Objectives: Our aim was to identify factors associated with overall survival and the efficacy of postrelapse treatment protocols and to determine whether pretransplant consolidation therapy and minimal residual disease status pose a survival benefit.
Materials and Methods: Patients with acute myeloid leukemia who underwent stem cell transplant between 2007 and 2018 were enrolled retrospectively. The effects of pretransplant cytogenetic and minimal residual disease status, pretransplant consolidation therapies, development of graft-versus-host disease, postrelapse treatment protocols, and type of con-ditioning regimens on overall survival were analyzed.
Results: In 76 study patients, the cumulative overall 1- and 5-year relapse probabilities were 67.8% and 58.7%, respectively. Overall survival rates at 3 and 5 years in patients with and without relapse were 23.5% and 0% and 95.9% and 91.1% (P < .001), respectively. Although mean postrelapse overall survival was better with intensive salvage plus donor lymphocyte infusion, no significant differences were shown versus other therapies (intensive salvage, nonintensive salvage, intensive salvage or nonintensive salvage plus donor lymphocyte infusion, or supportive therapy). Twenty-three patients (30.3%) died during the study period with a median survival of 9.6 months. Patients with favorable, intermediate, and unfavorable cytogenetic status showed overall survival of 46.6 ± 10.4, 54.6 ± 4.4, and 36.9 ± 5.9 months (P = .807). Patients with and without minimal residual disease and patients who received or did not receive consolidation therapy had similar overall survival. Relapse was an independent predictor of overall survival (increased mortality risk of 26.22). Patients who developed graft-versus-host disease showed decreased relapse.
Conclusions: Relapse is the most important predictor of overall survival and is associated with poor prognosis. Pretransplant minimal residual status and cytogenetic status showed no effect on relapse rates and overall survival, and consolidation therapy did not improve outcomes.
Key words : AML, Consolidation, Postrelapse treatment, Relapse
Acute myeloid leukemia (AML) is a clonal neoplasm derived from progenitor cells with varying outcomes.1 The initial goal is achieving complete remission (CR); however, without treatment after remission, most patients relapse.2 Although allogeneic stem cell transplant (allo-SCT) is an effective treatment option, even after allo-SCT, a significant number of patients (about 30%-40%) will eventually relapse.3 Relapse remains the major cause of treatment failure. Unfortunately, therapeutic strategies are limited, and there is no standardization of further therapy once relapse occurs.4 Thus, the prevention of relapse is important to improve survival rates, and novel strategies should be developed for this purpose.
To prevent relapse and increase survival, it is important to determine which patient should receive a transplant during the first remission and whether consolidation should be applied before allo-SCT. Consolidation therapy is known to be effective if a transplant procedure is not planned.5,6 However, controversy remains with respect to the need for consolidation chemotherapy before allo-SCT, exclu-ding when a donor has not yet been found. Despite evidence showing that consolidation therapy dec-reases the leukemic cell burden before transplant and may reduce relapse, the development of early relapse or toxicity due to consolidation therapy can prevent subsequent allo-SCT.5 To address this concern, we retrospectively evaluated 76 AML patients who were diagnosed and transplanted in our clinic. We explored the factors associated with overall survival (OS), the efficacy of postrelapse treatment protocols, and whether there is a survival benefit with pretransplant consolidation therapy and minimal residual disease (MRD) status.
Materials and Methods
All patients who were diagnosed with AML and underwent hematopoietic stem cell transplant (HSCT) between 2007 and 2018 at the Adana Baskent University Hospital Hematology Department were enrolled in this retrospective, single-center cohort study. In our center, the criteria for transplant are as follows: having a matched and suitable donor, having a good Eastern Cooperative Oncology Group performance status (0-1 status), being refractory to chemotherapy, and having an intermediate or unfavorable cytogenetic risk or having a favorable cytogenetic risk but being refractory to chemo-therapy. Patients younger than 18 years, those with relapsed AML within 30 days of transplant, and/or those with unavailable relapse date or conditioning regimens were excluded.
Data were obtained from standard forms that were created for transplant patients in the Nucleus Electronic Data Management System (version 9.3.39; Monad Software Co., Ankara, Turkey). Protocols were in accordance with the Joint Accreditation Committee: International Society for Cellular Therapy and the European Group for Blood and Marrow Transplantation Accreditation criteria. A data audit group checked all data. The study received approval from our institution’s ethics committee.
