Volume 14, Issue 1 , Pages 28-35, January 2008
A Prognostic Score for Patients with Acute Leukemia or Myelodysplastic Syndromes Undergoing Allogeneic Stem Cell Transplantation
Article Outline
Abstract
Allogeneic hematopoietic stem cell transplantation (SCT) has the potential to cure patients with acute leukemia or myelodysplastic syndromes (MDS), but a number of prognostic factors can influence the outcome of transplantation. At present, no transplantation-specific risk score exists for this patient population. We propose a simple scoring system for patients with acute myelogenous leukemia (AML), acute lymphoblastic leukemia (ALL), or MDS, based on a retrospective analysis of 445 patients undergoing SCT at our institution (divided into training and validation subsets). The score depends on 5 variables: age, disease, stage at transplantation, cytogenetics, and pretransplantation ferritin. It divides patients into 3 groups of comparable size, with 5-year overall survival (OS) of 56% (low risk), 22% (intermediate risk), and 5% (high risk). This prognostic score could be useful in making treatment decisions for individual patients, in stratifying patients entering clinical trials, and in adjusting transplantation outcomes across centers under the new federal reporting rules.
Key Words: Allogeneic stem cell transplantation, Acute myeloid leukemia, Acute lymphoblastic leukemia, Myelodysplastic syndrome, Cytogenetics, Iron overload
Introduction
Allogeneic hematopoietic stem cell transplantation (SCT) can be curative for patients with acute myelogenous leukemia (AML) [1], acute lymphoblastic leukemia (ALL) [2], or myelodysplastic syndromes (MDS) [3]. However, there is considerable heterogeneity in transplantation outcomes within this group of patients, with many patient-related, disease-related, and transplant-related factors influencing prognosis. A prognostic score that is specifically applicable to this patient population would be useful in several ways: first, it would help with prognostication at the time of transplantation consideration, which could have an impact on treatment decisions; second, it would allow stratification of patients entering clinical trials of transplantation. Successful prognostic scores developed for other diseases, such as the International Prognostic Index for patients with aggressive non-Hodgkin lymphoma (NHL) [4] or the International Prognostic Scoring System (IPSS) for patients with MDS [5], have been tremendously useful to clinicians and researchers alike for prognostication and clinical trial stratification. Unfortunately, there are at present few prognostic scores designed specifically for patients undergoing SCT. Three general risk scores have been proposed that are applicable to all patients undergoing transplantation [6, 7, 8]; however, by design, those scores rely minimally if at all on disease-specific factors (such as cytogenetics). The European Group for Blood and Marrow Transplantation (EBMT) developed a risk score for patients with chronic myelogenous leukemia (CML) [9]. No such score currently exists specifically for patients with acute leukemia or MDS undergoing SCT, despite the fact that they account for a substantial fraction of transplantation patients (52% of all allogeneic transplants reported from the International Bone Marrow Transplant Registry to the Center for International Blood and Marrow Transplant Research (CIBMTR) [10]). Although the IPSS has prognostic values for patients with MDS undergoing transplantation [11, 12], this score was derived on a cohort of patients who were not transplanted, and thus may not be an optimal prognostication tool in the context of SCT.
There is yet another benefit to transplant-specific prognostic scores. In December 2005, the U.S. Stem Cell Therapeutic and Research Act of 2005 created the C.W. Bill Young Cell Transplantation Program. As part of this program, U.S. centers will be required to report outcomes data for all allogeneic transplantation, and this data will be publicly available. It will therefore be critical to have well-defined risk stratification tools to allow calibration of outcomes across centers that may transplant patients with different risk profiles.
In this study, we propose and validate a simple prognostic score for patients with AML, ALL, or MDS undergoing SCT, based on the outcomes of 445 patients transplanted at our institution.
