Relapse of malignant disease remains the major complication in patients with acute myeloid leukemia (AML) or myelodysplastic syndrome (MDS) after hematopoietic cell transplantation (HCT) with reduced-intensity conditioning (RIC). In this study, we investigated the predictive value of disease-specific markers (DSMs), donor chimerism (DC) analysis of unsorted (UDC) or CD34+ sorted cells and Wilms’ tumor gene 1 (WT1) expression. Eighty-eight patients with AML or MDS were monitored after allogenic HCT following 2 Gy total-body irradiation with (n=84) or without (n=4) fludarabine 3 × 30 mg/m2, followed by cyclosporin A and mycophenolate mofetil. DSMs were determined by fluorescence in situ hybridization (FISH) and WT1 expression by real-time polymerase chain reaction. Chimerism analysis was performed on unsorted or CD34+ sorted cells, by FISH or short tandem repeat polymerase chain reaction. Twenty-one (24%) patients relapsed within 4 months after HCT. UDC, CD34+ DC and WT1 expression were each significant predictors of relapse with sensitivities ranging from 53 to 79% and specificities of 82–91%. Relapse within 28 days was excluded almost entirely on the basis of WT1 expression combined with CD34+ DC kinetics. Monitoring of WT1 expression and CD34+ DC predict relapse of AML and MDS after RIC-HCT.
Reduced-intensity conditioning (RIC) hematopoietic stem cell transplantation (HCT) relies mainly on immunological effects for disease control and is a promising treatment for patients with acute myeloid leukemia (AML) and myelodysplastic syndrome (MDS) not eligible for conventional conditioning regimens because of age or contraindications. Although a variety of RIC protocols has been developed,1, 2 recurrence of the disease remains the major problem,3, 4, 5 with relapse rates of 39% at 2 years in elderly patients conditioned with 2 Gy total body irradiation with or without fludarabine.6 A number of strategies are available for the treatment of relapse, including reduction of immunosuppression,7 chemotherapy or donor lymphocyte infusion.8 The success of the treatment depends heavily on the tumor load (molecular, cytogenetic or hematological relapse (HR)) and the kinetics of the disease.
As a consequence, the prediction of HR is an important factor in the management of AML patients following RIC-HCT. However, in contrast to Philadelphia chromosome-positive chronic myeloid leukemia, in which quantification of BCR-ABL1 (Abelson gene) can be used to detect early relapse or even predict relapse,9 no common genetic markers are available for AML. Only in subgroups of patients do specific mutations and fusion gene transcripts permit molecular monitoring of minimal residual disease (MRD).10, 11, 12 Nucleophosmin 1 is mutated in only 48%13, 14 and FMS-like tyrosine kinase 3 in only 25–30% of patients with AML and the specific FMS-like tyrosine kinase 3 mutation can be unstable during follow-up.10, 15, 16 Furthermore, the highly distinctive fusion genes such as AML1 -ETO, CBFβ-MYH11 and PML-RARα are rare.
In contrast, overexpression of the non-mutated Wilms’ tumor gene 1 (WT1) is found in 89–100%10, 17, 18, 19, 20 of patients with AML and MDS, making this marker an ideal tool for monitoring MRD. WT1 expression has already been shown to be a helpful marker in patients with AML after chemotherapy,19, 21, 22, 23 after autologous24 and after conventional allogeneic HCT.25, 26, 27
In this analysis, the ability of different markers to predict relapse following allogeneic HCT with RIC was analyzed both alone and in combination. Peripheral blood (PB) cell WT1 expression levels quantified by real-time polymerase chain reaction (RT-PCR) were compared with the detection of disease-specific (DS) markers by fluorescence in situ hybridization (FISH). Finally, levels of donor chimerism in unsorted cells (UDC) and in sorted CD34+ subpopulations were compared with the other parameters. The results clearly show that the combination of the CD34+ donor chimerism kinetics in bone marrow (BM) and WT1 expression in PB provides a very strong indicator of relapse within 28 days of sampling and can be used as a tool to detect early relapse in patients after allogeneic HCT following RIC.
Materials and methods
Between July 1999 and July 2005, 88 consecutive patients with a median age of 61 (range 22–74) years with either AML (n=68) or intermediate or high-risk MDS (n=20) in complete remission 1 (n=41), complete remission 2 (n=25) or in a more advanced disease status (n=22) were transplanted at the Department of Hematology and Oncology, University of Leipzig, Germany after written informed consent was obtained from all patients. Patient and transplant characteristics are summarized in Table 1. The conditioning regimen consisting of fludarabine 30 mg/kg body weight from day −4 to −2 (n=84) combined with low-dose total body irradiation (200 cGy, n=87 or 300cGy, n=1) at day 0 was followed by the infusion of granulocyte colony-stimulating factor-mobilized PB stem cells from related (n=22) or unrelated (n=66) donors. All patients received cyclosporin A and mycophenolate mofetil.28 Mycophenolate mofetil was stopped at day 27 following related HCT and tapered from days 40 to 96 following unrelated HCT. Cyclosporin A was reduced from day 84 or day 180 following related and unrelated HCT, respectively. Immunosuppression was prolonged in case of graft-versus-host disease grade ⩾II or rapidly reduced in case of HR, defined as >5% blasts in the BM by cytomorphology. Engraftment was documented by full donor T-cell chimerism. HRs diagnosed by cytomorphology occurred in 37 patients (43%) after a median time of 122 (range 18–1456) days after HCT. Immunosuppression was tapered in two AML patients following a decrease in CD34+ chimerism in the absence of elevated blasts. These two patients were excluded from the prediction analysis.
