Therapy-related myelodysplastic syndrome–acute myelogenous leukemia in patients treated for acute promyelocytic leukemia: an emerging problem

Acute promyelocytic leukemia (APL) is a subtype of acute myelogenous leukemia (AML) that is characterized by peculiar clinical and biologic features. These include severe hemorrhagic diathesis at presentation, specific chromosome translocation t(15;17) resulting in the fusion of promyelocytic (PML) and retinoic acid receptor α (RARα) genes, and unique in vitro and in vivo responses to the differentiating agent all-transretinoic acid (ATRA).1-3 Front-line use of ATRA combined with chemotherapy has recently contributed remarkable improvement in the prognostic outlook of APL, converting this once frequently fatal leukemia to a highly curable disease.4-12 

The development of therapy-related myelodysplasia or AML (tMDS–AML) after treatment for other tumors is one of the most serious complications occurring after chemotherapy for highly curable malignancies such as breast cancer, Hodgkin disease, non-Hodgkin lymphoma, and childhood acute lymphoblastic leukemia.13-15Among chemotherapy agents, alkylating drugs and topoisomerase II inhibitors such as anthracyclines and epipodophillotoxins have been frequently associated with the development of tMDS–AML. Regarding tMDS–AML occurring after treatment for APL, sporadic cases have been reported to date,16-24 but no studies have investigated this issue by analyzing large series of patients. We report here our experience with the development of tMDS–APL in a consecutive group of 77 patients with APL treated at a single institution.

Eighty-eight patients with APL consecutively diagnosed and treated at the Department of Cellular Biotechnology and Hematology of the University La Sapienza of Roma from January 1989 to September 1998 are included in this analysis. A minimum follow-up of 2 years after completion of induction therapy was considered for enrollment into the study. Diagnoses of APL were initially established by morphologic and cytochemical criteria following the French-American-British (FAB) guidelines25 and were confirmed in all patients by Southern blot analysis, reverse transcription–polymerase chain reaction (RT-PCR), or both using specific primers and probes as described.26,27 

Two consecutive protocols were used.

Twenty-eight patients who received diagnoses from January 1989 to March 1993 were randomly assigned to undergo induction treatment with idarubicin (IDA) alone (10 mg/m² for 4 days) versus IDA at the same dosage plus cytarabine (ARA-C) (200 mg/m² continuous infusion [c.i.] for 7 days).28 Patients in complete remission (CR) were administered 3 polychemotherapy consolidation courses as reported.4,28 At the end of consolidation, patients in CR were randomly assigned to undergo maintenance therapy with methotrexate (MTX, 15 mg/m² per week) and 6-mercaptopurine (6-MP, 90 mg/m² per day) for 2 years versus no further therapy.

Sixty patients given diagnoses from April 1993 to December 1998 underwent the AIDA regimen as reported.4 

Patients were monitored at regular time intervals after the end of consolidation therapy. Bone marrow samples were collected every 3 to 4 months and were analyzed by RT-PCR for PML–RARα amplification as reported elsewhere.4,26 The diagnosis of tMDS–AML was established according to the FAB criteria.29 

Karyotypic analyses were carried out in marrow samples collected at the time of evolution in tMDS–tAML in all patients using direct technique and short-term culture (24 hours). The GTG banding method was used, and karyotypes were defined according to standard nomenclature.

Of 88 consecutive patients with newly diagnosed APL, 8 died during induction, 2 died during consolidation therapy, one did not achieve molecular remission at the end of consolidation, and the remaining 77 (87.5%) obtained hematologic and molecular remission after induction and consolidation therapy. Five patients (3 of 53 or 5.6% in the AIDA 0493 study and 2 of 24 or 8.3% in the GIMEMA 0389 study) acquired tMDS–AML during follow-up. Initial clinical features and therapies for APL in these 5 patients are reported in Table1. Patient 1 had been reported previously.30 Three patients (patients 1-3) were in first CR when tMDS–AML was diagnosed, whereas 2 patients (patients 4-5) received further treatment for APL relapse before tMDS–AML developed, including alkylating agents as part of the conditioning regimen before autologous stem cell transplantation (AuSCT). Of the 2 latter patients, patient 4 received a diagnosis of tMDS–AML when in second CR, and patient 5 acquired tMDS–AML while in third CR. In all 5 patients, RT-PCR monitoring indicated molecular remission (ie negativity of the PML–RARα test) at the time of tMDS–AML diagnosis.

