Clinical impact of FLT3 mutation load in acute promyelocytic leukemia with t(15;17)/PML-RARA

Abstract

Background Combined treatment with all-trans-retinoic acid and chemotherapy is extremely efficient in patients with acute promyelocytic leukemia with t(15;17)/PML-RARA, but up to 15% of patients relapse.Design and Methods To further clarify the prognostic impact of parameters such as FLT3 mutations, we comprehensively characterized the relation between genetic features and outcome in 147 patients (aged 19.7–86.3 years) with acute promyelocytic leukemia.Results Internal tandem duplications of the FLT3 gene (FLT3-ITD) were detected in 47/147 (32.0%) and tyrosine kinase domain mutations (FLT3-TKD) in 19/147 (12.9%) patients. FLT3-ITD or FLT3-TKD mutation status did not have a significant prognostic impact, whereas FLT3-ITD mutation load, as defined by a mutation/wild-type ratio of less than 0.5 was associated with trends to a better 2-year overall survival rate (86.7% versus 72.7%; P=0.075) and 2-year event-free survival rate (84.5% versus 62.1%, P=0.023) compared to the survival rates of patients with a ratio of 0.5 or more. Besides the t(15;17), an additional chromosomal abnormality was detected in 57 of 147 cases and did not show a significant impact on survival. White blood cell counts of 10×10⁹/L or less versus more than 10×10⁹/L were associated with a better 2-year overall survival rate (88.3% versus 69.4%, respectively; P=0.015), as was male sex (P=0.040). In multivariate analysis, only higher age had a significant adverse impact.Conclusions Prospective trials should further investigate the clinical impact of the FLT3-ITD/wild-type mutation load aiming to evaluate whether this parameter might be included in risk stratification in patients with acute promyelocytic leukemia.

Introduction

Acute promyelocytic leukemia (APL) with the t(15;17)/PML-RARA is characterized by fusion of the promyelocytic leukemia (PML) and retinoic acid receptor alpha (RARA) genes. Administration of all-trans retinoic acid in parallel to anthracycline-based chemotherapy for induction therapy results in complete remission rates of more than 90% in newly diagnosed APL with other cases suffering mostly from early death due to hemorrhage. A 5-year overall survival rate of greater than 80% has been reported in adults.1 Although recent population-based studies suggested a higher early death rate in patients with APL,2–3 the high remission rates of patients who were able to complete treatment suggest that virtually all PML-RARA-positive APL are sensitive to all-trans retinoic acid and anthracycline-based chemotherapy. Still, up to 15% of patients with APL develop clinical or molecular relapses.4–5 A high white blood cell (WBC) count (>10×10/L) is supposed to be the main factor associated with relapse. The Sanz score subdivides APL patients according to peripheral blood counts into three risk groups: low (WBC ≤10×10/L and platelet count >40×10/L), intermediate (WBC ≤10×10/L and platelet count ≤40×10/L), and high (WBC >10×10/L).4 High-risk APL patients with a WBC count greater than 10×10/L were reported to achieve higher complete remission rates and better survival outcomes when cytarabine was included in the chemotherapy regimens, whereas for patients with a WBC count less than 10×10/L all-trans retinoic acid in combination with anthra-cyclines might be sufficient.6–7 Lengfelder et al. observed no significant differences in survival outcomes and relapse incidence in 142 APL patients (who all received cytarabine within their induction and consolidation protocols) when they were separated according to a WBC threshold of 10×10/L.5

The most suitable parameters for risk stratification in APL are, therefore, still under debate. It was discussed whether patients with the French-American-British (FAB) subtype M3v might have higher rates of early death because of hemorrhagic complications when compared to patients with the classical FAB M3 morphology,8–9 but Tallman et al. found that outcomes of patients with the two FAB subtypes did not differ significantly when adjustment for WBC counts or relapse risk scores was made.10 In fact, the FAB M3v subtype has been associated with higher frequencies of FLT3-internal tandem duplications (ITD), which may have a negative prognostic impact.11 FLT3-ITD occur in 12–38% of all APL patients and tyrosine kinase domain (TKD) mutations in 2–20%.12 The presence of an FLT3-ITD was reported to worsen prognosis in APL and to be associated with higher WBC counts by several study groups,11–12 but others found no adverse prognostic impact of this molecular marker in APL.13–14 In fact, there were too few patients in many studies in order to be able to draw final conclusions, and it remains unclear whether FLT3-ITD mutation status should be incorporated into risk-adapted therapeutic algorithms for APL patients.12,15 Other parameters, such as FLT3-ITD mutation level or length, and PML-RARA expression level have been described to be of prognostic relevance in APL.16

In this study, we investigated the impact of different pre-treatment parameters and the influence of additional cytogenetic or molecular genetic parameters which may predict outcome in 147 adult patients with newly diagnosed APL.

