FLT-3 aberrations in acute promyelocytic leukaemia: clinicopathological associations and prognostic impact


Dr Y. L. Kwong, University Department of Medicine, Professorial Block, Queen Mary Hospital, Pokfulam Road, Hong Kong, China. E-mail: ylkwong@hkucc.hku.hk


FLT-3 aberrations that occur as an internal tandem duplication (ITD) or a mutation at the activation-loop position 835, D835, are common in acute promyelocytic leukaemia (APL). We investigated the clinicopathological associations and prognostic impact of FLT-3 aberrations in a cohort of APL patients. FLT-3 exons 11 and 12 were amplified by polymerase chain reaction (PCR), and the ITD was recognized as an increase in the size of the PCR product. FLT-3 exon  17 was amplified, and D835 mutation was identified by loss of an EcoRV site, followed by DNA sequencing. Of 82 patients studied, FLT-3 aberrations were detected in 35 cases (43%) at diagnosis (ITD: 16; D835 mutation: 18; ITD + D835 mutation: 1). FLT-3 ITD, but not D835 mutations, was significantly associated with higher presentation white blood cell count (WBC) and microgranular morphology. Early/induction deaths were related to male sex and high presentation WBC. There was a trend for FLT-3 ITD to be associated with non-remission (P = 0·06). For disease-free survival, high WBC was the only significant adverse factor. Male sex, high WBC and FLT-3 ITD were significant adverse factors for overall survival. These findings have important implications on the possible use of FLT-3 inhibitors in the treatment of APL.

Acute promyelocytic leukaemia (APL) is characterized by t(15;17)(q22;q12) and the PML/RARα gene fusion. It is unique in its response to all trans-retinoic acid (ATRA) and arsenic trioxide (As2O3) (Wang, 2003). About 5–10% of patients die from haemorrhagic complications at initial presentation (Wang, 2003), and up to 30% of patients will eventually relapse (Tallman et al, 2002; Gilliland, 2003; Wang, 2003). Proposed adverse prognostic indicators include high presentation white blood cell count (WBC), older age (Pantic et al, 2000; Asou et al, 2001; Tallman et al, 2002), male sex (Tallman et al, 2002), PML/RARα fusion at PML BCR3 (Jurcic et al, 2001), the microgranular variant (Tallman et al, 2002), persistent minimal residual leukaemia (Grimwade et al, 1996; Jurcic et al, 2001), additional cytogenetic alterations (Hiorns et al, 1997), expression of the surface antigens CD34, CD56 and the multidrug resistant (MDR) gene (Murray et al, 1999; Candoni et al, 2003; Lee et al, 2003), and aberrant p15 promoter hypermethylation (Chim et al, 2003).

Recently, FLT-3 aberrations, in the form of an internal tandem duplication (ITD) or mutation at the activation loop position 835, D835, have been reported as the commonest genetic abnormality in acute myeloid leukaemia (AML) (Thiede et al, 2002). Two studies have shown FLT-3 aberrations in 20–45% of APL cases (Thiede et al, 2002; Shih et al, 2003). However, both reports failed to implicate FLT-3 aberrations as prognostic indicators.

The treatment of APL has evolved rapidly. Furthermore, the use of As2O3 has dramatically improved the treatment results of relapsed APL (Au et al, 2003a). Therefore, prognostic factors should be assessed in relationship to treatment modalities (Hu et al, 1999). In this study, we investigated the interplay of various clinicopathological parameters and FLT-3 aberrations in the prognostication of a consecutive cohort of Chinese APL patients treated with ATRA and As2O3.

Material and methods


All consecutive cases of APL diagnosed between 1995 and 2002 were studied. The diagnosis was confirmed with cytogenetic analysis, fluorescence in situ hybridization (Ma et al, 2000) and reverse transcription polymerase chain reaction (RT-PCR) as reported (Kwong et al, 1995). The short and long PML/RARα fusion products (PML break at BCR3 and BCR1 respectively) were determined from the size of RT-PCR product.