The main morphologic and karyotypic features of the tMDS–AML phase—time latency between APL and tMDS–AML diagnosis, treatment received for tMDS–AML, and patient outcomes—are reported in Table2. In all patients, progressive pancytopenia was detected before diagnosis of tMDS–AML. A 2-month phase of tMDS preceded tAML diagnosis in patients 3 and 4. In patient 1, trilineage myelodysplasia was diagnosed concomitantly with tAML (FAB M4). Evolution into tAML was not detected in patients 3 and 5 (the latter underwent allogeneic SCT shortly after MDS diagnosis). Cytogenetic characterization revealed numeric abnormalities involving chromosome 5 or 7 in 2 patients (patients 2, 4), balanced t(10;11)(p14;q21) in patient 1, and a normal karyotype in patient 5, whereas it failed because of lack of evaluable metaphases in patient 3. In patient 1, the involvement of the MLL gene was ruled out by Southern blot analysis, as reported elsewhere.30 

In light of poor performance status, 3 patients (patients 2-4) received only supportive care as therapy for tMDS. Of these, one is alive with tMDS 18 months after diagnosis of tMDS, and 2 died of progressive disease shortly after tMDS development. Patient 1 underwent reinduction and consolidation therapy followed by allogeneic SCT and died on day +50 from hepatic GvHD. Finally, patient 5, who acquired MDS in third CR, received supportive care for 5 months and then underwent allogeneic SCT. She is alive and in CR from APL or MDS 12 months after SCT.

Since the advent of ATRA, APL is increasingly reported as curable.4-12 Thus, larger numbers of long-term survivors of this disease are expected in the near future, and, as a consequence, more patients will be at risk for late complications related to antileukemic treatment. In fact, with few exceptions,31conventional chemotherapy is still part of the protocol used in the front-line therapy for the disease. Furthermore, chemotherapy is considered essential to obtain sustained molecular remission, which in turn correlates with prolonged survival and potential cure.1-3 

Current APL chemotherapy protocols usually include high-dose anthracyclines, mitoxantrone, and epipodophillotoxins—in other words, topoisomerase II inhibitors whose leukemogenic potential is well established.32 All these agents were administered as part of the initial treatment in the 2 protocols reported in our study. In addition, 2 of the 5 patients described received further treatment with alkylating agents because of APL relapse. As reported in other tAML studies,13-15 evolution from tMDS to overt tAML was rapid in 2 patients. However, we also observed a patient (patient 2) with a prolonged and indolent tMDS phase that lasted more than 18 months despite the fact that the tMDS clone harbored a poor prognostic aberration (monosomy 7). Interestingly, such a lesion would have led to a diagnosis of alkylating agent–related tMDS according to the World Health Organization (WHO) classification,33 yet this patient was never administered alkylating drugs. The latter finding is in agreement with a recent report by Au et al,24 who described a patient with APL treated without alkylating agents who acquired tMDS with chromosome 5 and 7 abnormalities. Together, these observations suggest that the pathobiologic associations proposed by the WHO to distinguish epipodophillotoxin-related from alkylating-related tMDS may not always hold true. Other clinical and biologic features of patients in our study were similar to those usually reported for tMDS–AML, with a median time latency of 46 months and evidence of unexplained pancytopenia and macrocytosis preceding tMDS diagnosis.

Three patients in the current study acquired tMDS while in prolonged first hematologic and molecular remission; 2 of them died, one of disease progression and the other of transplant-related toxicity. It is conceivable that these patients were cured of APL. Because studies have also suggested that a relevant proportion of patients with newly diagnosed APL are overtreated,9,31,34 we emphasize the need for better tailoring of treatment intensity in this disease by identifying risk categories at diagnosis and by prospective minimal residual disease monitoring. Besides identifying patients at risk for relapse and in need of further treatment, the use of RT-PCR of PML–RARα during remission might spare the development of unnecessary toxicity in potentially cured patients.