Sampling of BM and PB
BM aspirations for cytomorphology, UDC, CD34+ DC and, where possible, DS-FISH were performed at days 28, 56 and 84 after HCT or until relapse as a routine follow-up monitoring schedule. WT1 expression was assessed retrospectively from stored PB samples from the same time points.
DC in unsorted and in CD34+ BM cells
Chimerism analyses were performed by FISH- or PCR-based amplification of short tandem repeat sequences for sex-mismatched or -matched donor recipient pairs, respectively.29 Briefly, 300 interphases were analyzed by X/Y FISH (SO CEP X/SG CEP Y; Vysis, Stuttgart, Germany) with a lab-specific sensitivity of one cell in a background of 299 cells. Short tandem repeat-PCR was performed on DNA of 107 total white cells with a similar sensitivity.
CD34+ cells were sorted by fluorescence-activated cell sorting using the FACSCalibur Flow Cytometer (Becton Dickinson, Heidelberg, Germany) and the CD34PE8G12 antibody (Becton Dickinson, Heidelberg, Germany).
FISH of DS markers
FISH analyses were performed from BM cells in 38 patients with molecular markers according to standard protocols on a minimum of 200 interphases.30 In detail, CEP 8 SO, LSI 5q EGR 1SO/D5S23 SG, LSI D7S522 SO/CEP 7SG, LSI MLL Dual Color, LSI CBFB, inv16 DualColor, LSI D20S108 SO, LSI D13S319 SO and LSI p53SO were used to detect abnormalities. The cutoffs for positivity, determined in the background of healthy individuals, range between 1 and 4.5%. All probes were purchased from ABBOTT (Wiesbaden, Germany).
Quantitative assessment of WT1 expression in PB
Mononuclear cells were isolated from PB and stored at −196 °C until use. RNA was extracted from 1 × 107 cells and reverse transcribed as described previously.31 RT-PCR for WT1 was performed in duplicate and expressed as WT1/10.000 ABL1 transcripts.19 Samples containing <0.5 WT1 transcripts together with >5000 ABL1 transcripts were scored as negative, whereas WT1-negative samples with <5000 ABL1 transcripts were excluded. The assay consistently detected 10 WT1 plasmids in an aqueous dilution series and showed an interassay variation of 0.9 log for low WT1 transcripts (median 1.3, n=8) and 0.4 log for high WT1 transcripts (median 26.4 WT1 per 104 ABL1, n=8), respectively.
WT1 expression in healthy volunteers
The expression levels of WT1 in PB cells of healthy volunteers (n=10) were compared with the expression levels of patients with MDS (n=3) or AML (n=10). The WT1 expression levels in healthy individuals were significantly lower (median 0.7, range 0.01–3.1 WT1 per 104 ABL1) than that in patients with MDS (median 65.0, range 36.8–88.7 WT1 per 104 ABL1, P=0.007) or in patients with AML (median 548.5, range 5.5–4649.1 WT1 per 104 ABL1, P<0.0001) as shown in Figure 1a.
Prediction of relapse
As we were interested in predicting relapse early after RIC-HCT and during immunosuppression taper, the 4-month evaluation period was chosen as the most suitable to assess the predictability of imminent relapse (defined as occurring within 28 days between evaluations). Of the analyzed patients, 21 (24.4%) relapsed during this period, accounting for 56.7% of all relapses after a median follow-up of 15.9 (range 1.3–119) months after HCT. Six patients developed relapse before the first evaluation point at day +28 and 15 patients thereafter. All measurements were analyzed as absolute levels and as kinetics between two subsequent measurements (changes over the 28-day period between two evaluation points), with the exception of the FISH results, in which low sample numbers prevented kinetic assessment.