From our database, we selected consecutive patients who (1) had a diagnosis of AML according to World Health Organization criteria,7 (2) had under-gone allo-SCT, and (3) were treated between 2007 and 2018. A total of 76 patients who provided written informed consent according to institutional policy fulfilled these criteria and were included in the study.
The Southwest Oncology Group cytogenetic classification was used for cytogenetic risk analysis,8 and patients were grouped as favorable, intermediate, and unfavorable in terms of cytogenetic status. Induction therapy was 3 days of idarubicin 12 mg/m2/day plus cytarabine 100 mg/m2/day followed by 4 more days of only cytarabine 100 mg/m2/day. Chemotherapy as consolidation after remission included 0 to 3 cycles of high-dose cytarabine and 2 × 1.5 or 2 or 3 g/m2/day on days 1, 3, and 5. Fludarabine-cytarabine-idarubicin therapy in refractory patients was as follows: fludarabine 30 mg/m2/day, cytarabine 2000 mg/m2/day for 5 days, and idarubicin 10 mg/m2/day for 3 days. Complete remission during transplant was defined as < 5% bone marrow blasts.
The method of MRD analysis in our clinic has been previously published.9 Briefly, all flow cytometry measurements were performed using the blue (wavelength = 488 mm) and red (wavelength = 633 mm) lasers of an 8-parameter Becton Dickinson FACS CANTO II device (BD Bioscience, San Jose, CA, USA). Leukemia-related immune typing was performed at the time of diagnosis. Three tubes containing diagnostic markers of AML (CD45APC H7, CD13 APC, CD33 PerCp, CD7 PE, CD19 AlexaFluor, CD56 PE, CD123 PerCp, CD34 PE-Cy7, CD11b Alexa Fluor, HLA DR APC, CD38 FITC, CD15 FITC, CD64 PE, and CD14 Alexa Fluor) were used. Each tube received approximately 106 cells, which were evaluated using FACS DIVA Software (BD Bioscience). Minimal residual disease was defined as the presence of cell groups exhibiting abnormal antigen expression during maturation of bone marrow cell lines. If possible, the same monoclonal antibodies and fluorochromes were used for MRD follow-up after diagnosis. The MRD bone marrow-positive cutoff was 0.1%. All antibodies were supplied by BD Bioscience.9
All patients were treated with a myeloablative or a reduced-intensity conditioning (RIC) regimen. A myeloablative regimen was used in patients who received autologous or allogeneic transplant. Donor sources included matched related, 1 mismatched related, haploidentical, matched unrelated, or 1 mismatched unrelated. For patients who had allogeneic transplant procedures, the myeloablative regimen included total busulfan of 12.8 mg/kg intravenously (on days -5, -4, -3, -2), total fludarabine of 150 mg/m2 (on days -6, -5, -4, -3, -2), and total antithymocyte globulin of 30 mg/kg (on days -4,-3, -2) (total body irradiation of 400 cGy), with graft-versus-host disease (GVHD) prophylaxis of cyc-losporine (starting on day -2) and methotrexate 8 mg/m2 (on days 2, 4, and 8). For patients who had haploidentical transplant, the regimen included intravenous busulfan 9.6 mg/kg on days -5, -4, and 3; fludarabine 150 mg/m2 on days -7, -6, -5, -4, -3; and cyclophosphamide 14.5mg/kg/day on days -7 and -6, with GVHD prophylaxis of cyclophosphamide 50 mg/kg on days 3 and 4, tacrolimus starting on day 5, and mycophenolate mofetil. For RIC regimen, patients received total intravenous busulfan of 9.6 mg/kg (on days -4, -3, -2), total fludarabine of 150 mg/m2 (days -6, -5, -4, -3, -2), antithymocyte globulin of 30 mg/kg (days -4, -3, -2), and GVHD prophylaxis of cyclosporine and methotrexate.
