); 11 patients received a bone marrow transplantation. The median survival of complete remitters was 26.5 months; the probability of remaining in CR at respectively 1 and 2 years was 75% and 62%. The only factor significantly correlated with the outcome of remission induction, survival and duration of CR was age. Patients less than 30 years fared significantly better than those 30 years or older; no difference in outcome was observed between patients aged 30-50 and those over 50 years. In patients less than 30 years the CR rate was 95%; 75% of them were still alive at 2 years and only one (5%) has relapsed. In contrast, in patients 30 years or older the CR rate was 60% and the median survival only 11.5 months, 50% of the complete remitters in this age group have relapsed. Morbidity from intensive consolidation therapy was considerable; more than 50% of consolidation courses were complicated by high fever, needing urgent admission; only four (3%) courses had a fatal event. It is concluded that intensive consolidation therapy may be considered as a major advance in the treatment of younger patients with AML, while its role in older individuals remains questionable. A possible explanation for the completely different outcome in younger and older patients with AML is discussed.     

Recurrent childhood anaplastic large cell lymphoma: a retrospective analysis of registered cases in Japan

This report presents a retrospective study of 26 Japanese children with recurrent anaplastic large cell lymphoma. The first relapses were documented at a median of 10.5 months after the initial diagnosis. Twenty-four patients achieved a second remission. After a median follow-up period of 47 months, 18 patients are still alive: 15 patients are in second complete remission (CR), three patients are in third CR or later. The 5 year overall and relapse-free survival rates were 61 +/- 12% and 51 +/- 12% respectively. The patients who received allogeneic haematopoietic stem cell transplantation during second CR showed a superior outcome to other patients.     

Bone marrow transplantation for acute lymphoblastic leukaemia: a survey of the European Group for Bone Marrow Transplantation (E.G.B.M.T.)

Between 1979 and 1982, 192 patients with acute lymphoblastic leukaemia were given allogenic bone marrow transplants from HLA-identical siblings. Data on 5 7 patients transplanted in first remission were compared with the results in 96 patients transplanted in second remission, and 39 in third or subsequent remission. The majority of these patients were of the non-B non-T-ALL and T-ALL subtypes. The 2-year actuarial survival for non-B non-T-ALL was 58% for patients transplanted in first remission, 34% for second remission patients and 33% for patients transplanted in third or subsequent remission ( P = O60). For patients with T-ALL, the survivals at 2 years for first and second remission transplant patients were 61% and 10%, respectively ( P =0-04). Multifactorial analysis demonstrated a significantly higher probability of survival for patients grafted in first remission (compared with more advanced remission states) and with bone marrow from young (age ≤ 20 years) donors. The 2-year actuarial risk of relapse for patients with non-B non-T-ALL transplanted in first remission was 0%, 32% and 69% for second and third or subsequent remission patients, respectively ( P = 0-04). For T-ALL the actuarial risk of relapse at 2 years is 10% for first remission grafts and 48% for second remission grafts ( P = 0-07). Only patients (with both non-B non-T-/and T-ALL) transplanted in first remission have a high and stable probability of remaining in remission.     

BAVC regimen and autologous bone marrow transplantation in patients with acute myelogenous leukemia in second remission

Twenty-one acute myelogenous leukemia (AML) patients were submitted in second remission (II CR) to BAVC conditioning regimen followed by unpurged ABMT. Transplant was done after a median of 2 months from II CR (range, 1 to 13). Median first remission (I CR) duration was 16 months (range, 1-35). Conditioning regimen was well tolerated, with no major extra-medullary toxicity. One patient died during aplasia from fungal sepsis. Of the 20 evaluable patients, nine relapsed after a median time of 6 months (range, 2 to 18). Eleven patients are in continuous complete remission (CCR) after a median follow-up of 40 months (range, 24 to 63). The duration of II CR has exceeded the duration of I CR in all patients in CCR. Projected probability of disease-free survival is 52% at 63 months.     

Allogeneic SCT in refractory or relapsed adult ALL is effective without prior reinduction chemotherapy

We present 60 patients with refractory (n=8) or relapsed (n=52) adult ALL who received allogeneic hematopoietic SCT (HSCT) with (n=41) or without (n=19) prior reinduction chemotherapy. In our center, omission of reinduction is recommended if a suitable donor is promptly available, tumor burden is moderate and disease features suggest a highly aggressive course. Overall survival (OS) of the whole cohort at 1, 2 and 5 years was 42, 33 and 28%, respectively. Leukemia-free survival at 1, 2 and 5 years was 37, 33 and 24%. Deaths were due to relapse (n=25), acute or chronic GVHD (n=7), infections (n=8) or toxicity (n=4). Interestingly, patients who did not receive reinduction before HSCT had better outcomes than patients who received reinduction with OS at 1, 2 and 5 years being 58 vs 34%, 47 vs 25% and 47 vs 18%, respectively (P=0.039). Importantly, even achievement of a second CR after reinduction was not associated with improved survival compared to patients directly proceeding to HSCT. We conclude that patients who undergo HSCT for refractory or relapsed ALL can achieve long-term survival. In selected patients, reinduction chemotherapy can be omitted if immediate HSCT is feasible.     

Autologous bone marrow transplantation for acute myeloid leukemia using monoclonal antibody-purged bone marrow

We report our experience from a clinical trial of autologous bone marrow transplantation (ABMT) in the treatment of 30 patients with acute myeloid leukemia (AML) using monoclonal antibody (MoAb) and complement-treated bone marrow. All patients were in complete remission (CR) at the time of transplant: 6 patients were in first CR, 18 in second CR, and 6 in third CR. The median age of all patients was 42 years (range 11 to 57 years). For marrow ablation, 28 patients were treated with cyclophosphamide and total body irradiation. One patient was treated with busulfan and cyclophosphamide and one was treated with busulfan and VP-16. Each patient was then transfused with autologous bone marrow that had been harvested previously and treated with two MoAbs, PM-81 and AML-2-23, and rabbit complement. Median time to recovery of neutrophils (500/microL) was 30 days, and platelets (20,000/microL) was 45 days. Median time for initial erythrocyte engraftment, assessed by a flow cytometric reticulocyte assay, was 13 days. Median overall and relapse-free survival of first CR patients was at least 17.4 months post-ABMT and the 2- and 3-year actuarial overall and relapse-free survival was 67% (+/- 19%). Median survival for the 24 patients in second or third CR was 6.8 months post-ABMT and 9.3 months since CR; however, six patients survived disease-free from 16 to 61 months post-ABMT. For the second and third CR group it was observed that six patients (5 of the 6 survivors) showed "inversions," when their post-ABMT remission lasted longer than any previous one. Actuarial 2- and 3-year disease-free and overall survival of patients in second and third CR was 25% (+/- 9%) and 18% (+/- 9%), and 29% (+/- 9%) and 23% (+/- 9%), respectively. ABMT avoids the problems of graft-versus-host disease and of finding suitable donors for allogeneic marrow transplantation.