Pearson's correlation was used to analyze correlation between WT1 values and BM blasts. Analyses of differences were calculated by the Mann–Whitney U-test or Student's t-test for unpaired data (e.g. comparison of patients with and without relapse) and with Wilcoxon's signed rank test or paired Student's t-test for paired data (e.g. comparison of subsequent measurements). To analyze the diagnostic power of the investigated techniques, receiver-operating characteristics (ROC) were used. By this technique, the percentage of true-positive findings (sensitivity) is plotted against the percentage of false-positive findings (1-specificity) for all possible cutoff values of a diagnostic variable. Random allocation would result in an area under the resulting curve of 0.5. Values greater than 0.5 indicate positive diagnostic value. For the prediction of relapse within the 4-month post-HCT evaluation period, absolute values and changes between two subsequent measurements (kinetics) of WT1 expression, UDC and CD34+ chimerism were analyzed. For diagnostic variables, we considered the highest differences between subsequent measurements before relapse or during the observed time period for patients without relapse. Alternatively, we considered the absolute values before relapse or the peak value of measurements for patients without relapse, respectively. With the help of the ROC curves, we selected suitable cutoff values of the diagnostic variables by maximizing the sum of sensitivity and specificity. Cutoffs were rounded for clinical use. Characteristics of these tests were determined by the analysis of corresponding 2 × 2 contingency tables of test- positive and -negative cases (with relapse) and controls (without relapse). As small numbers were expected in the contingency tables, we also calculated exact 95% confidence intervals for both sensitivity and specificity estimates. Likewise, we calculated conditional maximum-likelihood estimates for the odds ratios and corresponding exact confidence intervals and P-values (Fisher's exact test). These calculations were performed using the statistical software package ‘R’ (www.r-project.org).32 All other analyses were performed by Excel 7.0 for Windows and SPSS 12.0 for Windows software package (SPSS, Chicago, IL, USA).
To investigate whether the identified markers are independent predictors, we used regression techniques to analyze the univariate and multivariate influence of additional baseline and follow-up risk factors (see Table 3 for a list of factors) on the end points: (1) 4-month relapse-free survival, (2) overall relapse-free survival and (3) overall survival. We used exact logistic regression for the end point 4-month relapse-free survival and Cox regression for the other end points. In case of univariate significance, the corresponding factors were multivariately analyzed using a forward model selection algorithm. Calculations were performed using ‘R’, except for the exact logistic regression for which LogXact-8 (Cytel Inc., Cambridge, MA, USA) was applied.
WT1 expression pattern before/after HCT and correlation with BM blasts
We first determined whether WT1 transcripts measured in the PB (median 3.0 (range 0.01–4649.1) WT1 per 104 ABL1, n=260) correlated with BM blasts (median 3.5 (range 0–82)%, n=265). A strong correlation was observed in the paired 239 assessment points (r=0.707, P<0.0001, data not shown).
WT1 transcripts showed characteristic fluctuations in both AML and MDS patients. Immediately before HCT, with the majority of patients in complete remission, median WT1 transcripts were in the normal range of healthy volunteers (median 0.7 (range 0.01–3.1) WT1 per 104 ABL1, n=10 vs 1.5 (range 0.01–810.1), n=53, and 1.7 (range 0.01–2322.2), n=14 for AML and MDS, respectively (P=NS)) and clearly lower than that in AML or MDS patients with HR (Figures 1a–c). On day 28, WT1 expression in patients with an impending relapse increased to a median of 11.9 (range 0.01–589.9) in AML and median of 6.5 (range 4.2–15.2) WT1 per 104 ABL1 in MDS patients (Figures 1b and c). There was a clear difference in transcripts between patients subsequently relapsing and non-relapsing. This difference was significant for AML patients (P<0.05), but not for MDS patients (P=0.06). In contrast, the difference in WT1 transcript levels for patients with or without HR at day 84 showed borderline association for AML (P=0.058), but not for MDS (Figures 1b and c). Patients without relapse maintained WT1 transcript levels comparable with those of healthy individuals throughout follow-up (Figures 1b and c). In contrast, patients with relapse showed continuously increasing WT1 levels peaking at the time of relapse. Importantly, increases in WT1 transcripts were already detected 28 days before HR (Figure 1d). Subsequently, WT1 levels 28 days before relapse were included in the model to identify the most suitable technique to predict relapse.
Unsorted donor chimerism
Donor chimerism analyses were performed on unsorted cells from 86 patients. Of the samples, 19 (7.3%) were excluded because of relapse before sampling and 18 (7.0%) because of insufficient sample quality, leaving a total of 221 samples eligible for analysis. Patients with a relapse within 28 days of sampling (n=15) showed a significantly lower median UDC of 89 (range 35–100)% than those without relapse (median 98 (range 21.7–100)%, n=63, P=0.01). In contrast, UDC kinetics in all eight evaluable relapsed patients showed no decrease (median +0.4 (range −15.3 to +8)%) 28 days before relapse and were similar to the kinetics of the patients without relapse (median 0 (range −20 to +56.3)%, n=60, P=0.883). ROCs were then used to determine the best clinical cutoff with corresponding sensitivities and specificities for the prediction of relapse. Absolute UDC values ⩽90% predict a relapse with a sensitivity of 53% and a specificity of 86% (Table 2), with seven of 15 patients (AML: nos. 41, 43, 45, 46, 47, 48; MDS: no. 51) having had UDC >90% before relapse. For a patient with UDC of ⩽90%, the odds ratio for relapsing within the next 28 days was calculated to be 6.6 (Table 2). In contrast, UDC kinetics was a poor predictor of relapse with a low sensitivity (25%) and a specificity of 92%.