Intensive therapy was defined as induction-type cytoreductive chemotherapy plus donor lymphocyte infusion (DLI) and/or a second allograft. HLA typing for recipients with unrelated donors was classified using published Center for International Bone Marrow Transplant Registry (CIBMTR) criteria.10 The intensity of conditioning regimens was classified according to CIBMTR definitions.11,12
All patients were assessed for response according to International Working Group criteria.13 Complete response was defined as bone marrow blasts < 5%, absence of blasts with Auer rods, absence of extramedullary disease, absolute neutrophil count > 1.0 × 109/L, and platelet count > 100 × 109/L. Overall survival was calculated from the day of diagnosis until death from any cause or the last follow-up. Patients who achieved CR with a remission induction therapy were categorized as having CR1, and patients who achieved CR only after salvage therapy were categorized as having CR2.
The diagnosis and grading of acute GVHD were performed according to the consensus conference on acute GVHD.14 Chronic GVHD was diagnosed and graded based on published criteria.15 Relapse after allo-SCT was confirmed based on bone marrow aspirate according to standard criteria.16 Treatment of relapse consisted of (1) an intensive salvage regimen with high or intermediate dose level of cytarabine, associated or not with gemtuzumab and/or an anthracycline17; (2) a nonintensive salvage (low-dose cytarabine, demethylating agents, com-bination of oral chemotherapy or investigational drugs); and (3) supportive care with or without oral cytoreductive therapy such as hydroxyurea. The use of DLI was determined according to the availability of donor cells, previous history of GVHD, and clinical status. Patients in CR following salvage chemotherapy received cytarabine-based consolidation regimens and/or second transplant and/or DLI. Criteria in accordance with Cheson and colleagues were used to assess the response after salvage therapy.13
Statistical analysis was performed with SPSS version 20 (SPSS, Chicago, IL, USA). Nominal data are shown as percentages, normally distributed contin-uous data are shown as means ± standard deviation, and nonnormally distributed continuous data are shown as median (minimum to maximum or range). Mean survival time is shown as mean ± standard error (95% confidence interval [95% CI]). Chi-square tests, t tests, and Mann-Whitney U tests were used where appropriate. Kaplan-Meier analysis was performed for survival analysis. Comparisons between groups with respect to survival were performed using the log-rank test. Cox regression analysis was used to determine the independent predictors of relapse and OS. A P value of ≤ .05 was considered significant.
Patient and disease characteristics are shown in Table 1. Of total patients, 33 (43.3%) had CR MRD-positive status before transplant, 27 (35.5%) had CR MRD-negative status, and 3 (3.9%) showed pro-gressive disease. Of those in CR, 13 patients (17.1%) had unknown MRD status. During the study period, relapse occurred in 27 patients (35.5%) with hematologic relapse being the most common (31.6%). The median time to development of relapse was 5.5 months (range, 1.5-37 mo). Of 27 patients with relapse, 7 (25.9%) were alive during the study period; of 49 patients without relapse, 46 (93.9%) were alive (P < .001). The mean estimated time for development of relapse was 48.7 ± 3.9 months (95% CI, 40.9-56.5 mo). The cumulative 1- and 5-year OS rates for patients with relapse were 67.8% and 58.7%, respectively. The mean length of survival in patients with relapse was 22.4 ± 4.1 months (3-year OS of 23.5%, 5-year OS of 0%), and the mean length of survival in those without relapse was 70.5 ± 2.5 months (3-year OS of 95.9%, 5-year OS of 91.1%) (log-rank chi-square test 42.3; P < .001; Figure 1).
Treatment protocols for relapse are shown in Table 1. Figure 2 compares the mean OS after relapse in patients in terms of treatment protocols used after relapse. No significant differences were found between treatment groups with respect to mean postrelapse OS (log-rank chi-square test 2.512; P = .643). The mean estimated OS times after relapse in patients who received intensive salvage, nonintensive salvage, intensive salvage plus DLI, nonintensive salvage plus DLI, and supportive therapy were 9.3 ± 1.7, 8.0 ± 3.3, 13.4 ± 4.1, 4.9 ± 1.9, and 7.7 ± 4.8 months, respectively. Although not significant, the mean OS after relapse in patients in the intensive salvage plus DLI group was better than that shown in the other groups.
Myeloablative and RIC regimens were the conditioning regimens used before transplant (Table 1). Relapse occurred in 36.8% (21/57) of patients who received myeloablative therapy, and this rate was 31.6% (6/19) in patients who received RIC therapy (P = .451). Of 57 patients in the myeloablative group, 28 patients (49.1%) developed GVHD; in the RIC group, 57.9% (11/19) developed GVHD (P = .346).