Design and Methods

Patients

The study was based on 147 patients with APL at diagnosis. There were 85 males and 62 females (median age, 53.9 years; range, 19.7 – 86.3 years). One hundred and thirty-six patients had de novo APL and 11 had therapy-related disease (t-APL). Bone marrow and/or peripheral blood samples were sent from different hematologic centers between 8/2005–07/2010 to the MLL Munich Leukemia Laboratory for routine diagnostic purposes. Patients were selected according to availability of cytogenetic data and parallel information on molecular genetics including PML-RARA and FLT3-ITD and FLT3-TKD mutation status. All patients received all-trans retinoic acid in combination with intensive chemotherapy according to standard study protocols.1,5,17 This cohort is completely independent of a previously published one.11 All patients gave informed consent to the use of laboratory data for research studies. The study was approved by the Internal Review Board and adhered to the Declaration of Helsinki. Details and further characterization of the cohort are shown in Online Supplementary Table S1.

Cytomorphology

Bone marrow/peripheral blood smears were available in 115 cases and were stained using the May-Grunwald Giemsa method. Cytochemistry was performed for myeloperoxidase and non-specific esterase.18 Cases were classified as M3/hypergranular type or M3v/microgranular variant according to the FAB19 and World Health Organization (WHO)15 classifications.

Cytogenetics

Chromosome banding analysis and fluorescence in situ hybridization (FISH) were performed in all 147 cases according to standard techniques.20

Molecular genetics

Following extraction of mRNA and cDNA synthesis, the different PML-RARA fusion transcripts were detected by reverse transcription polymerase chain reaction (RT-PCR) analysis.11 PML-RARA expression was quantified based on real-time PCR. Expression of PML-RARA was given as a ratio defined as %PML-RARA/ABL1.21 Fragment analysis was used to screen and quantify the FLT3-ITD mutation load, and determine the length of the ITD (GeneScan, 3130 sequence detection system, ABI, Darmstadt, Germany).22 The FLT3-ITD load was quantified as the ratio of the mutated allele to the wild-type allele (FLT3-ITD/wt ratio). Ratios of 1 or more were indicative of complete or partial loss of the wild-type allele (FLT3wt) in at least some of the cells. The FLT3-TKD were analyzed by LightCycler® (ROCHE, Mannheim, Germany) based melting curve analysis.23

Immunophenotyping

Immunophenotyping by multiparameter flow cytometry was done by 5-fold staining in a subgroup of 43 cases (Beckman Coulter, Krefeld, Germany).24 For subgroup analysis cases were also investigated for CD56 and classified as CD56-positive when at least 20% of the leukemic cell population was positive by comparison with the isotype.

Statistical analysis

Overall survival and event-free survival were calculated according to the Kaplan-Meier method and compared by two-sided log rank tests. Cox regression analysis was performed for survival outcomes with different parameters as covariates. Parameters which were significant in univariate analysis were included into multivariate analysis. Dichotomous variables were compared between different groups using the χ-test and continuous variables by Student’s T-test. Spearman’s rank correlation was used to analyze correlations between continuous parameters. All P values reported are two-sided. SPSS (version 14.0.1, Chicago, IL, USA) was used.

Results

Frequency of FLT3 mutations

FLT3 mutations were detected in 65/147 (44.2%) cases: FLT3-ITD in 47/147 (32.0%) and FLT3-TKD in 19/147 (12.9%). As one case had an FLT3-ITD and an FLT3-TKD mutation in parallel, patients with mutated FLT3 status (either ITD or TKD) were called the "FLT3-mutated cohort". Patients without any FLT3 mutation (neither ITD nor TKD) were defined as the "FLT3-wt cohort”.

FLT3-internal duplication mutation load

In the 47 FLT3-ITD-positive APL patients the mean FLT3-ITD/wt ratio was 0.51 (range, 0.02 – 1.06; median, 0.54). The mean FLT3-ITD/wt ratio was significantly lower in APL patients than in a previous cohort of 197 FLT3-ITD-positive patients with normal karyotype AML25 (0.51 versus 1.98, respectively; P<0.001). The median length of the ITD was 46 bp (range, 15–105 bp) which was in the same range as that in the normal karyotype cohort.