Methods
Patients
We studied 660 consecutive adult patients with AML, ALL, or MDS who underwent allogeneic stem cell transplantation at the Dana-Farber/Brigham and Women’s Hospital transplant program between January 1997 and December 2005. Patients with therapy-related AML or AML arising out of MDS were included. Patients with CMML were excluded. The following pretransplantation variables were collected: age, sex, disease type and stage, prior leukemogenic therapy, cytogenetics, donor HLA match, stem cell source, conditioning regimen, graft-versus-host disease (GVHD) prophylaxis regimen, cytomegalovirus (CMV) serostatus of donor and recipient, sex of donor, and pretransplantation serum ferritin [13]. Ferritin was included because we have recently demonstrated that it has an important prognostic impact for patients with MDS or acute leukemia receiving myeloablative conditioning [13]. For patients receiving reduced intensity conditioning (RIC), ferritin was not included as a prognostic factor in the multivariable models. Patients with missing data were excluded; this left 445 patients, who form the basis of this study. Institutional review board (IRB) approval was obtained from the Office for the Protection of Research Subjects (OPRS) at Dana-Farber/Harvard Cancer Center, in accordance with the principles of the Declaration of Helsinki.
Cytogenetics
Cytogenetics were determined by standard metaphase karyotype analysis or fluorescence in situ hybridization (FISH; without requiring central review). For ALL, t(9;22) and t(4;11) are considered adverse, and all others intermediate. We used the MRC classification scheme for AML [14], and the cytogenetic classification scheme of the IPSS for MDS [5]. For risk group assignment purposes, we used the cytogenetics at diagnosis for patients in remission, and the pretransplantation cytogenetics for patients with active disease.
Transplantation
Patients were transplanted under several treatment and investigational protocols over the 9-year period covered by this study. Myeloablative conditioning regimens consisted mostly of cyclophosphamide (3600 mg/m2 or 120 mg/kg) plus either total body irradiation (TBI) (1400 cGy in 7 fractions) or busulfan (16 mg/kg by mouth or 12.8 mg/kg i.v.). Reduced intensity regimens consisted mostly of fludarabine (120 mg/m2) with busulfan (3.2 mg/kg). Patients received either bone marrow or peripheral blood stem cells, from matched or mismatched, related or unrelated donors. Thirteen patients received umbilical cord blood (UCB) transplants. GVHD prophylaxis regimens consisted mostly of a combination of calcineurin inhibitor and methotrexate (MTX), tacrolimus plus sirolimus, with or without low-dose methotrexate, or T cell depletion (TCD).
Statistics
Patient baseline characteristics were reported descriptively. Overall survival (OS) and disease-free survival (DFS) were calculated using the Kaplan-Meier method. OS was defined as the time from stem cell infusion to death from any cause. Patients who were alive or lost to follow-up were censored at the time last seen alive. DFS was defined as the time from stem cell infusion to relapse or death from any cause, whichever occurred first. Patients who were alive without relapse were censored at the time last seen alive and relapse-free. The log-rank test was used for comparisons of Kaplan-Meier curves. Cumulative incidence curves for nonrelapse death and relapse with or without death were constructed reflecting time to relapse and time to nonrelapse death as competing risks. The difference between cumulative incidence curves in the presence of a competing risk was tested using the Gray method [15]. Time to relapse and time to nonrelapse death were measured from the date of stem cell infusion.
The entire cohort was split randomly 2:1 into a training set (n = 297) and a validation set (n = 148). Potential prognostic factors for survival were examined in the proportional hazards model, as were relevant interaction terms. The proportional hazards assumption was tested for all important variables both graphically and analytically [16]. Regression assumptions were tested using Martingale and deviance residual analysis. Transformation for continuous variables to categoric variables was done by visual analysis of residual plots. A parsimonious model was built using a stepwise variable selection method, with an entry criterion of P ≤ .1 and a retention criterion of P < .05. To create the risk score, each variable was assigned an integral number of points based roughly on its associated hazard ratio (HR) from the Cox model, as described in the text. We calculated the c-statistic for the validation cohort [17], defined as the proportion of informative patient pairs whose outcomes are correctly ordered by the risk score. All calculations were done using SAS 9.1 (SAS Institute Inc, Cary, NC), R (version 2.3.1), Splus (version 3.4), and Matlab 6.5 (Mathworks, Natick, MA).