All patients received an initial induction treatment with ATRA (45 mg/m2/d) until remission (Table I). This was supplemented with daunorubicin 50 mg/m2/d ×3 and cytosine arabinoside 100 mg/m2/d ×7, which might be given up-front in patients with high presentation WBC and symptoms of the ATRA syndrome (Chim et al, 1996). Postremission consolidation comprised two monthly courses of daunorubicin 50 mg/m2/d ×2 and cytosine arabinoside 100 mg/m2/d ×5. Maintenance treatment with methotrexate (20 mg weekly), mercaptopurine (100 mg/d) and ATRA (45 mg/m2/d) was given for 18 months in all cases, except those who refused or suffered side effects. Before 1997, relapses were treated with ATRA, cytosine arabinoside and either daunorubicin or mitoxantrone. After 1997, relapses were treated with As2O3 (10 mg/d i.v. or p.o.) until remission. This was supplemented with idarubicin if leucocytosis >10 × 109/l occurred during As2O3 treatment (Kwong et al, 2001). Remitting patients received idarubicin consolidation, and maintenance treatment with oral As2O3 ± ATRA as reported (Au et al, 2002, 2003b).

Table I.  Clinicopathological features and treatment results of 82 cases of acute promyelocytic leukaemia.
Characteristics and treatment outcomeNumber
  1. ATRA, all trans-retinoic acid; As2O3, arsenic trioxide; CR, complete remission.

Patient demographics
 Median age, years (range)39 (7–87)
 Median white cell count (× 109/l)4·4 (0·4–184)
Induction therapy with ATRA + chemotherapy
 Died at presentation7
 Induction death12
 First Relapse (median time in months, range)39 (15, 6–67)
 CR1 (median follow-up in months, range)24 (60, 29–100)

Molecular analysis

Two regions of the FLT-3 gene were amplified for detection of ITD and D835 mutation as reported (Au et al, 2004). A 329 base pair (bp) fragment between exons 11 and 12 was amplified, and ITD was detected as an increase in the size of the PCR product. Another 114 bp fragment at exon  17 was amplified, followed by restriction enzyme digestion with EcoRV. The PCR product from the wild-type allele would be cleaved into 68 and 46 bp fragments, whereas that from the D835 mutant would have a loss of the EcoRV site. Direct DNA sequencing was then used to confirm the D835 mutation.

Statistical analysis

Disease-free survival (DFS) was measured from complete remission (CR) to first relapse. Overall survival (OS) was measured from diagnosis to death or last follow-up. Survivals were analysed with the Kaplan–Meier method, and univariate analysis for risk factors (WBC, age >50 years, FLT-3 aberrations, sex, microgranular morphology, PML breakpoint and additional cytogenetic changes) was performed by log-rank analysis. Factors with P < 0·1 on univariate analysis were further analysed by logistic regression (SPSS, Chicago, IL, USA). P < 0·05 was regarded as statistically significant.



Eighty-two patients (men 39, women 43) at a median age of 39 (7–87) years were studied (Table I). A total of 19 patients died either before or during induction treatment, mainly because of bleeding complications (Table I). Sixty-three patients achieved CR. Relapses occurred in 39 of 63 cases at a median of 15 months into the first remission (CR1). Eight patients were treated with ATRA/chemotherapy, of whom six died of leukaemia or therapy-related causes. Thirty-one patients were treated with As2O3, and all achieved a second CR (CR2). Three patients died from subsequent post-As2O3 relapses, while three patients died of unrelated causes. At the latest follow-up, a total of 51 patients were still alive and in remission, at a median follow-up of 69 months. For the 63 patients who achieved CR1, the 3-year DFS and OS were 42% and 83% respectively. The difference between DFS and OS was due to the high efficacy of As2O3-based salvage strategies (Au et al, 2002).