Biological characteristics

Sex ratio, mean age, and history of APL (de novo APL or t-APL) did not differ significantly between the different subgroups defined by FLT3-ITD or -TKD mutation status (Online Supplementary Table S1). The age of patients with t-APL or de novo APL did not different significantly (mean age, 51.5 versus 54.1 years, respectively). In contrast, the FLT3-mutated cohort had higher mean WBC counts when compared to FLT3-negative patients (26.8 versus 4.7×10/L, respectively; P<0.001). In more detail, FLT3-ITD-positive patients had higher mean WBC counts than either FLT3-ITD-negative patients (P<0.001) or FLT3-wt patients (P<0.001). In contrast, there were no significant differences in mean WBC counts between FLT3-TKD-mutated patients, FLT3-TKD negative patients and FLT3-wt cases. The mean platelet count was lower in FLT3-mutated patients than in FLT3-negative patients (30 versus 71×10/L, respectively; P<0.001), without significant differences between FLT3-ITD and FLT3-TKD mutated patients. Mean hemoglobin levels were lower in FLT3-ITD and FLT3-TKD patients than in the FLT3-wt cohort (Online Supplementary Table S1).

Morphological characterization

According to the FAB classification, 68 cases were classified as M3 (59.1%) and 47 as M3v (40.9%) (FAB subtypes were known for 115 cases). The proportion of M3v subtype was higher in the FLT3-ITD positive cohort than in the FLT3-ITD negative cases (25/34; 73.5% versus 22/81; 27.2%; P<0.001). Moreover, the M3v subtype was more frequent in FLT3-ITD-positive cases than in FLT3-TKD-positive cases (25/34; 73.5% versus 5/16; 31.3%; P=0.004) (Online Supplementary Table S1). The AML M3v patients had significantly higher mean WBC counts when compared to the M3 patients (34.5×10/L versus 4.5×10; P<0.001).

PML-RARA transcript types

The most frequent PML-RARA transcript type was bcr1 which was detected in 89 cases (60.5%), followed by bcr3 in 52 cases (35.4%). The bcr2 type was detected in 6 cases (4.1%) only. The distribution of the different bcr transcript types for FLT3 mutations was heterogeneous (P<0.001): the bcr1 transcript type was correlated with FLT3-wt status (63/82; 76.8% in FLT3-wt versus 26/65; 40.0% in FLT3-mutated patients), while bcr3 was more frequent than bcr1 or bcr2 in FLT3-mutated patients than in FLT3-wt patients (38/65; 58.5% versus 14/82; 17.1%; P<0.001). FLT3-ITD-positive patients more frequently showed bcr3 breakpoints than did FLT3-TKD-positive patients (32/47; 68.1% versus 6/19; 31.6%; P=0.011) (Online Supplementary Table S1).

PML-RARA expression

The mean ratio of %PML-RARA/ABL1 expression was 25.8; the median was 18.5 and there was a wide range of 0.6 – 96.7. No significant differences of mean PML-RARA expression levels were observed between the different molecular subgroups defined by the different FLT3 mutations (Online Supplementary Table S1). The mean PML-RARA/ABL1 expression was higher in patients with WBC counts above 1.0×10/L than in those with WBC counts of 1.0×10/L or less (33.6 versus 23.4; P=0.031). When peripheral WBC counts and PML-RARA/ABL1 were considered as continuous parameters by Spearman’s analysis, they were significantly correlated (P=0.042). Mean PML-RARA/ABL1 expression was lower in cases with bcr1 breakpoints than in patients with bcr2 and bcr3 combined (23.2 versus 30.3; P=0.059).

Additional chromosomal alterations

Additional chromosomal abnormalities (ACA) were detected in 57 patients (38.8% of the total cohort) without significant differences between the FLT3-mutated and the FLT3-wt patients. There was a trend to a higher rate of ACA in the FLT3-TKD-positive cohort than in the FLT3-TKD-negative patients (11/19; 57.9% versus 46/128; 35.9%; P=0.080), and FLT3-TKD-positive patients had ACA more frequently than had FLT3-ITD-positive patients (11/19; 57.9% versus 15/47; 31.9%; P=0.05).