Transplant was performed in 64 patients classified as CR1 and 9 patients classified as CR2. The median time from transplant to relapse was 18.4 months (range, 1.5-75.0 mo) in CR1 and 14.8 months (range, 1.6-71.7 mo) in CR2 (P < .001).
Median follow-up was 33.9 months (range, 5.4-75 mo) for living patients. A total of 23 patients (30.3%) died during the study period, with median survival of 9.6 months (range, 1.6-44.5 mo). In the study population, mean survival time was 52.8 ± 3.8 months (95% CI, 45.4-60.2 mo) and the cumulative 1- and 5-year OS rates were 80.7% and 63.2%, respectively. Of study patients, 36.4%, 28.8%, and 30.8% with favorable, intermediate, and unfavorable cytogenetic status died, respectively, during the study period. The survival distributions for patients with favorable, intermediate, and unfavorable cytogenetic status are compared in Figure 3. Patients with favorable, intermediate, and unfavorable cytogenetic status showed OS of 46.6 ± 10.4 months (95% CI, 26.2-66.9 mo), 54.6 ± 4.4 months (95% CI, 45.9-63.2 mo), and 36.9 ± 5.9 months (95% CI, 25.4-48.5 mo), respec-tively, with no significant differences shown between groups (log-rank chi-square test 0.430; P = .807).
The MRD statuses of 60 patients were known. At the end of the study period, 75.8% of patients with CR MRD-positive status and 70.4% of patients with CR MRD-negative status remained alive. Figure 4 shows a comparison of survival of patients in CR with respect to MRD status. The overall time until relapse was 47.8 ± 6.2 months (95% CI, 35.8-59.9 mo) in the CR MRD-positive patients and 54.4 ± 6.1 months (95% CI, 42.6-66.3 mo) in CR MRD-negative patients, with no significant difference shown between these groups (log-rank chi-square test 0.614; P = .433; Figure 4A). The OS time was 48.4 ± 5.8 months (95% CI, 37.1-59.7) in CR MRD-positive patients and 54.7 ± 5.9 months (95% CI, 43.1-66.2 mo) in CR MRD-negative patients (log-rank chi-square test 0.763; P = .382). The 1- and 5-year OS rates were 84.2% and 50.6%, respectively, in CR MRD-positive patients and 88.1% and 68.7%, respectively, in CR MRD-negative patients (not significant; Figure 4B).
Patients who developed GVHD had similar OS versus patients who did not develop the disease (47.0 ± 5.7 vs 57.2 ± 4.6 mo; log-rank chi-square test 2.48; P = .115). Patients with chronic extensive GVHD had similar OS versus patients who had no GVHD or acute GVHD (49.3 ± 4.6 vs 58.2 ± 5.7 mo; log-rank chi-square 1.218; P = .27).
A total of 66.7% of patients with progressive disease before transplant died during the study period, and 27.1% of patients with CR died. The mean OS was 11.3 ± 2.9 months for patients with progressive disease and 55.2 ± 3.8 months for patients with CR (log-rank chi-square test 7.802; P = .005).
The pretransplant therapy protocols are shown in Table 2. A total of 41 patients (53.9%) received consolidation treatment, and 13 (31.7%) died; 10 of the remaining 35 patients (28.6%) who did not receive consolidation therapy died (P = .483). Based on Kaplan-Meier analysis, the mean survival time was 52.6 ± 5.0 months (95% CI, 42.7-62.4 mo) for patients who received consolidation therapy and 44.2 ± 4.4 months (95% CI, 35.6-52.8 mo) for patients who did not receive consolidation therapy (no significant difference; log-rank chi-square test 0.007; P = .932). To analyze the additive effect of consolidation on survival, patients who received only induction and consolidation therapy and patients who received reinduction were compared. Eight of 21 patients (38.1%) who received induction and consolidation therapy died, whereas none of the 7 patients who received reinduction treatment died (P = .018) during the study period. When patients who received induction + consolidation + fludarabine-cytarabine (FLAG) treatment (n = 21) were compared with patients who only received induction + FLAG (n = 6) with respect to survival, in both groups, 33.3% of patients died (P = .695), and mean OS in the induction + consolidation + FLAG group was 50.2 ± 13.4 months (95% CI, 23.9-76.3 mo). The mean OS was 43.1 ± 5.5 months (95% CI, 32.4-53.9 mo) in the induction + FLAG group (log-rank chi-square 0.035; P = .852).