FLT-3 aberrations

FLT-3 aberrations were detected in 35 cases (43%) at diagnosis (ITD: 16; D835 mutation: 18; ITD + D835 mutation: 1) (Figs 1 and 2). The amino acid changes were Asp835 → Tyr (n = 11), Asp835 → Val (n = 4), Asp835 → His (n = 3) and Asp835 → Ala (n = 1). FLT-3 aberrations were significantly associated with higher median WBC and microgranular morphology, but not significantly related to sex, age, PML breakpoint or additional cytogenetic changes (Table II). Further analysis of the type of FLT-3 aberration showed that high WBC and microgranular morphology were related to ITD but not D835 mutations (Table II).

Figure 1.

FLT-3 aberrations in patients with acute promyelocytic leukaemia. Polymerase chain reaction (PCR) for internal tandem duplication (ITD). Amplification of a fragment between exons 11 and 12 gives a 329 base pair (bp) product. The ITD is shown by a larger PCR product (upper panel). An ITD was found in patient 1 at presentation (P1P) and at first relapse (P1R1), but was absent at the second and third relapse (P1R2 and P1R3). No ITD was found in patient 2 at presentation (P2P) or relapse (P2R1). In patient 3, ITD was present at presentation (P3P), was lost at first relapse (P3R1), but reappeared at second relapse (P3R2). The lower panel shows PCR for the D835 mutation. Amplification of a 114 bp fragment at exon  17 was followed by restriction enzyme digestion with EcoRV. Successful digestion of the PCR product by EcoRV indicated absence of the D835 mutation.

Figure 2.

Changes in the FLT-3 D835 mutation at presentation and relapse in a patient with acute promyelocytic leukaemia. (A) At presentation, the D835 mutation was GAT to TAT (Asp → Tyr). (B) At relapse, the D835 mutation became GAT to GTT (Asp → Val).

Table II.  Clinicopathological associations of FLT-3 aberrations.
Clinicopathological featuresFLT-3 aberrationP-valueType of FLT-3 aberrationP-value
Present (n = 35)Absent (n = 47)ITD (n = 16)*D835 (n = 18)*
  1. * One case had concomitant ITD and D835 was excluded from subgroup analysis.

  2. † 16 cases had no growth/normal metaphases.

  3. ITD, internal tandem duplication; NS, not significant.

Median presentation white cell count (× 09/l)38·1 8·5<0·00158·524·1<0·001
Blast morphology
 Classical (n = 73)2647 917 
 Microgranular variant (n = 9)90<0·001710·014
 Male (n = 39)2019 911 
 Female (n = 43)15280·1877NS (1·00)
 >50 years (n = 22)913 45 
 <50 years (n = 60)26340·821312NS (0·89)
 Long isoform (n = 49)1831 512 
 Short isoform (n = 33)17160·15107NS (0·17)
Additional cytogenetic abnormalities†
 Present (n = 15)69 15 
 Absent (n = 51)23280·57139NS (0·16)

Prognostic factors impacting on early/induction deaths

Early/induction deaths were significantly related to male sex (P = 0·003) and high presentation WBC (P < 0·001), but not associated with ITD (P = 0·06) (Table III).

Table III.  Prognostic factors for early/induction death in 82 cases of acute promyelocytic leukaemia.
ParametersInductionOdds ratio95% CIP-value*
  1. CR, complete remission; ITD, internal tandem duplication.

  2. * Two-tailed P-values.

  3. † One case, with concomitant ITD and D835, was excluded from this analysis.