Recurrent ACA (+8, 9q-, 17q-alterations) were found in 41 cases (27.9% of the total cohort; 71.9% of all ACA). The most frequent abnormality was trisomy 8 or gain of 8q (n=24); followed by 17q alterations (n=11) and 9q deletions (n=6). Infrequent and non-recurrent ACA were detected in 16 cases (10.9% of the total cohort; 28.1% of patients with ACA) (Online Supplementary Table S1 and Table 1). In detail these abnormalities were additional translocations (n=7), insertions or deletions (n=4), a complex karyotype (n=1), other alterations (n=1) and –Y (n=3).

Chromosomal gains and losses due to ACA are depicted in Online Supplementary Figure S1A and breakpoints from ACA in Online Supplementary Figure S1B according to CyDAS.

Based on a report by Slack et al., who described a significant association of the presence of ACA and the PML-RARA S isoform,26 we compared the frequencies of the different bcr transcript types depending on the presence of ACA, but did not detect any significant correlations (data not shown).

Survival analysis

The presence of ACA had no significant impact on survival outcomes when compared to the outcomes of patients with a sole t(15;17)/PML-RARA (Table 2).

The FLT3-ITD (Figure 1A,B) or FLT3-TKD mutation status (positive versus negative) had no significant impact on survival outcomes. The median overall and event-free survival of patients with FLT3-ITD, FLT3-TKD, and FLT3-wt also did not differ significantly (Figure 1C,D; Table 2). In contrast, when taking the mutation load expressed as FLT3-ITD/wt ratio into account, patients with a FLT3-ITD/wt ratio less than 0.5 (meaning FLT3-ITD-positive patients with a ratio <0.5 and FLT3-ITD-negative patients combined) showed better 2-year overall survival (86.7% versus 72.7%; P=0.075) (Figure 1E) and event-free survival rates (84.5% versus 62.1%; P=0.023) (Figure 1F) than those with a ratio of 0.5 or more. Results for 0.25 and 0.75 thresholds are shown in Table 2, demonstrating that the 0.25 threshold had no effect on survival. Thus, only an FLT3-ITD load of 0.5 or more had an adverse impact on survival in patients with PML-RARA positive APL.

Subsequently, we defined the influence of FLT3 mutations on induction death, i.e. within the first 30 days following the start of therapy. The 30-day overall survival rate of FLT3-ITD mutation carriers and of patients with a negative mutation status did not differ significantly (85.7% versus 91.1%; Figure 2A). The FLT3-TKD mutation also had no significant impact on the 30-day overall survival rate (89.5% versus 88.9%; P=n.s.). Likewise, the presence of any of the FLT3 mutation types did not significantly affect the 30-day overall survival rate when compared to that in FLT3 mutation-negative patients (91.8% versus 86.4%; P=n.s.). However, using a threshold of the mutation/wild-type of 0.5, the 30-day overall survival rate was significantly better for those with a FLT3-ITD/wt ratio less than 0.5 compared to those with a ratio of 0.5 or above (91.7% versus 78.3%; P=0.039; Figure 2B).

Based on previous observations that expression of CD56 was correlated with higher relapse risk in APL,27–28 we evaluated this parameter in 43 patients with available immunophenotypic data. CD56 expression greater than 20% was seen in 6/43 cases (14.0%) which was comparable to the frequency in the mentioned studies. In this small subcohort the 2-year overall and event-free survival rates did not differ significantly between patients with and without CD56 expression (75.0% versus 90.0% and 75.0 versus 90.9%, P=n.s. respectively; Table 2).

Survival data were available for 132/147 cases of the total cohort. The 2-year overall survival rate was 84.3% and the median follow-up was 767 days. Male patients had better 2-year overall survival rates compared to female patients (92.7% versus 78.3%; P=0.040) (Figure 3A, Table 2). No significant difference was found regarding 2-year event-free survival (Figure 3B, Table 2). M3 versus M3v FAB subtypes, history of APL (de novo APL versus t-APL), and the different PML-RARA fusion transcript types (bcr1–3) had no significant impact on survival outcomes.

We further separated patients according to the Sanz score,4 based on WBC and platelet counts. Patients with WBC counts of 10×10/L or below had a better 2-year overall survival rate than those with WBC counts above 10×10/L (88.3% versus 69.4%; P=0.015) (Figure 4A) and a better 2-year event-free survival rate (85.7% versus 60.0%; P=0.006) (Figure 4B). Patients with a platelet count above 40×10/L had a better 2-year overall survival rate than patients with a platelet count of 40×10/L or below (92.7% versus 78.1%; P=0.060) (Figure 4C), and a better 2-year event-free survival rate (84.0% versus 76.5%; P=n.s.) (Figure 4D).