Results
Patient Characteristics
We retrospectively reviewed the posttransplantation outcomes of 445 adult patients with AML, ALL, or MDS transplanted at our institution between 1997 and 2005. Median follow-up for survivors was 25 months (range: 6-102). Table 1 lists the baseline characteristics of those patients. The median age in this cohort was 45 years (range: 18-71). Forty-three percent of patients had AML, 18% had ALL, and 39% had MDS (including 18% with secondary AML). Twelve percent of patients had therapy-related disease. Forty-eight percent of patients were transplanted in first complete remission (CR1) or with untreated MDS, whereas 33% had active disease at the time of transplantation. Sixty-nine percent had myeloablative conditioning (mostly cyclophosphamide + TBI), whereas 31% had nonmyeloablative conditioning (mostly low-dose busulfan + fludarabine); 41% of patients received a graft from an HLA-matched sibling; 61% received peripheral blood stem cells (PBSC), whereas 36% received bone marrow and 3% received cord blood.
Table 1. Baseline Characteristics of the Patients
| Variable | Total, No. (%)⁎ |
|---|---|
| Number of patients analyzed | 445 (100%) |
| Median age in years (range) | 45(18-71) |
| Disease | |
| ALL | 80(18) |
| AML | 191(43) |
| MDS | 174(39) |
| RA/RARS/RCMD | 79(18) |
| RAEB | 17(4) |
| AML from MDS | 78(18) |
| Therapy-related disease | 54(12) |
| Stage | |
| Untreated MDS | 86(19) |
| Untreated AML/ALL | 10(2) |
| CR1 | 131(29) |
| CR ≥ 2 | 74(17) |
| Induction failure | 75(17) |
| Active relapse | 69(16) |
| Cytogenetics† | |
| Favorable | 102(23) |
| Intermediate | 210(47) |
| AML/MDS adverse | 101(23) |
| ALL adverse | 32(7) |
| Graft source | |
| Peripheral blood | 271(61) |
| Bone marrow | 161(36) |
| Cord blood | 13(3) |
| Donor HLA match | |
| MRD | 184(41) |
| MUD | 193(43) |
| Mismatched related | 9(2) |
| Mismatched unrelated | 59(13) |
| Conditioning | |
| Myeloablative | 309(69) |
| Cy/TBI | 297(67) |
| Bu/Cy | 9(2) |
| Other | 3(0) |
| Reduced intensity | 136(31) |
| Bu/flu | 127(29) |
| Other | 9(2) |
| GVHD prophylaxis | |
| CnI ± steroids ± MMF | 50(11) |
| CnI + MTX ± steroids | 172(39) |
| CnI + Siro ± MTX | 165(37) |
| TCD | 58(13) |
| Sex‡ | |
| Female -> Male | 98(22) |
| Male -> Female | 112(25) |
| Male -> Male | 137(31) |
| Female -> Female | 95(21) |
| CMV seropositivity | |
| Recipient | 182(41) |
| Donor | 164(37) |
⁎Percentages may not add to 100 because of rounding. |
†See Methods section for details. |
‡Donor sex information was missing for 3 patients. |
Cytogenetics was available for all 445 patients. AML and MDS cytogenetics were classified according to the MRC [14] and IPSS [5] grouping schemes, respectively. For ALL, we considered t(9;22) and t(4;11) to be adverse, and all others to be intermediate. When classified in this way, 47% of the patients had intermediate cytogenetics, 23% had favorable cytogenetics, and 30% had adverse cytogenetics.