Independent predictors of OS were assessed using Cox regression analysis (Table 3). Relapse, pretransplant cytogenetic and MRD status, GVHD development, and pretransplant consolidation therapy were included in the model, and it was shown that only relapse was an independent predictor of OS since it increased mortality risk 26.22 times (Table 3). Because relapse was the only independent predictor of OS, factors associated with relapse were assessed (Table 4). In the model, pretransplant cytogenetic and MRD status, GVHD development, and pretransplant consolidation were included. As shown in Table 4, GVHD was found to be associated with relapse, and the development of GVHD was found to decrease relapse 0.29 times.
In this retrospective study, we showed that MRD or cytogenetic status of AML patients before transplant did not have any effect on the relapse rate or OS. We also demonstrated that consolidation therapy did not improve outcomes. Overall survival in patients who developed GVHD during the posttransplant period was similar to that shown in patients who did not develop GVHD. There were no significant dif-ferences shown between having favorable, unfa-vorable, or intermediate cytogenetic risk status. A total of 66.7% of patients with progressive disease died during the study period. This rate was 27.1% in patients with CR status.
It has been reported that assessment of MRD after induction therapy and at different points thereafter has a prognostic importance5 and that achieving a negative MRD status before allo-HSCT reduces relapse risk.18 Thus, achieving MRD-negative status with additional therapy is believed to be associated with improved survival. In our institution, MRD status was assessed after induction therapy, and if MRD-negative status could not be achieved reinduction therapy was administered. We aimed to perform transplant in patients who were MRD negative; however, if MRD negativity could not be achieved with reinduction therapy, patients undergo transplant. In the present study, although the relapse rate was decreased and OS was improved in CR MRD-negative patients, the differences were not significant. In addition, MRD status was not an independent predictor of relapse or OS. Hourigan and associates reported that MRD is used in clinical practice to guide the care of individual patients, but more data are required to establish the use of MRD as a surrogate endpoint for clinical trials in AML.19 Recently, in the consensus document from the European LeukemiaNet MRD Working Party, it was reported that, although adding MRD status to remission assessment can optimize long-term outcomes, these approaches have not yet been qualitatively or quantitatively standardized, making it difficult to use clinically.20 These results show that the predictive role of MRD status on clinico-pathologic progress of AML remains unclear.
The cytogenetic status is considered one of the most important prognostic markers predicting remission, relapse, and OS.21 In the current study, clinical outcomes associated with the Southwest Oncology Group coding pretreatment risk criteria/cytogenetic scores were assessed and it was found that, although not significant, the OS in patients with unfavorable cytogenetic status was worse than that shown in the favorable and intermediate status groups. However, in the regression model, cytogenetic status was not found to be an independent predictor of OS or relapse. Our results are similar to other studies in which cytogenetic status was not found to be a prognostic factor.21,22 Tallman and associates reported similar disease-free survival and OS in 3 cytogenetic risk groups and reported that transplant-related mortality was higher in their favorable and intermediate risk groups. They suggested that transplant-related mortality might have negated any advantage that favorable or intermediate cytogenetic risk groups might have otherwise had because of lower relapse rates.22 In the present study, except for GVHD, other factors that may be related with transplant-related mortality, such as cytomegalovirus viremia, were not included in our statistical analyses. In addition, although transplant is not performed in patients with favorable cytogenetic risk, patients refractory to chemotherapy with favorable cyto-genetics are considered as having adverse risk. In these patients, there may be undetectable mutations.
Outcomes of HSCT in CR2 are generally poorer than outcomes of HSCT in CR1.23-25 In the present study, patients in the CR1 group showed an improved OS compared with patients in the CR2 group. In general, the probability of achieving CR2 after chemotherapy is low and earlier allo-SCT would be the optimal approach. Therefore, we believe that transplant should be performed on time and should not be reserved for CR2 patients.