Table 1.Karyotypes of 13 cases from the subgroup of infrequent additional chromosomal abnormalities (the three cases with -Y as sole abnormality are not shown).

Table 2.Kaplan-Meier estimates for survival outcomes according to different biological, morphological, cytogenetic, and molecular genetic parameters.

The following parameters were tested in univariate analysis with respect to their influence on survival outcomes: gender, WBC count (threshold: 10×10/L) and platelet count (threshold: 40×10/L) - both limits set according to the Sanz score),4 hemoglobin level, age as a continuous variable, de novo APL versus t-APL, FAB M3 versus M3v subtype, FLT3-ITD mutant status, FLT3/ITD/wt ratios of 0.25 or more and of 0.5 or more, FLT3-TKD mutant status and the presence of ACA. A negative influence on overall survival was documented for female sex (P=0.051), higher age (P=0.001), and FLT3-ITD/wt ratio of 0.5 or more (P=0.084). WBC counts greater than 10×10/L (P=0.021) and platelet counts less than 40×10/L (according to the Sanz score; P=0.074) were associated with worse overall survival. In contrast, the other parameters listed above (including FLT3-ITD and FLT3-TKD mutant status or FLT3/ITD/wt ratio with a threshold of ≥0.25) had no significant impact on overall survival. In multivariate analysis for overall survival, significance was reached only for age as a continuous parameter (P<0.001). There was borderline significance for gender (P=0.055) and WBC counts greater than 10×10/L (P=0.059).

For event-free survival, significant parameters in univariate analysis were WBC counts greater than 10×10/L (P=0.009), age (P<0.001), and FLT3-ITD/wt ratio of 0.5 or more (P=0.029). In multivariate analysis for event-free survival, only age as a continuous parameter (P<0.001) was statistically significant (Table 3). We further divided patients into subgroups under 60 years old and those 60 years or more and found the same correlations as in the combined group with respect to age (data not shown).

Discussion

In current study protocols patients with APL are assigned to different therapeutic regimens (e.g. with regard to application of cytarabine) according to peripheral WBC counts,6–7 but it remains under debate whether other parameters should be included in risk stratification in patients with this subtype of acute myeloid leukemia. We, therefore, investigated the clinical impact of FLT3-ITD and FLT3-TKD, and other parameters in 147 patients with APL at diagnosis. First, we were able to confirm the high frequency of FLT3 mutations in APL,12,29 as 44.2% of patients had either an ITD or TKD or both (one case). The presence of FLT3-ITD was associated with specific characteristics, i.e. higher WBC counts, lower platelet counts, a preponderance of the M3v subtype, and of the bcr3 PML-RARA fusion transcript (P<0.001 for all parameters) when compared to ITD-negative patients, confirming findings of other study groups.29–30 FLT3-TKD mutated patients differed from FLT3-TKD negative cases by having lower platelet counts (P=0.006) and a higher frequency of ACA (P=0.080).

Subsequently, we evaluated the prognostic impact of FLT3-ITD in our cohort of patients with APL. The presence of FLT3-ITD per se had no significant impact on survival outcomes, but a FLT3-ITD/wt ratio of 0.5 or more was prognostically adverse in univariate analysis (event-free survival, P=0.029; overall survival, P=0.084) compared to a low FLT3-ITD burden below 0.5% or to wild-type FLT3 status. A high FLT3-ITD burden is, therefore, clinically relevant in patients with APL, and may contribute to explaining the differences in survival in patients with t(15;17)/PML-RARA. We were further able to confirm the 0.5 threshold for the FLT3-ITD/wt ratio to be prognostically relevant specifically in the induction period within the first 30 days from the start of therapy. Consistent with our results, Chillon et al. described an increasing FLT3-ITD/wt ratio of greater than 0.66 to be related with shorter 5-year relapse-free survival in patients with APL, whereas overall survival was only weakly influenced by FLT3-ITD mutation status.16 Most other studies focused on the FLT3-ITD mutation status in APL only and did not investigate the clinical impact of a certain FLT3-ITD mutation load. Gale et al. found higher rates of induction death in patients with mutant FLT3, but no significant adverse effect of FLT3-ITD mutated status on overall survival or relapse risk of patients with APL.29 This was similar to the findings of Au et al. who described a clearly adverse impact of the FLT3-ITD on the achievement of remission (P=0.06), but failed to demonstrate a significant impact of the FLT3-ITD mutation status on disease-free survival in APL patients.31 Noguera et al. reported a non-significant trend for worse disease-free survival or relapse risk in patients with FLT3-ITD-positive APL.30 In accordance with Chillon et al.16 and previous data from our group11 we found no significant impact of the FLT3-TKD mutant status in our APL cohort. In contrast, Gale et al. described a worse overall survival, of borderline statistical significance, in APL patients with FLT3-TKD (P=0.05).29