Chimerism of CD34+ sorted cells
Eighty-six patients were monitored for CD34+ sorted donor cell chimerism (CD34+ DC). A total of 195 samples from 86 patients were available for the prediction analysis, 63 samples (24.4%) being excluded because of previous relapse (n=26, 10.1%) or insufficient sample quality (n=37, 14.3%). In AML, but not in MDS, there was a clear prognostic association between impending HR and a low CD34+ DC at day 28 (P<0.001), 56 (P<0.01) and 84 (P=0.014) (data not shown). Furthermore, the median CD34+ DC was found to be significantly lower in patients undergoing subsequent relapse within 28 days (median 73.3 (range 0–100)%, n=13) compared with those without relapse (median 99.3 (range 28–100)%, n=62, P=0.001). Accordingly, an analysis of CD34+ DC kinetics revealed a significant decrease (median –15 (range –76 to 0)%) in patients undergoing relapse within 28 days (n=7) compared with those without relapse (median 0 (range –20 to +15.5)%, n=58, P<0.0001). The ROC analysis revealed an absolute value of ⩽90% CD34+ DC and ⩾5% decrease in CD34+ DC between two measurements to be the best clinical cutoff values with sensitivities of 62 and 71% and specificities of 90 and 91%, respectively (Table 2). Five of the 13 patients relapsing with 28 days of CD34+ DC determination (AML: nos. 41, 44, 45, 47, 48) nonetheless had a high CD34+ DC of >90%, whereas only two of the seven patients for whom pre-relapse kinetic data were available (AML: nos. 45, 47) failed to show a decrease in CD34+ chimerism ⩾5% before relapse. An estimation of the odds ratios of relapsing within the next 28 days revealed CD34+ DC reduction to yield a higher odds ratio (23.9) than did the CD34+ DC absolute values (14.0; Table 2).
FISH of DS markers in BM
A total of 42 (48%) patients had karyotypic aberrations. Although FISH probes were available for 38 of these, 82 of the 152 samples were excluded for the prediction model owing to previous disease recurrence (n=13, 8.6%) or were unavailable owing to violations of the MRD monitoring protocol (n=69, 45.4%). Analysis performed on data from 30 patients, five of whom underwent relapse, revealed no significant increase above cutoff values 28 days preceding relapse. This is a reflection of the low sensitivity (40%) and high specificity (96%) of the technique (Table 2). FISH was therefore unable to predict relapse within 28 days in our study population and has been excluded from further analysis.
WT1 expression in PB
WT1 expression levels were determined in a total of 260 samples from 86 patients. Eighty-four samples were not available for the prediction model owing to previous relapse (n=16, 4.6%), MRD protocol violations (n=48, 14%) or insufficient sample quality (n=20, 5.8%).
Data from 14 relapsed patients were available for the analysis of absolute WT1 transcript levels, revealing significantly higher (median 31.4 (range 3–1970.5) WT1 per 104 ABL1) WT1 transcript levels in samples taken 28 days before relapse compared with the rest (median 5.7 (range 0.01–192.5) WT1 per 104 ABL1, n=64, P<0.0001). Similarly, patients (n=12) undergoing relapse within 28 days had a higher increase in WT1 transcript kinetics (median 1.1 (range 0–2.4) log) compared with patients without relapse (median 0.4 (range 0–1.6) log, n=60, P=0.023). ROC analysis identified transcript levels of ⩾24 WT1 per 104 ABL1 and an increase of ⩾1 log between sampling points as the most appropriate cutoffs for predicting relapse within 28 days with a sensitivity of 79 and 58% and a specificity of 89 and 82%, respectively (Table 2).
Only three of 14 patients (AML: nos. 40, 50; MDS: no. 42) relapsed within 28 days with WT1 per 104 ABL1 transcripts <24, whereas five of 12 patients (AML: nos. 40, 41, 49; MDS: nos. 42, 53) had no increase ⩾1 log WT1 transcripts before relapse. The odds ratio of a relapse within 28 days of sampling for patients above the estimated cutoffs is 27.6 for the WT1 absolute values and 6.0 for the WT1 kinetics (Table 2).
Comparison of the prediction factors and combinations
Both the absolute levels and kinetics of WT1 expression, as well as CD34+ DC and absolute UDC are each associated with relapse within 28 days of sampling. For each technique, the corresponding area under the ROC (area under the curve) corresponds to the diagnostic power to predict imminent relapse (Figure 2, Table 2). Thus, CD34+ DC kinetics (area under the curve=0.895) and WT1 transcripts (area under the curve=0.855) were identified as the most powerful parameters in the ROC model, consistent with the highest odds ratios (Table 2) for predicting a relapse within 28 days. The combination of these parameters was then analyzed to further optimize sensitivity or specificity of prediction.
Sixty-four patients, including seven who developed relapse, were available to analyze the predictive power of combining WT1 expression (cutoff ⩾24 WT1 per 104 ABL1 transcripts) and decrease in CD34+ DC (cutoff ⩾5% decrease). All seven patients with relapse were preceded by at least one positive indication, resulting in a sensitivity of prediction of 100%, with a specificity of 84%. On the other hand, double positives were extremely rare in those cases in which relapse did not occur within the next 28 days, corresponding to a specificity of 98% with a sensitivity of 57% (Table 2). It is therefore possible to exclude a relapse almost completely if WT1 transcripts are lower than ⩾24 WT1 per 104 ABL1 and CD34+ DC do not decrease ⩾5%. Only one patient out of 34 failed these criteria, having WT1 levels of 192.5 WT1 per 104 ABL1 with a decrease of 18% in CD34+ DC without developing subsequent relapse.