Outcomes and Prognostic Factors
For the entire cohort, the 5-year OS was 32%, and the 5-year DFS was 31%; the 5-year cumulative incidence of relapse (CIR) was 42%, and the 5-year cumulative incidence of nonrelapse mortality (NRM) was 27%. We divided the patients 2:1 by random assignment into a training set and a validation set. All subsequent analyses to establish the risk score were performed on the group of patients in the training set (n = 297). Within this group, we performed proportional hazards univariate and multivariate analyses to determine the prognostic factors for OS. All the variables from Table 1 were included in the model. We also included pretransplantation serum ferritin for patients receiving myeloablative conditioning (as this is the patient subgroup for which ferritin has a significant prognostic impact [13]) in the model, as well as year of transplantation. The results are shown in Table 2. The variables that had a significant impact on OS in multivariate analysis were: age, disease, stage at transplantation, cytogenetics, pretransplantation ferritin (for patients undergoing myeloablative transplantation), and year of transplantation. We also built a parsimonious model using stepwise variable selection; in this analysis, the same 5 variables were selected (data not shown). Donor HLA match, conditioning regimen intensity, graft source, GVHD prophylaxis regimen, sex mismatch, CMV serostatus, and therapy-related disease were not significant in this model. We also examined the effect of pretransplantation Eastern Cooperative Oncology Group (ECOG) performance status on outcome. This data was available for 32% of the patients in this study, with a median value of 0 (range: 0-3). Performance status had no significant effect on posttransplantation outcome (HR for mortality 1.3, P = .2).
Table 2. Univariate and Multivariate Analyses for Overall Survival
| Univariate Analysis | Multivariate Analysis | |||
|---|---|---|---|---|
| Variable | HR | P | HR | P |
| Age ≥40 | 1.4 | .036 | 2.0 | .0007 |
| Disease | ||||
| AML | 1.0 | ref | 1.0 | ref |
| ALL | 1.0 | 1.0 | 1.0 | .9 |
| Low-risk MDS⁎ | 0.8 | .4 | 1.0 | 1.0 |
| High-risk MDS⁎ | 2.4 | .040 | 3.1 | .023 |
| AML arising from MDS | 1.6 | .034 | 1.6 | .071 |
| Therapy-related disease | 1.4 | .2 | 1.2 | .7 |
| Cytogenetics | ||||
| Favorable | 0.8 | .2 | 0.6 | .037 |
| Intermediate | 1.0 | ref | 1.0 | ref |
| Adverse ALL | 1.1 | .8 | 1.3 | .4 |
| Adverse AML/MDS | 1.8 | .002 | 1.6 | .051 |
| Stage | ||||
| CR1/Unt MDS | 1.0 | ref | 1.0 | ref |
| CR >1 | 1.0 | .9 | 1.5 | .2 |
| Untreated AML/ALL | 1.9 | .2 | 2.1 | .2 |
| Induction failure | 2.2 | <.0001 | 2.2 | .0007 |
| Relapse | 2.0 | .003 | 2.6 | .0001 |
| Ferritin >2500† | 1.4 | .067 | 1.5 | .043 |
| Graft source | ||||
| Bone marrow | 1.0 | ref | 1.0 | ref |
| Peripheral blood‡ | 1.0 | .8 | 1.2 | .4 |
| Donor match | ||||
| MRD | 1.0 | ref | 1.0 | ref |
| Non-MRD | 1.0 | .9 | 1.1 | .8 |
| Conditioning regimen | ||||
| Ablative | 1.0 | ref | 1.0 | ref |
| Reduced intensity | 1.1 | .5 | 1.0 | .9 |
| GVHD prophylaxis | ||||
| CnI ± MTX ± MMF | 1.0 | ref | 1.0 | ref |
| CnI + Siro ± MTX | 0.7 | .055 | 0.8 | .4 |
| TCD | 0.7 | .2 | 0.9 | .6 |
| Gender mismatch | 1.2 | .3 | 1.1 | .7 |
| Donor CMV seropositive | 1.2 | .2 | 1.1 | .5 |
| Recipient seropositive | 1.0 | .9 | 1.0 | .9 |
| Year of transplantation | ||||
| 1997-2000 | 1.0 | ref | 1.0 | ref |
| 2001-2005 | 0.7 | .081 | 0.6 | .034 |
⁎Low-risk MDS was defined as RA, RARS, and RCMD; high-risk as RAEB-1 and RAEB-2. |
†For patients receiving myeloablative conditioning (see text). |
‡Including umbilical cord blood recipients. |
Results were similar if we grouped cytogenetics according to our recently proposed grouping scheme [18]. The model fit was slightly stronger (Akaike’s information criterion (AIC) 1685 versus 1735, with a lower score indicating a better model fit), but the conclusions were unaltered. We also reclassified ALL cytogenetics according to the results of the MRC UKALLXII/ECOG 2993 study [19]; this had no substantial effect on the model fit (AIC 1735).