Figure 1.(A-B) Overall survival and event-free survival of FLT3-ITD-positive patients versus FLT3-ITD-negative patients. (C-D) Overall survival and event-free survival of patients with FLT3-ITD, FLT3-TKD, and FLT3 wild-type. (E-F) Overall survival and event-free survival of patients with a FLT3-ITD/wt load ≥0.5 versus those with an ITD load <0.5.

Figure 2.(A) Overall survival of the APL patients within the first 30 days after starting therapy comparing those with FLT3-ITD and without. (B) Comparison of patients with FLT3-ITD ratio ≥0.5 and below this threshold.

Figure 3.(A) Overall survival and (B) event-free survival of male versus female patients.

Figure 4.(A) Overall survival and (B) event-free survival according to WBC count thresholds following the Sanz score4 (≤10x10⁹/L versus >10x10⁹/L). (C) Overall survival and (D) (EFS) according to platelet count thresholds following the Sanz score4 (>40x10⁹/L versus ≤40x10⁹/L)

The mean FLT3-ITD/wt ratio was significantly lower (P<0.001) in APL patients than in a cohort of 197 patients with FLT3-ITD-positive normal karyotype acute myeloid leukemia25 which we had previously analyzed. It remains speculative whether this aspect could contribute to explain the weaker prognostic impact of the FLT3-ITD in patients with APL than in those with normal karyotype acute myeloid leukemia.

The frequency of ACA in our cohort (38.8%) was similar to that in previous studies.32–33 In general we did not find that ACA had a significant impact on prognosis, which is in accordance with our previous study in 50 patients with APL.34 Cervera et al. from the PETHEMA study group found that ACA were associated with higher rates of coagulopathy, lower platelet counts, and higher relapse risk scores in APL, but remission rates of patients with and without ACA were nearly the same, being 90% and 91%, respectively. In their study, no specific ACA was an independent risk factor for relapse.33 The European APL group also did not find that ACA had a significant impact on prognosis,32 and Slack et al. found no difference in overall survival between APL patients with an isolated t(15;17)/PML-RARA and patients with ACA.26 Accordingly, we were not able to show a significant influence of ACA on outcome in patients with APL.

Finally, we were able to confirm the prognostic power of a WBC count threshold of 10×10/L, as introduced by Sanz et al., to separate different risk groups.4 Other than that, male gender showed a borderline significance for better overall survival in univariate and multivariate analyses, and age as a continuous parameter was a strong independent prognostic parameter (P<0.001 for overall and event-free survival in multivariate analysis) in our study. Survival outcomes of patients with de novo APL or t-APL did not differ significantly in our study, which was in accordance with the results of a previous study on 106 patients with t-APL by Beaumont et al.35 Nevertheless, given that for all subtypes of AML, clinical outcomes of patients with therapy-related disease were found to be worse than those of patients with de novo acute myeloid leukemia, as recently described in a large cohort including 200 patients with therapy-related acute myeloid leukemia and 2653 with de novo acute myeloid leukemia by the AMLSG Study Group,36 the potential clinical decisions in t-APL should be further studied before definite conclusions are drawn.

In conclusion, we were not able to show a significant impact of the FLT3-ITD mutation status per se on prognosis in APL, but a higher FLT3-ITD/wt ratio (≥0.5%) was prognostically adverse. Prospective trials should further investigate the clinical impact of the FLT3-ITD/wt mutation load aiming to evaluate whether this parameter might be included in risk stratification in APL.

Table 3.Results of univariate and multivariate analyses according to Cox regression analysis (* thresholds according to the Sanz score4).

Footnotes

• The online version of this article has a Supplementary Appendix.
• Authorship and Disclosures The information provided by the authors about contributions from persons listed as authors and in acknowledgments is available with the full text of this paper at www.haematologica.org.
• Financial and other disclosures provided by the authors using the ICMJE (www.icmje.org) Uniform Format for Disclosure of Competing Interests are also available at www.haematologica.org.

• Received June 17, 2011.
• Revision received August 11, 2011.
• Accepted August 16, 2011.