Prediction of relapse within 3 months
Having shown that CD34+ DC kinetics and WT transcripts provide a reliable prediction of relapse within 28 days, we tested their ability to predict relapse up to 84 days following sampling. Sixty-three patients were available for the analysis of WT1 transcripts alone. Only one of the five relapsing patients already had a WT1 expression level above the cutoff 84 days before relapse. Similarly, six relapsed patients were available for the CD34+ analysis, and only one of these showed a CD34+ DC decrease exceeding the cutoff 84 days before relapse. Neither WT1 transcript levels, CD34+ DC kinetics nor the combination of both provide a reliable prediction of relapse within the following 84 days.
Analysis of independent factors to predict relapse by 4 months after HCT, progression-free survival and overall survival
As CD34+ DC kinetics and WT1 transcripts provide the best prediction of relapse within 28 days, other baseline and follow-up factors have been analyzed to identify independent prognostic factors for relapse by 4 months, progression-free survival and overall survival (Table 3). In univariate analysis, a CD34+ DC decrease of ⩾5% and WT transcripts of ⩾24 WT1 per 104 ABL1 remained the only significant and independent parameters predictive for relapse by 4 months after HCT and progression-free survival. In contrast, a decrease of >50% in platelets, the use of an HLA-mismatched donor and a increase of white blood cells by more than 50% in the model with WT1 were additional independent variables for overall survival.
In this analysis, we have investigated the prediction of HR in patients with AML and MDS following HCT, comparing techniques of unsorted and CD34+ sorted donor chimerism in BM, DSM-FISH in BM and WT1 expression in PB. The overall aim of this analysis was to identify the most informative parameter or parameter combination upon which to base early intervention before HR occurs. UDC has the advantage of being independent of DS characteristics, and has been widely used to detect relapse after HCT.33 We did indeed find that UDC ⩽90% is associated with relapse within 28 days of sampling, and also that the sensitivity of this test was only 53%. A decrease in UDC >10% and also UDC values from PB (data not shown) had an even lower positive predictive value and was not associated with HR. In patients harboring an appropriate marker, DSM-FISH analysis showed a positive 40% ability to predict HR within 28 days. On this basis, we conclude that UDC and FISH are insufficient to predict HR in advance, consistent with observations of other groups.34, 35
The majority of AML cells express the CD34 antigen, and we and others have reported an association between a low or decreasing CD34+ DC and relapse after HCT with conventional conditioning,36 2 Gy total body irradiation,7 other RIC regimens37 or both RIC and conventional conditioning.33 We report here that an absolute level of ⩽90% CD34+ DC and a decrease of >5% in CD34+ DC provide sensitive and specific predictions of relapse within 28 days. The sensitivities of ⩽90% CD34+ DC or >5% decrease in CD34+ DC were 62 and 71%, respectively.
Taken together, these sensitivities are comparable with the 80% reported from a cutoff of ⩽80% CD34+ PB DC after a different conditioning regimen.33 According to our ROC results, impending relapse is associated more strongly with the kinetic of CD34+ DC reduction than with the absolute levels of CD34+ DC, although both parameters provide a good indication of imminent relapse. Of note, we obtained this result in an AML and MDS patient cohort after a homogeneous 2 Gy total body irradiation-based conditioning and similar immunosuppression in the absence of information concerning the CD34 status at diagnosis. Interestingly, our results were not altered by exclusion of the CD34− AML patients and those for whom initial CD34 status was unknown (data not shown). Patients who developed a relapse in our study cohort had significantly higher WT1 expression levels in blood cells at day +28 than did those without an impending relapse. Combining absolute levels and increases in WT1 mRNA levels in PB provides an even stronger indicator for an imminent relapse. The best indication was derived from WT1 levels in excess of the cutoff of 24 WT1 per 104 ABL1 transcripts: a level which is almost 10-fold higher than the highest level of WT1 expression in our cohort of healthy volunteers. Specifically, if a patient had ⩾24 WT1 per 104 ABL1 transcripts, the risk of developing HR within 4 weeks was 79%. Only three patients had no ⩾24 WT1 per 104 ABL1 transcripts 4 weeks before HR. This may be explained in part by the recently recognized phenomenon of WT1 exon 7 mutations, which might have prevented sufficient primer binding and compromised the performance of the RT-PCR amplification.20, 37, 38 WT1 expression above the cutoff corresponds to an odds ratio of impending relapse of 27.6. This is comparable with the predictive power of MRD levels determined via T-cell receptor and immunoglobulin gene rearrangement in standard risk acute lymphoblastic leukemia, in which an MRD level higher than 10−4 by 16 weeks was associated with a relapse risk of 95% within 3 years.39 When comparing our most informative parameters, the absolute levels of WT1 expression appear to be a better predictor of relapse than the WT1 kinetics, whereas the opposite was true for CD34+ DC. This suggests that the parameters of gene expression level averaged over all cells on the one hand and frequency of a distinct cell population on the other may provide complementary information concerning impending relapse. Indeed, if both parameters (decrease of >5% in CD34+ DC and a WT1 transcript level of >24/104 ABL1 transcripts) are taken into consideration, 100% of patients with relapse were identified with a specificity of 84%. Similarly, when both markers are below the cutoffs, relapse within 28 days can be excluded almost entirely (specificity 98%, sensitivity 57%). Unsurprisingly, neither of the parameters tested could provide an indication of relapse 84 days in advance. Although both CD34+ DC kinetics and WT1 levels remained also the only two independent factors for relapse at 4 months and progression-free survival in a multivariate model, the use of a mismatched donor, a decrease in platelets by more than 50% and a increase in WBC by more than 50% in the model with WT1 only proved to be additional factors associated with a worse overall survival. This confirms the finding that a mismatched donor is an independent factor for non-relapse mortality in a very similar patient cohort using the same HCT technique.40 The decrease in platelets and the increase of WBC has not been previously identified, but should be included in future evaluations.