 Male (n = 39)2415   
 Female (n = 43)3944·131·50–11·40·003
Age (years)
 >50 (n = 22)148   
 <50 (n = 60)49111·500·78–2·890·27
Presentation white cell count (×109/l)
 >10 (n = 31)1912   
 <10 (n = 51)4473·711·87–7·40<0·001
Blast morphology
 Microgranular (n = 9)63   
 Classical (n = 73)57161·660·46–6·000·42
Additional cytogenetic abnormalities
 Present (n = 15)123   
 Absent (n = 51)36150·660·21–2·090·74
PML/RARα isoform
 Short BCR3 (n = 33)258   
 Long BCR1 (n = 49)38111·060·55–1·951·00
FLT-3 aberration
 Present (n = 35)2411   
 Absent (n = 47)3981·520·93–2·500·19
Type of FLT-3 aberration†
 ITD (n = 16)97   
 D835 (n = 28)1441·620·83–3·200·27
Presence of ITD
 Present (n = 17)107   
 Absent (n = 65)53122·321·02–5·260·06

Prognostic factors impacting on relapse and survival

For DFS, both univariate and multivariate analyses showed that high presentation WBC was the only significant adverse factor (P = 0·0064 and P = 0·026 respectively) (Table IV). The presence of ITD was not significantly related to DFS on univariate and multivariate analysis (P = 0·052 and P = 0·264 respectively). For 3-year OS, on univariate analysis, male sex, age and high presentation WBC were adverse factors. FLT-3 aberrations were also associated with an inferior OS, which was due to the presence of ITD instead of D835 mutations. On multivariate analysis, the prognostic significance of FLT-3 aberrations (P = 0·77) and sex (P = 0·06) were lost, while age (P = 0·01) and high WBC (P = 0·05) remained significant.

Table IV.  Analysis of prognostic factors for disease-free survival (DFS) and overall survival (OS) in 82 cases of acute promyelocytic leukaemia.
ParametersDFS (month)95% CIP-value*3-year OS (%)†RR95% CIP-value*
  1. CI, confidence interval; RR, relative rates of survival at three years with 95% CI; ITD, internal tandem duplication.

  2. P-value obtained by univariate analysis.

  3. † 3-year OS with standard errors (5-year OS not reached yet).

  4. ‡ Differences due to one case with concurrent ITD and D835 aberrations being excluded from comparison between ITD and D835.

 Male13·71·2–26·2 56 ± 8   
 Female20·312·6–28·00·2973 ± 72·20·89–5·70·044
Age (years)
 >5015·17·5–22·7 50 ± 10   
 <5017·211·4–27·40·9971 ± 62·50·92–6·90·042
Presentation white cell count (×109/l)
 >1013·75·4–21·9 51 ± 9   
 <1024·613·9–35·30·006475 ± 62·91·1–7·50·023
Blast morphology
 Microgranular 7·60·9–14·3 80 ± 17   
 Classical17·914·4–24·40·5154 ± 81·60·4–6·60·15
Additional cytogenetic aberrations
 Present24·615·9–34·3 80 ± 10   
 Absent15·810·3–23·30·5154 ± 30·40·1–1·30·15
PML/RARα isoform
 Short BCR313·910·3–17·7 67 ± 9   
 Long BCR119·410·8–29·80·1362 ± 82·40·9–6·20·83
FLT-3 aberration
 Present13·99·4–18·4 54 ± 8   
 Absent20·713·3–27·80·273 ± 72·41·0–6·20·049
Type of FLT-3
 ITD 5·3‡0–18·8 31 ± 12‡   
 D83514·18·1–20·10·1772 ± 75·71·3–25·10·009
Presence of ITD
 Present 7·6‡0·0–22·9 35 ± 12‡   
 Absent20·38·1–20·10·05273 ± 65·21·7–16·20·006

Serial changes of FLT-3 aberrations

In 20 of 39 relapsing cases, 29 serial APL marrow samples (from diagnosis to relapses) were studied. There was no change in the FLT-3 genetic profile in 13 patients. However, a change in status was found in seven cases (35%), including loss of ITD (n = 3, one after chemotherapy, two after As2O3 therapy), loss of the D835 mutation (n = 2, both after chemotherapy) and change in D835 mutation (n = 1, after chemotherapy) (Fig 2). One case presented with an ITD at initial diagnosis. At first relapse (R1) after chemotherapy, there was a transient loss of the ITD. The R1 was successfully treated with As2O3. A further relapse (R2) occurred later, with re-emergence of the ITD (Fig 1). None of the cases with normal FLT-3 at presentation acquired FLT-3 aberrations at relapse.


