Prognostic significance of additional cytogenetic abnormalities and FLT3 mutations in acute promyelocytic leukemia

Acute promyelocytic leukemia (APL) is now a very curable malignancy. Clinical trials report long-term disease-free survival in approximately 85% of patients treated with all-trans retinoic acid (ATRA) and anthracycline chemotherapy (reviewed [1]), or with regimens including arsenic trioxide (ATO) [2–4]. Nevertheless, approximately 10% of patients will experience a relapse, including some who are refractory to further therapy or die from disease- or treatment-related complications, particularly hemorrhage and differentiation syndrome. It would therefore be very helpful to be able to prospectively identify patients at increased risk of relapse to inform therapy, and in whom close molecular monitoring for PML–RARA fusion transcripts by real-time quantitative polymerase chain reaction (RT-qPCR) assays could be employed to allow early intervention in the event of treatment failure [5]. Conversely, patients identified as unlikely to relapse could be spared repeated bone marrow aspirations and possibly selected for de-intensified treatment.

The Sanz score, which is based on the presenting peripheral leukocyte and platelet counts, forms the mainstay for risk stratification in many APL clinical trials; patients with a white blood cell (WBC) count of ≥ 10 × 10⁹/L are deemed high risk, with an approximately 30% chance of relapse following ATRA + anthracycline-based chemotherapy [6]. Other factors identified to be associated with an adverse outcome include CD56 expression [7] and a polymorphism in the promoter of the CD95 gene [8]. However, none of these factors accurately identifies all patients destined to relapse.

Another pretreatment parameter which has been extensively investigated as a potential prognostic indicator is mutation in the FLT3 gene, which is found in a significant proportion of cases of APL. Internal tandem duplications in exon 14 (FLT3 ITD) are associated with a high presenting WBC count, a greater risk of induction death and have variably predicted an increased risk of relapse following ATRA and anthracycline-based therapy [9,10,23]. Additionally, an intriguing relationship exists between the presence of FLT3 ITD, the short or BCR3 isoform of PML–RARα, hypogranular (or M3 variant) morphology [9], and the expression of CD34 [11] and T-lineage markers such as CD2 and CD3 [12]. These features are poorly understood, but likely reflect differences in underlying disease biology (reviewed [13]).

The study by Poiré et al. in this issue [14] examines the association between FLT3 mutations, cytogenetic abnormalities and relapse in a subset of 245 patients with PML–RARA+ APL from the North American Intergroup trial C9710. This study for newly diagnosed APL evaluated the impact of the addition of two courses of ATO in consolidation, showing improved survival in comparison to patients randomized to receive just ATRA and chemotherapy as frontline therapy. Forty-eight percent of the cases of APL with available material for molecular analysis had mutations in FLT3, including 31% with FLT3 ITD. This is consistent with previous reports [9,10,15], and provides further evidence that formation of the PML–RARA fusion by the t(15;17) is insufficient for APL pathogenesis, requiring additional mutations which may themselves influence the disease phenotype.

In contrast to previously published studies [9,10], Poiré et al. did not observe an association between the presence of FLT3 mutation and early death; however, the authors noted that induction death was not increased in patients with high presenting WBC count in the study cohort, in contrast to the trial as a whole, which may be a reflection of selection bias and/or smaller sample size. There was also no association between the presence of FLT3 mutations and overall or disease-free survival, regardless of whether patients were treated with ATO or with standard therapy. This further suggests that FLT3 mutation status does not provide a reliable predictor of relapse risk in APL, where outcomes are generally good, in marked contrast to the situation in other types of acute myeloid leukemia (AML).

Poiré and colleagues also examined the effect of additional chromosome abnormalities on outcome. In the majority of cases of APL, t(15;17) occurs in isolation, with additional changes seen in about ˜25%. Some studies report an inverse relationship between the presence of FLT3 mutations and additional chromosome abnormalities [9,16,17], leading to speculation that they too may be cooperating events in leukemia initiation. The most frequent additional chromosome abnormality in APL is a gain of chromosome 8 or 8q, one of the effects of which is to increase the copy-number of MYC. Poiré et al. report the presence of a single additional cytogenetic abnormality in 21% of patients, including 11% with gains in chromosome 8 or 8q. Their analysis confirms previously published reports showing no prognostic significance of a single additional chromosome abnormality [18,19].

Approximately 7% of patients reported by Poiré and colleagues had more than one additional chromosome abnormality, consistent with previous reports [20,21]. Interestingly, these patients (n = 15) had a poorer outcome, with lower CR rate (73% vs. > 90% in patients with < 3 cytogenetic abnormalities) and decreased overall survival (53% vs. 81% at 5 years, p = 0.001). Seven of the 15 patients died, with deaths spread across both treatment groups, including four from relapse. It is notable that three of the only seven relapses observed in patients treated with arsenic in the entire study occurred in patients with more than one additional chromosome abnormality.

The cytogenetic changes seen in the cases with three or more abnormalities are interesting to note, and merit further study. It is striking that nearly half of the patients had additional chromosome 17 abnormalities. Four out of 15 had ider(17q), an abnormality in which an isochromosome is formed from the derivative chromosome 17 after the 15;17 translocation has taken place. Isochromosome 17 is reported in 1–2% of AML [18] and ˜1% of cases of APL [17]. ider(17q) has the effect of removing one copy of TP53; additionally, an extra copy of the reciprocal RARA–PML fusion is generated [22]; it is unclear whether this is biologically significant. A further two of these patients had monosomy 17, which as well as reducing TP53 copy number would remove the remaining normal RARA allele. No separate outcome data are provided for patients with chromosome 17 abnormalities; further data on these patients would be of interest, as TP53 inactivation is a potential mechanism of therapy resistance.

Additionally, seven of 15 patients with APL with three or more chromosome abnormalities had evidence of at least two distinct cytogenetic clones. It remains unproven whether patients with APL with multiple separate clones have an inferior outcome, although theoretically such patients may be more likely to develop therapy resistance through clonal selection.

The study from Poiré and colleagues indicates that patients with two or more cytogenetic abnormalities accompanying the t(15;17) have a significant risk of relapse or treatment failure, even when treated with highly efficacious therapy including ATO. These findings merit investigation in much larger cohorts of patients, and lend further support to the routine use of cytogenetics in the diagnostic work-up of APL [1]. While it is premature to intensify therapy in APL cases with three or more cytogenetic abnormalities on the basis of current evidence, this may represent a subgroup in whom regular molecular monitoring during and after treatment may be of benefit.

Acknowledgements

RD and DG gratefully acknowledge support from Leukaemia & Lymphoma Research and thank Christine Harrison for helpful discussions.

Potential conflict of interest

Disclosure forms provided by the authors are available with the full text of this article at www.informahealthcare.com/lal.