Derivation of the Prognostic Score
For each of the significant variables in the multivariate analyses above, we created categoric variables using proportional hazards model or visual analysis of Martingale/deviance residuals against each continuous predictor. The results (not shown) allowed us to categorize age into 2 groups (<40 versus ≥40); disease type into 2 groups (low-risk MDS, AML, or ALL, versus high-risk MDS or MDS transformed to AML); stage into 3 groups (AML/ALL in CR1 or untreated MDS, versus CR > 1 or induction failure, versus active relapse or untreated leukemia); cytogenetics into 3 groups (favorable versus intermediate versus adverse); and pretransplantation ferritin into 2 groups (myeloablative conditioning and ferritin ≥2500 ng/dL, versus RIC or ferritin <2500 ng/dL). The resulting multivariat model using those categoric variables is shown in Table 3. The results were essentially identical when using an ALL cytogenetics classification based on the MRC/ECOG ALL study [19] (not shown).
Table 3. Final Multivariable Model for Overall Survival
| Variable | HR | P |
|---|---|---|
| Age | ||
| <40 | 1.0 | ref |
| ≥40 | 1.8 | .001 |
| Disease | ||
| Low-risk MDS/AML/ALL | 1.0 | ref |
| High-risk MDS/transformed MDS | 2.0 | .0005 |
| Cytogenetics | ||
| Favorable | 1.0 | ref |
| Intermediate | 1.7 | .014 |
| Adverse | 3.2 | <.0001 |
| Stage | ||
| CR1 or untreated MDS | 1.0 | ref |
| CR >1 or Induction failure | 1.8 | .001 |
| AML/ALL untreated or active relapse | 2.4 | <.0001 |
| Ferritin | ||
| RIC OR ferritin <2500 | 1.0 | ref |
| Ablative AND ferritin ≥2500 | 1.7 | .007 |
We then created a risk score by assigning each variable a number of points. We strove for a simple risk score that could be calculated easily during a clinical encounter. Therefore, we chose integer values for the points, and assigned them based on the HR in the proportional hazards model from Table 3. To variables with HR between 1.5 and 2.0, we assigned a score of 1; and to those with HR above 2.0, we assigned a score of 2. The scoring system thus obtained is described in Table 4. Each variable is assigned a score; the sum of the individual components is the overall prognostic score. This defines 3 risk categories, low (score ≤2), intermediate (score = 3), and high (score ≥4). The OS of the 297 patients in the training set, stratified according to their prognostic score, is plotted in Figure 1A.
Table 4. Calculation of the Prognostic Score
| Variable | Score |
|---|---|
| Age | |
| <40 | 0 |
| ≥40 | 1 |
| Disease | |
| Low-risk MDS,⁎ AML, or ALL | 0 |
| High-risk MDS⁎ or transformed MDS | 1 |
| Cytogenetics† | |
| Favorable | 0 |
| Intermediate | 1 |
| Adverse | 2 |
| Stage | |
| CR1 or untreated MDS | 0 |
| CR >1 or induction failure | 1 |
| AML/ALL untreated or active relapse | 2 |
| Ferritin | |
| RIC OR ferritin <2500 ng/dL | 0 |
| Ablative AND ferritin ≥2500 ng/dL | 1 |
| Final score‡ | Risk group |
| ≤2 | Low |
| 3 | Intermediate |
| ≥4 | High |
⁎Low-risk MDS is defined as RA, RARS, or RCMD; high-risk MDS is defined as RAEB. |
†For ALL, t(9;22) and t(4;11) are considered adverse, and all others intermediate. AML cytogenetics are grouped according to the MRC classification scheme, and MDS cytogenetics according to the IPSS scheme. |
‡This is calculated as the sum of the individual components. |
Figure 1.
OS of patients stratified by risk score. A, Training set; B, validation set.