Despite recent advances in the treatment of APL, early mortality and relapses still occur. Therefore, prognostic markers may be useful in identifying high-risk patients for closer monitoring and possibly more prolonged therapy and maintenance treatment.

Unfortunately, most studies to date have failed to show a prognostic significance for FLT-3 aberrations. A small German study of 21 APL patients showed an FLT-3 ITD incidence of 38%, which did not have a significant influence on OS, CR or DFS (Kainz et al, 2002). A Japanese study showed an ITD incidence of 20·3% in 17 APL patients, which was related to high WBC but not survival (Kiyoi et al, 1997). These studies involved only small numbers of patients, so that the results should be interpreted with caution. Two recent studies have investigated both ITD and D835 mutations (Noguera et al, 2002; Shih et al, 2003). In 107 consecutive Chinese patients, the frequency of ITD and D835 mutations were 20·6% and 18·7% respectively. FLT-3 ITD, but not D835 mutations, was associated with PML breakpoint at BCR3 and microgranular morphology. The outcome (CR, OS and DFS) was unrelated to FLT-3 aberrations (Shih et al, 2003). However, the incidence of early death and the salvage protocol for relapses was unclear. In 90 Italian APL patients treated with the AIDA (ATRA and idarubicin) protocol, the incidence of ITD and D835 mutations was 36·7% and 7·8% respectively. Again, FLT-3 ITD, but not D835, was associated with high WBC, microgranular morphology and PML breakpoint at BCR3. ITD cases tended to have more relapses and an inferior DFS (Noguera et al, 2002). More recent data reported from 157 APL patients treated in the Medical Research Council AML 10 and 12 trials showed that ITD and D835 mutations occurred in 36% and 11% of cases respectively. ITD was associated with microgranular morphology, and FLT-3 mutations were a significant negative prognostic factor, owing to its association with a high WBC and induction deaths (Grimwade et al, 2003).

In this study, the frequency and subtype of FLT-3 aberrations were almost identical to those reported previously in Chinese patients (Shih et al, 2003). Our results showed that FLT-3 ITD was associated with high WBC and microgranular morphology, but not PML breakpoint at BCR3. In this series, there was a relatively high frequency of early deaths and relapses. Against this backdrop, FLT-3 ITD was associated an inferior OS, owing to its impact on induction deaths rather than increased relapse. Multivariate analysis showed that the effect of ITD was probably secondary to its association with high WBC.

The effect of FLT-3 aberrations on relapse seems to be limited. First, all studies to date have failed to show an impact of FLT-3 aberrations on relapse rates and DFS. Furthermore, FLT-3 aberrations might only exist in subclones (Whitman et al, 2001), or represent secondary changes in relapses (Kiyoi et al, 1997; Shih et al, 2003). Finally, FLT-3 aberrations might be lost at relapse. Therefore, FLT-3 up-regulation is unlikely to be a major cause of treatment failure with current therapeutic strategies in APL.

These results have therapeutic implications. As early induction death is significantly related to the male sex, high WBC, microgranular morphology and FLT-3 ITD, the use of FLT-3 antagonists upfront in patients possessing these adverse factors might be explored. As induction deaths tend to occur very early after diagnosis, empirical upfront use of FLT-3 antagonists followed by subsequent confirmation of the presence or not of FLT-3 aberrations might be a strategy to allow more time for the drugs to act. However, our results appear to indicate that the use of FLT-3 antagonists in APL patients in relapse might not be warranted. In view of the relatively small number of cases investigated in our study, further prospective trials of larger number of patients will be required to address this issue.


The authors would like to thank Stanley Yeung for statistical advice and technical assistance. This study was supported by the Kadoorie Charitable Foundation.