On the basis of this study, we have adapted our MRD monitoring protocol for patients with AML and MDS after HCT with RIC: during the 6 months post-HCT period with the highest relapse incidence,3, 4, 5 monthly monitoring of MRD by the detection of CD34+ DC from aspirated BM together with the assessment of WT1 transcripts from PB can provide an accurate indication of impending relapse. The identification of impending relapse in this way provides a basis for optimizing treatment of relapse, either by tapering of immunosuppression over 3–4 weeks with the risk of severe graft-versus-host disease or by chemotherapeutic intervention,41 in which the application of low toxicity agents such as sorafenib42 and demethylating agents43 may provide new treatment perspectives.
Storb R . Can reduced-intensity allogeneic transplantation cure older adults with AML? Best Pract Res Clin Haematol 2007; 20: 85–90.
Niederwieser D, Lange T, Cross M, Basara N, Al-Ali H . Reduced intensity conditioning (RIC) haematopoietic cell transplants in elderly patients with AML. Best Pract Res Clin Haematol 2006; 19: 825–838.
Tauro S, Craddock C, Peggs K, Begum G, Mahendra P, Cook G et al. Allogeneic stem-cell transplantation using a reduced-intensity conditioning regimen has the capacity to produce durable remissions and long-term disease-free survival in patients with high-risk acute myeloid leukemia and myelodysplasia. J Clin Oncol 2005; 23: 9387–9393.
Hallemeier CL, Girgis MD, Blum WG, Brown RA, Khoury HJ, Devine SM et al. Long-term remissions in patients with myelodysplastic syndrome and secondary acute myelogenous leukemia undergoing allogeneic transplantation following a reduced intensity conditioning regimen of 550 cGy total body irradiation and cyclophosphamide. Biol Blood Marrow Transplant 2006; 12: 749–757.
Schmid C, Schleuning M, Schwerdtfeger R, Hertenstein B, Mischak-Weissinger E, Bunjes D et al. Long-term survival in refractory acute myeloid leukemia after sequential treatment with chemotherapy and reduced-intensity conditioning for allogeneic stem cell transplantation. Blood 2006; 108: 1092–1099.
Hegenbart U, Niederwieser D, Sandmaier BM, Maris MB, Shizuru JA, Greinix H et al. Treatment for acute myelogenous leukemia by low-dose, total-body, irradiation-based conditioning and hematopoietic cell transplantation from related and unrelated donors. J Clin Oncol 2006; 24: 444–453.
Al-Ali HK, Nehring C, Krahl R, Becker C, Leiblein S, Edelmann J et al. Donor CD34+ cell chimerism at day 28 and chronic graft-versus-host disease (GvHD) but not high-risk cytogenetics influence outcome of allogeneic hematopoetic cell transplantation (HCT) following reduced intensity conditioning (RIC) in patients with AML and MDS. ASH Annual Meeting Abstracts 2006; 108: 547.
Mielcarek M, Martin PJ, Maloney DG, Storb R, Sandmaier BM . Outcomes among patients with recurrent high-risk hematologic malignancy after nonmyeloablative versus myeloablative allogeneic hematopoietic cell transplantation. ASH Annual Meeting Abstracts 2006; 108: 262.
Lange T, Deininger M, Brand R, Hegenbart U, Al-Ali H, Krahl R et al. BCR-ABL transcripts are early predictors for hematological relapse in chronic myeloid leukemia after hematopoietic cell transplantation with reduced intensity conditioning. Leukemia 2004; 18: 1468–1475.
Kern W, Schoch C, Haferlach T, Schnittger S . Monitoring of minimal residual disease in acute myeloid leukemia. Crit Rev Oncol Hematol 2005; 56: 283–309.
Gozzetti A, Le Beau MM . Fluorescence in situ hybridization: uses and limitations. Semin Hematol 2000; 37: 320–333.