The rationale supporting consolidation therapy before transplant is that it may reduce relapse rates and thus improve OS by decreasing the leukemic cell number. In addition, previous studies have supported a benefit of this therapy.26,27 However, in a review by Yeshurun and associates, it was stated that the value of postremission consolidation chemotherapy before allo-HSCT remains controversial.5 They reported that considerable data have been gathered from retro-spective studies from large transplant centers establishing a lack of any survival benefit of additional pretransplant consolidation cycles once the patient achieves CR.5 Kassim and associates stated that HSCT is an effective postremission consolidation treatment.28 Our results showed that consolidation therapy before allo-SCT showed no association with relapse and OS. In addition, the number of consolidation therapies had no effect (the 5-year relapse and OS rates in the no consolidation group, one consolidation therapy group, and ≥ 2 consolidation therapies group were 48.2%, 75%, and 67% and 66.1%, 70%, and 56%, respectively). Therefore, it would be logical to perform transplant therapy as soon as the patient achieves CR with induction therapy. With this approach, the patient would not experience toxic side effects of the consolidation therapy, which does not appear to improve outcomes. If a suitable donor is readily available for SCT and the patient has achieved the first CR, there is no need for consolidation. As stated by Yeshurun and associates, toxicity after an unnecessary consolidation chemotherapy would make transplant impossible or increase transplant-related mortality and morbidity.5 When a donor is not available, the patient should receive bridging consolidation prior to allo-SCT to prevent early relapse before transplant.
Our results confirm that patients with AML who relapsed after allo-SCT have poor outcomes. The mean survival time from relapse was estimated to be 22.4 ± 4.1 months and 3- and 5-year OS rates were 23.5% and 0%, respectively. In regression analysis, having a GVHD was found to have a 3-fold protective effect for relapse development. No other risk factors were found to be related to relapse development in the present study.
There is no consensus on a standardized post-relapse treatment protocol.4 Different institutions use different protocols; however, intensive chemotherapy with or without DLI, nonintensive chemotherapy with DLI, a second allo-SCT, and supportive care are among the commonly preferred treatment options.29 In our institution, we also prefer these treatment options based on the development of GVHD, cytogenetic status, age, and the clinical conditions of the patient. In our study, relapse was seen in 27 patients; when postrelapse treatments were compared with respect to OS, it was found that none of the treatment protocols had significant superiority over others. However, although not statistically signi-ficant, patients who received intensive salvage therapy plus DLI had a longer mean survival time than the other treatment groups. Previous studies reported that the best outcomes are achieved in patients who receive intensive chemotherapy followed by a second allo-SCT or DLI.30-33
Relapse is the most important predictor of OS, after which the prognosis is poor. The most important strategy is the prevention of relapse. We found that pretransplant MRD and cytogenetic status did not affect relapse rates and OS. We also observed no improved outcomes when we compared conso-lidation therapies. Progressive disease status prior to allo-SCT was associated with poor prognosis, and patients who were transplanted in CR1 had better OS than patients who received transplants in CR2.
DOI : 10.6002/ect.2018.0352
From the Adana Baskent University Department of Hematology, Adult Bone Marrow
Transplantation Unit, Adana, Turkey
Acknowledgements: The authors have no sources of funding for this study and have no conflicts of interest to declare.
Corresponding author: Pelin Aytan, Adana Baskent University, Dr. Turgut Noyan Uygulama ve Araştırma Hastanesi, Dadaloğlu Mah, 2591 Sok. 4/A, 01250, Yüreğir, Adana, Turkey
Phone: +90 532 7803572
Figure 1. Overall Survival of Patients With and Without Relapse
Figure 2. Comparison of Postrelapse Treatment Protocols in Terms of Overall Survival
Figure 3. Comparison of Survival Probability With Respect to Cytogenetic Status
Figure 4. Comparison of Survival of Patients in Complete Remission With Respect to Minimal Residual Disease Status
Table 1. Patient and Disease Characteristics
Table 2. Pretransplant Therapy Protocol
Table 3. Cox Regression Model for Identification of Factors Related to Overall Survival: Analysis of Maximum Likelihood Estimates
Table 4. Cox Regression Model for Identification of Factors Related to Relapse: Analysis of Maximum Likelihood Estimates