Validation of the Risk Score
We calculated the risk score for the 148 patients in the validation set, whose outcomes were not used for the derivation of the score. Their OS is shown in Figure 1B. The prognostic score remained a highly statistically significant in this group (log-rank P < .0001). We also calculated the c-statistic in the validation cohort. This statistic indicates how often informative pairs of patients are correctly ordered by the risk score with regard to their survival. A score of 1.0 indicates perfect agreement with the data. For the validation group, using time of death as a continuous variable, the c-statistic was 0.66 (for comparison, in the validation set of Sorror et al.’s comorbidity score study the corresponding c-statistic was 0.626). If we considered the binary outcome of survival to 2 years (as in the Pretransplantation Assessment of Mortality [PAM] score of Parimon and colleagues [8]), the c-statistic was 0.79 (for comparison, it was 0.72-0.76 in the PAM study [8]).
Outcomes by Risk Group
We calculated the prognostic score for all 445 patients (including both those in the training and those in the validation set). As shown in Table 5, the patients are thus divided into 3 groups of comparable size (each comprising between 28% and 41% of patients). Figure 2 shows the outcomes of all patients stratified by their risk score. OS is shown in Figure 2A, and DFS in Figure 2B. The 5-year values for OS, DFS, cumulative CIR, and NRM for each risk group are given in Table 5.
Table 5. Summary of Outcomes in the Entire Cohort By Risk Group
| Risk Group | % of pts | 5-y OS | 5-y DFS | 5-y CIR | 5-y NRM |
|---|---|---|---|---|---|
| Low | 41% | 56%(48-64) | 51%(43-59) | 25%(18-32) | 24%(17-30) |
| Intermediate | 31% | 22%(10-33) | 22%(12-32) | 53%(42-65) | 25%(17-32) |
| High | 28% | 5%(0-12) | 6%(0-12) | 58%(48-67) | 36%(26-47) |
Figure 2.
Outcomes for the entire cohort (training and validation sets combined) stratified by risk score. A, OS; B, DFS.
Discussion
We have created a new prognostic score for patients with acute leukemia or MDS undergoing allogeneic stem cell transplantation (SCT). This score depends on only 5 patient and disease characteristics: age, disease, stage at transplantation, cytogenetics, and pretransplantation serum ferritin (for patients undergoing myeloablative transplantation). By combining these 5 variables, our scoring system stratifies patients into 3 groups of roughly comparable size at very different risk of death after transplantation (with 5-year OS ranging from 56% in the low-risk group to 5% in the high-risk group), through an effect on both relapse and NRM. This scoring system was based on retrospective transplantation results at a single institution. Nonetheless, it retained its high prognostic significance when applied to a validation set of randomly selected patients who were not included in the derivation of the score (see Figure 1B). The c-statistic, which quantifies the agreement of the prognostic score with the outcome of interest (here, OS) was at least as high as that of other prognostic scores [6, 8], which were not designed specifically for patients with acute leukemia or MDS. In its current form, this scoring system does not include pretransplantation comorbidity score, which was not available in our cohort but could be incorporated in future versions. It must also be noted that many of the patients included in this study were part of the cohorts used in our previous work on the prognostic impact of cytogenetics and ferritin levels [13, 18]. However, the intent of the present study is not to validate our prior results, but to integrate them into a more versatile scoring system.
All of the variables included in this score have been previously shown to have a prognostic impact after transplantation. Age is often considered an adverse prognostic factor through an increase in NRM [20, 21] (although at least 2 retrospective multicenter studies failed to document a significant effect on OS [22, 23]). The possibility of a superior outcome for low-risk MDS compared to high-risk MDS has been previously suggested [24, 25]. More advanced disease stage is always associated with worse outcome (as demonstrated, eg, by the data of the CIBMTR). Pretransplantation serum ferritin has also been shown to carry prognostic significance for patients undergoing ablative transplantation, through an effect on NRM [13, 26]. Finally, the prognostic importance of cytogenetics has been confirmed in several studies [27, 28, 29, 30, 31], including our own, where we demonstrated that cytogenetics influenced the risk of relapse, and proposed a transplantation-specific cytogenetics grouping scheme [18]. Although our grouping scheme allowed a better model fit to the data, it was derived on a cohort that included most of the patients in the current study. Because we have not yet validated our grouping scheme in a multi-institution study, we did not use it for the present risk score. In the future, the recognition of prognostic subtypes based on molecular analyses within the normal karyotype category will undoubtedly alter the scheme [32] and subdivide this category, which still includes the majority of patients.