Thiede C . Diagnostic chimerism analysis after allogeneic stem cell transplantation: new methods and markers. Am J Pharmacogenom 2004; 4: 177–187.
Dohner K, Schlenk RF, Habdank M, Scholl C, Rucker FG, Corbacioglu A et al. Mutant nucleophosmin (NPM1) predicts favorable prognosis in younger adults with acute myeloid leukemia and normal cytogenetics: interaction with other gene mutations. Blood 2005; 106: 3740–3746.
Gorello P, Cazzaniga G, Alberti F, Dell’Oro MG, Gottardi E, Specchia G et al. Quantitative assessment of minimal residual disease in acute myeloid leukemia carrying nucleophosmin (NPM1) gene mutations. Leukemia 2006; 20: 1103–1108.
Schnittger S, Schoch C, Kern W, Hiddemann W, Haferlach T . FLT3 length mutations as marker for follow-up studies in acute myeloid leukaemia. Acta Haematol 2004; 112: 68–78.
Scholl S, Krause C, Loncarevic IF, Muller R, Kunert C, Wedding U et al. Specific detection of Flt3 point mutations by highly sensitive real-time polymerase chain reaction in acute myeloid leukemia. J Lab Clin Med 2005; 145: 295–304.
Cilloni D, Saglio G . WT1 as a universal marker for minimal residual disease detection and quantification in myeloid leukemias and in myelodysplastic syndrome. Acta Haematol 2004; 112: 79–84.
Keilholz U, Menssen HD, Gaiger A, Menke A, Oji Y, Oka Y et al. Wilms’ tumour gene 1 (WT1) in human neoplasia. Leukemia 2005; 19: 1318–1323.
Ostergaard M, Olesen LH, Hasle H, Kjeldsen E, Hokland P . WT1 gene expression: an excellent tool for monitoring minimal residual disease in 70% of acute myeloid leukaemia patients—results from a single-centre study. Br J Haematol 2004; 125: 590–600.
Cilloni D, Renneville A, Hermitte F, Hills RK, Daly S, Jovanovic JV et al. Real-time quantitative polymerase chain reaction detection of minimal residual disease by standardized WT1 assay to enhance risk stratification in acute myeloid leukemia: a European LeukemiaNet study. J Clin Oncol 2009; 27: 5195–5201.
Weisser M, Kern W, Rauhut S, Schoch C, Hiddemann W, Haferlach T et al. Prognostic impact of RT-PCR-based quantification of WT1 gene expression during MRD monitoring of acute myeloid leukemia. Leukemia 2005; 19: 1416–1423.
Lapillonne H, Renneville A, Auvrignon A, Flamant C, Blaise A, Perot C et al. High WT1 expression after induction therapy predicts high risk of relapse and death in pediatric acute myeloid leukemia. J Clin Oncol 2006; 24: 1507–1515.
Cilloni D, Messa F, Arruga F, Defilippi I, Gottardi E, Fava M et al. Early prediction of treatment outcome in acute myeloid leukemia by measurement of WT1 transcript levels in peripheral blood samples collected after chemotherapy. Haematologica 2008; 93: 921–924.
Osborne D, Frost L, Tobal K, Liu Yin JA . Elevated levels of WT1 transcripts in bone marrow harvests are associated with a high relapse risk in patients autografted for acute myeloid leukaemia. Bone Marrow Transplant 2005; 36: 67–70.
Ogawa H, Ikegame K, Kawakami M, Tamaki H . WT1 gene transcript assay for relapse in acute leukemia after transplantation. Leuk Lymphoma 2004; 45: 1747–1753.
Ogawa H, Tamaki H, Ikegame K, Soma T, Kawakami M, Tsuboi A et al. The usefulness of monitoring WT1 gene transcripts for the prediction and management of relapse following allogeneic stem cell transplantation in acute type leukemia. Blood 2003; 101: 1698–1704.
Candoni A, Tiribelli M, Toffoletti E, Cilloni D, Chiarvesio A, Michelutti A et al. Quantitative assessment of WT1 gene expression after allogeneic stem cell transplantation is a useful tool for monitoring minimal residual disease in acute myeloid leukemia. Eur J Haematol 2009; 82: 61–68.
Niederwieser D, Maris M, Shizuru JA, Petersdorf E, Hegenbart U, Sandmaier BM et al. Low-dose total body irradiation (TBI) and fludarabine followed by hematopoietic cell transplantation (HCT) from HLA-matched or mismatched unrelated donors and postgrafting immunosuppression with cyclosporine and mycophenolate mofetil (MMF) can induce durable complete chimerism and sustained remissions in patients with hematological diseases. Blood 2003; 101: 1620–1629.
Bryant E, Martin PJ . Documentation of engraftment and characterization of chimerism following hematopoietic cell transplantation. In: Thomas ED, Blume KG, Forman SJ (eds). Hematopoietic Cell Transplantation, 2 edn. Boston, 1999, pp 197–206.
Bumm T, Muller C, Al-Ali HK, Krohn K, Shepherd P, Schmidt E et al. Emergence of clonal cytogenetic abnormalities in Ph− cells in some CML patients in cytogenetic remission to imatinib but restoration of polyclonal hematopoiesis in the majority. Blood 2003; 101: 1941–1949.
