Slow relapse in acute myeloid leukemia with inv(16) or t(16;16)

Acute myeloid leukemia (AML) with inv(16)(p13q22) or t(16;16)(p13;q22), resulting in CBFβ-MYH11 fusion transcript detectable by RT-PCR and RQ-PCR, is associated with an overall good prognosis, relapses however still occur in 30–35 % of patients, and with higher frequency in older patients.1,2 Molecular relapse generally precedes hematologic relapse in AML with balanced translocations such as t(15;17) or t(8;21), but generally by only a few weeks or a few months.3–9 Fewer data are available in AML with inv(16)/t(16;16). In 4 of the 5 relapses we observed in AML with inv(16)/t(16;16), the interval between molecular and hematologic relapse was prolonged.

Between 2005 and 2009, 9 AML patients with inv(16) or t(16;16), with a median age of 60 years (range, 21–75) who had reached CR using French co-operative AML trials (anthracycline–cytarabine induction chemotherapy followed by consolidation chemotherapy with high-dose cytarabine, or intermediate dose cytarabine in elderly patients) at our center were prospectively monitored for minimal residual disease (MRD) based on CBFβ-MYH11 fusion transcript levels in bone marrow samples.

RQ-PCR was performed on bone marrow cells according to the Europe Against Cancer (EAC) Program recommendations for CBFβ-MYH11 fusion transcripts (type A, D or E), using Taqman technology, on an ABI PRISM 7000 (Applied Biosystems).10 Quantification of CBFb-MYH11 fusion transcripts was normalized to the housekeeping ABL gene. Results were expressed by the ratio CBFβ-MYH11 copy number/ABL copy number x 100 (%).10 Median follow-up after CR achievement was 18 months (range, 3–33) and median number of MRD analyses per patient was 6 (range 1–9). Molecular relapse was defined as a 10-fold or greater increase in CBFβ-MYH11 transcript level compared to the lowest level achieved.

Table 1.Main characteristics of the 4 relapsing patients analyzed.

Five of the 9 patients relapsed, after 11–23 months, in the bone marrow (no extramedullary relapse was seen). In one of them, the interval between molecular relapse and hematologic relapse was short (one month). The 4 other hematologic relapses occurred slowly, and were preceded in all cases by molecular relapse detected in bone marrow samples, by ten (patient n. 1), six (patient n. 2), seven (patient n. 3) and eight (patient n. 4) months, respectively. Baseline characteristics of those 4 patients are shown in Table 1. Patients ns. 2 and 3 had c-KIT mutation in exon 8 and c-KIT D816V mutation, respectively (versus none of the patients who did not relapse) and patient n. 3 had N-RAS mutation, while no patient had FLT3-ITD or FLT3-835/I836 mutation. All 4 patients had achieved at least a 3-log reduction of the fusion transcript level, after induction therapy in 3 of them, and after the first consolidation course in patient n. 1 (Figure 1). During the period of isolated molecular relapse, blood counts and marrow aspirates remained normal in patients ns. 3 and 4, while cytopenias reappeared in patient n. 1 and abnormal marrow eosinophils in patient n. 2. A second CR was achieved in the 4 patients with chemotherapy, combined to gemtuzumab in 3 cases. Three of them were subsequently allografted, and all 4 patients were alive 2–11 months after hematologic relapse. In acute promyelocytic leukemia (APL), the median interval between molecular and hematologic relapse was 3–4 months in published litterature3–5 while in AML with t(8;21) it was generally less than six months6–8 and at a median of three months in our experience.9

Figure 1.Minimal residual disease (MRD) sequentially measured by RQ-PCR in bone marrow samples in the 4 relapsing patients. MRD was expressed in CBFB-MYH11 copies per /100 ABL gene copies. Dots on curves represent MRD examinations. Full arrows indicate molecular relapse and empty arrows indicate hematologic relapse. Follow-up is in months.

In AML with inv(16)/t(16;16), few studies are available: Schnittger et al. reported 6 relapses of CBFβ-MYH11 AML with an interval between molecular and hematologic relapse ranging from 1–5 months,8 similar to what they observed in AML with PML-RAR and AML1-ETO. By contrast, Stentoft et al. reported, in 4 relapsing inv(16)/t(16;16) AML, an interval between molecular and hematologic relapse of approximately one year, with a slow molecular progression rate of about 1-log per 100 days.8 In our patients ns. 2, 3 and 4, the increase in fusion transcript levels had comparable kinetics, while in our patient n. 1, it was even slightly slower, with a molecular progression rate of about 1-log per 130 days.

Thus, AML with inv(16)/t(16;16) AML may frequently relapse more slowly than other types of AML with balanced translocations. This interval between molecular and hematologic relapse may justify frequent MRD monitoring in those patients, and therapeutic intervention before overt relapse.

Acknowledgments

we thank the staff of the Department of Clinical Haematology, Hôpital Avicenne, Bobigny for the care of the patients in this study.