This scoring system could be used in several ways. First, it should allow clinicians to easily estimate the prognosis of patients with MDS or acute leukemia who are candidates for SCT. This could be helpful in discussing the value of transplantation and in choosing a treatment course. It also identifies a subgroup of patients (high-risk) with a dismal prognosis (5-year OS of 5%), for whom the value of transplantation may be questionable, and who may especially benefit from enrollment into clinical trials. Second, this scoring system could be used to stratify patients entering clinical trials of transplantation. At present, there is no consensus as to how to perform such stratification, and some variables that are commonly used, such as donor HLA match, may not be nearly as prognostically important today as less commonly used variables (such as cytogenetics). Similar prognostic scores, such as the IPI for aggressive lymphoma and the IPSS for MDS, are widely used in clinical trial design. Finally, our prognostic score could be used under the new federal reporting requirements for transplant centers to adjust for differences in patient mix and allow a fairer comparison of outcomes across centers.
Prospective validation of this scoring system through a multicenter collaboration would confirm its utility and possibly fine-tune the score assignments. Further in the future, it is possible that improvements in the care of specific patient subgroups (eg, chelation of iron-overloaded patients, or the use of tyrosine kinase inhibitors after transplantation for Ph+ ALL) would require reevaluation of the scoring scheme. This reevaluation could be accomplished through a periodic reexamination of multicenter transplantation data, such as those that will be reported to the Stem Cell Therapeutic Outcome Database, which the CIBMTR will administer. This process could even be semiautomated such that clinicians worldwide would have ready access to up-to-date, international prognostic scoring systems for patients with a variety of diseases undergoing SCT.
Acknowledgments
This work was funded in part by P01 HL070149 from the National Heart, Lung and Blood Institute. P.A. designed and performed the research, analyzed the data, and wrote the paper. H.T.K. analyzed the data and edited the paper. C.S.C. collected data and edited the paper. V.T.H. collected data and edited the paper. J.K. collected data and edited the paper. J.R. collected data and edited the paper. E.P.A. collected data and edited the paper. J.H.A. designed the research, collected data, and edited the paper. R.J.S. designed the research, collected data, and wrote the paper.
References
- . Role of allogeneic stem cell transplantation in current treatment of acute myeloid leukemia. Hematology Am Soc Hematol Educ Program. 2005;151–155
- . Haemopoietic stem cell transplantation for acute lymphoblastic leukaemia. Cancer Treat Rev. 2003;29:3–10
- . Optimization of transplant regimens for patients with myelodysplastic syndrome (MDS). Hematology Am Soc Hematol Educ Program. 2005;167–173
- A predictive model for aggressive non-Hodgkin’s lymphoma (The International Non-Hodgkin’s Lymphoma Prognostic Factors Project). N Engl J Med. 1993;329:987–994
- International scoring system for evaluating prognosis in myelodysplastic syndromes. Blood. 1997;89:2079–2088
- Hematopoietic cell transplantation (HCT)-specific comorbidity index: a new tool for risk assessment before allogeneic HCT. Blood. 2005;106:2912–2919
- Performance status and comorbidity predict transplant-related mortality after allogeneic hematopoietic cell transplantation. Biol Blood Marrow Transplant. 2006;12:954–964
- . A risk score for mortality after allogeneic hematopoietic cell transplantation. Ann Intern Med. 2006;144:407–414
- Risk assessment for patients with chronic myeloid leukaemia before allogeneic blood or marrow transplantation (Chronic Leukemia Working Party of the European Group for Blood and Marrow Transplantation). Lancet. 1998;352:1087–1092
- CIBMTR Progress Report, January-December 2006.