Lange T, Niederwieser DW, Deininger MW . Residual disease in chronic myeloid leukemia after induction of molecular remission. N Engl J Med 2003; 349: 1483–1484.
Ihaka R, Gentleman R . R a language for data analysis and graphics. J Comp Graph Stat 1996; 5: 299–314.
Bornhauser M, Oelschlaegel U, Platzbecker U, Bug G, Lutterbeck K, Kiehl MG et al. Monitoring of donor chimerism in sorted CD34+ peripheral blood cells allows the sensitive detection of imminent relapse after allogeneic stem cell transplantation. Haematologica 2009; 94: 1613–1617.
Bacher U, Kern W, Schoch C, Schnittger S, Hiddemann W, Haferlach T . Evaluation of complete disease remission in acute myeloid leukemia: a prospective study based on cytomorphology, interphase fluorescence in situ hybridization, and immunophenotyping during follow-up in patients with acute myeloid leukemia. Cancer 2006; 106: 839–847.
Mancini M, Cedrone M, Diverio D, Emanuel B, Stul M, Vranckx H et al. Use of dual-color interphase FISH for the detection of inv(16) in acute myeloid leukemia at diagnosis, relapse and during follow-up: a study of 23 patients. Leukemia 2000; 14: 364–368.
Zeiser R, Spyridonidis A, Wasch R, Ihorst G, Grullich C, Bertz H et al. Evaluation of immunomodulatory treatment based on conventional and lineage-specific chimerism analysis in patients with myeloid malignancies after myeloablative allogeneic hematopoietic cell transplantation. Leukemia 2005; 19: 814–821.
Nyvold CG, Stentoft J, Braendstrup K, Melsvik D, Moestrup SK, Juhl-Christensen C et al. Wilms’ tumor 1 mutation accumulated during therapy in acute myeloid leukemia: Biological and clinical implications. Leukemia 2006; 20: 2051–2054.
Gaidzik VI, Schlenk RF, Moschny S, Becker A, Bullinger L, Corbacioglu A et al. Prognostic impact of WT1 mutations in cytogenetically normal acute myeloid leukemia (AML): A study of the German–Austrian AML Study Group (AMLSG). Blood 2009; 113: 4505–4511.
Bruggemann M, Raff T, Flohr T, Gokbuget N, Nakao M, Droese J et al. Clinical significance of minimal residual disease quantification in adult patients with standard-risk acute lymphoblastic leukemia. Blood 2006; 107: 1116–1123.
Gyurkocza B, Storb R, Storer BE, Chauncey TR, Lange T, Shizuru JA et al. Nonmyeloablative allogeneic hematopoietic cell transplantation in patients with acute myeloid leukemia. J Clin Oncol 2010; 28: 2859–2867.
Bethge WA, Storer BE, Maris MB, Flowers ME, Maloney DG, Chauncey TR et al. Relapse or progression after hematopoietic cell transplantation using nonmyeloablative conditioning: effect of interventions on outcome. Exp Hematol 2003; 31: 974–980.
Metzelder S, Wang Y, Wollmer E, Wanzel M, Teichler S, Chaturvedi A et al. Compassionate use of sorafenib in FLT3-ITD-positive acute myeloid leukemia: sustained regression before and after allogeneic stem cell transplantation. Blood 2009; 113: 6567–6571.
Cross M, Jaekel N, Krahl R, Junghanss C, Maschmeyer G, Tran T et al. Low levels of global (LINE) and CDH13 methylation at diagnosis and rapid clearance of marrow blasts correlate with a better haematological response to azacitidine in patients with newly diagnosed and refractory/relapsed AML not eligible for or resistant to chemotherapy: a multi-centre phase I/II-study of the East German Haematology and Oncology Study Group (OSHO). ASH Annual Meeting Abstracts 2009; 114: 2642.
We thank Janet Bogardt, Dagmar Crohn, Christa Döring, Christina Franke, Christine Günther, Evelyn Hennig, Ines Kovacs, Rainer Krahl, Christel Müller and Scarlet Musiol for their excellent assistance.
About this article
Cite this article
Lange, T., Hubmann, M., Burkhardt, R. et al. Monitoring of WT1 expression in PB and CD34+ donor chimerism of BM predicts early relapse in AML and MDS patients after hematopoietic cell transplantation with reduced-intensity conditioning. Leukemia 25, 498–505 (2011). https://doi.org/10.1038/leu.2010.283
Current Treatment Options in Oncology (2020)
Comparison of non‐myeloablative and reduced‐intensity allogeneic stem cell transplantation in older patients with myelodysplastic syndromes
American Journal of Hematology (2019)
Blood Cancer Journal (2019)
Prognostic Impact of Blood MN1 Copy Numbers Before Allogeneic Stem Cell Transplantation in Patients With Acute Myeloid Leukemia
Quantitative chimerism in CD3-negative mononuclear cells predicts prognosis in acute myeloid leukemia patients after hematopoietic stem cell transplantation