- . Allogeneic bone marrow transplantation for myelodysplastic syndrome: outcomes analysis according to IPSS score. Leukemia. 1998;12(Suppl 1):S25–S29
- . Allogeneic stem cell transplantation in patients with myelodysplastic syndrome: outcome analysis according to the International Prognostic Scoring System. Acta Med Port. 2006;19:343–347
- Prognostic impact of elevated pretransplantation serum ferritin in patients undergoing myeloablative stem cell transplantation. Blood. 2007;109:4586–4588
- The importance of diagnostic cytogenetics on outcome in AML: analysis of 1,612 patients entered into the MRC AML 10 trial (The Medical Research Council Adult and Children’s Leukaemia Working Parties). Blood. 1998;92:2322–2333
- . A class of K-sample tests for comparing the cumulative incidence of a competing risk. Ann Stat. 1988;16:1140–1154
- . Proportional hazards tests and diagnostics based on weighted residuals. Biometrika. 1994;81:515–526
- . Regression modelling strategies for improved prognostic prediction. Stat Med. 1984;3:143–152
- Impact of cytogenetics on outcome of de novo and therapy-related AML and MDS after allogeneic transplantation. Biol Blood Marrow Transplant. 2007;13:655–664
- Karyotype is an independent prognostic factor in adult acute lymphoblastic leukemia (ALL): analysis of cytogenetic data from patients treated on the Medical Research Council (MRC) UKALLXII/Eastern Cooperative Oncology Group (ECOG) 2993 trial. Blood. 2007;109:3189–3197
- Elderly age and prior autologous transplantation have a deleterious effect on survival following allogeneic peripheral blood stem cell transplantation with reduced-intensity conditioning: results from the Spanish multicenter prospective trial. Bone Marrow Transplant. 2004;33:477–482
- Allogeneic marrow transplantation for refractory anemia: a comparison of two preparative regimens and analysis of prognostic factors. Blood. 1996;87:51–58
- Outcome after allogeneic bone marrow transplant for leukemia in older adults. JAMA. 1993;270:57–60
- Allogeneic bone marrow transplantation for acute leukemia in patients over the age of 40 years (Acute Leukemia Working Party of the European Group for Bone Marrow Transplantation (EBMT)). Leukemia. 1997;11:416–419
- Allogeneic and syngeneic marrow transplantation for myelodysplastic syndrome in patients 55 to 66 years of age. Blood. 2000;95:1188–1194
- . Allogeneic hemopoietic stem cell transplantation in patients with myelodysplastic syndrome or myelofibrosis. Int J Hematol. 2002;76(Suppl 2):29–34
- Iron overload might increase transplant-related mortality in haematopoietic stem cell transplantation. Bone Marrow Transplant. 2002;29:987–989
- Karyotype in acute myeloblastic leukemia: prognostic significance for bone marrow transplantation in first remission: a European Group for Blood and Marrow Transplantation study (Acute Leukemia Working Party of the European Group for Blood and Marrow Transplantation (EBMT)). Blood. 1997;90:2931–2938
- Impact of cytogenetic abnormalities on outcome of bone marrow transplants in acute myelogenous leukemia in first remission. Bone Marrow Transplant. 1995;16:203–208
- Parameters for predicting allogeneic PBSCT outcome of acute myeloid leukemia: cytogenetics at presentation versus disease status at transplantation. Ann Hematol. 2005;84:25–32
- Impact of cytogenetics on outcome of stem cell transplantation for acute myeloid leukemia in first remission: a large-scale retrospective analysis of data from the Japan Society for Hematopoietic Cell Transplantation. Int J Hematol. 2004;79:495–500
- Cytogenetic abnormalities in primary myelodysplastic syndrome are highly predictive of outcome after allogeneic bone marrow transplantation. Blood. 1998;92:1910–1917
- . Clinical relevance of mutations and gene-expression changes in adult acute myeloid leukemia with normal cytogenetics: are we ready for a prognostically prioritized molecular classification?. Blood. 2007;109:431–448
PII: S1083-8791(07)00457-0
doi:10.1016/j.bbmt.2007.07.016
© 2008 American Society for Blood and Marrow Transplantation. Published by Elsevier Inc. All rights reserved.
Volume 14, Issue 1 , Pages 28-35, January 2008
