In search of combinations for the treatment of patients with high-risk acute myeloid leukemia


  • Pellegrino Musto MD,

    Corresponding author
    1. Unit of Hematology and Stem Cell Transplantation, IRCCS-CROB, Reference Cancer Center of Basilicata, Rionero in Vulture, Italy
    • IRCCS - CROB, Reference Cancer Center of Basilicata, Hematology and Stem Cell Transplantation, Strada Provinciale, 8, Rionero in Vulture, Pz 85028, Italy
    Search for more papers by this author
    • Fax: (011) 0039-0972726217

  • Felicetto Ferrara MD

    1. Division of Hematology and Stem Cell Transplantation Unit, Cardarelli Hospital, Naples, Italy
    Search for more papers by this author

  • See referenced original article on pages 2090–6, this issue.


Although relapse still represents the most relevant obstacle to achieving a cure in acute myeloid leukemia, it offers an opportunity to investigate new therapeutic strategies, mainly in patients who are expected to gain negligible advantage from conventional salvage chemotherapy. In this setting, clofarabine represents 1 of the most promising drugs.

The clinical outcome of acute myeloid leukemia (AML) is extremely variable, ranging from survival of a few days to cure, depending on different patient clinical characteristics, disease biology, and treatment intensity.1 Causes of therapeutic failure include death in induction and, more frequently, resistant or relapsed disease. Over the past few years, the application of novel cytogenetic and molecular techniques has markedly improved our knowledge on the pathophysiology of the disease, resulting in new potential therapies.2 Despite this, current therapeutic results are still unsatisfactory, particularly in older patients and/or in those with primary refractory or relapsed disease, because of difficulties in translating new pathogenetic insights from the bench to the bedside. Accordingly, salvage chemotherapy still represents the best option, at least for young adult patients who relapse after achieving first complete remission (CR1).3 In contrast, the actual advantage of salvage chemotherapy is more controversial where older AML patients are concerned, especially in those with a CR1 duration of less than 12 months. In this subset, the second complete remission (CR2) rate is usually lower than 30%, with substantial toxicity and mortality and few possibilities of receiving stem cell transplantation (SCT), although more patients in the last few years are being considered for reduced intensity allogeneic SCT.4 Gemtuzumab ozogamycin (GO), an immunoconjugate composed of a humanized anti-CD33 antibody linked to the potent antitumor antibiotic calicheamicin, results in a 20% to 25% CR2 rate in elderly patients with relapsed AML, when used as a single agent; however, the majority of patients will ultimately relapse.5

Overall, most therapeutic results for relapsed AML derive from retrospective studies based on salvage regimens including intermediate- or high-dose cytosine arabinoside (Ara-C), with percentages of CR2 ranging from 30% to 60%, and possibility of long-term survival strictly related to eligibility for SCT. At the moment, there is no gold standard salvage chemotherapy, and the therapeutic choice mostly depends on personal experience concerning efficacy and toxicity. The addition to chemotherapy of GO has been reported as potentially useful in increasing CR rate and duration; however, toxicity of the combination is not negligible and can be substantial at the hepatic level.5 Whatever the salvage therapy, best results are achievable when SCT is feasible. On this basis, characteristics of an ideal salvage regimen should include both antileukemic efficacy and acceptable extrahematologic toxicity. In clinical practice, it is indeed not rare that toxicity from previous therapy may preclude the feasibility of SCT.

The wide differences in CR2 rates and durations in various phase 2 single-arm trials have clearly shown that biological and clinical characteristics of the patients rather than different treatment modalities account for discordant results. Relevant prognostic factors at relapse are age, cytogenetics at presentation, previous SCT, and duration of CR1.3 In particular, the prognosis of AML patients who relapse less than 12 months from CR1 achievement is particularly poor after conventional salvage chemotherapy; therefore, new strategies should be designed in this patient subset.

Although induction chemotherapy regimens have not substantially changed in over 20 years, exciting results have emerged from several phase 2 studies, based on the adoption of new agents either in relapsed/refractory AML patients or in newly diagnosed older ones with unfavorable characteristics at diagnosis and/or who are not eligible for intensive treatment. Among the new different cytotoxic agents that have been investigated recently in high-risk AML, clofarabine, a second generation purine nucleoside analogue designed to overcome the limitations and to incorporate the best qualities of both cladribine and fludarabine, is particularly notable. In 2004, the drug was approved by the United States Food and Drug Administration for the treatment of pediatric relapsed/refractory acute lymphoblastic leukemia after at least 2 prior regimens.6 However, clofarabine has also shown significant efficacy in high-risk AML and myelodysplastic syndrome (MDS), either alone or in combination in adult patients. In particular, exciting results have been reported by using clofarabine as a first-line single agent in elderly unfit patients, with a response rate approximating 50%, including patients with unfavorable cytogenetics at diagnosis.7 In refractory/relapsed disease, the drug has been mainly used in combination with Ara-C, given that there is a strong biological rationale for adopting such an association.6 Preclinical studies, in fact, have clearly shown that pretreatment with clofarabine leads to increases in intracellular Ara-C triphosphate levels via indirect up-regulation of deoxycytidine kinase (dCK) because of decreased 2′deoxynucleoside-S′triphosphate pools caused by clofarabine's inhibition of ribonucleotide reductase. In addition, because clofarabine inhibits both ribonucleotide reductase and DNA polymerases, logical combination regimens would include clofarabine and agents that damage DNA and inhibit DNA repair, such as anthracyclines. Clofarabine could also be combined with agents inducing apoptosis, because it has been shown to activate caspases, with potential therapeutic efficacy in the treatment of indolent lymphoproliferative disorders.

In this issue of Cancer, Faderl and coworkers designed a phase 1 study of clofarabine ± Ara-C, plus idarubicin.8 Patients with primary refractory or first-relapse AML were assigned to either clofarabine plus idarubicin (CI) if previously exposed to Ara-C with a response lasting <12 months, or clofarabine and idarubicin plus Ara-C (CIA) for responses ≥12 months or if never exposed to Ara-C. Interestingly, in both subgroups there was a considerable percentage of patients (38% and 39%, respectively) with primary refractory disease, and patients relapsing after allogeneic SCT were also included. In addition, 20% of patients in the entire cohort had experienced prior malignancies, and all but 1 of these patients had received prior chemotherapy and/or radiotherapy. Finally, the median age for both subgroups was older than 55 years.

A standard “3+3” phase 1 design was followed to define maximum tolerated dose. Response rate was higher for patients receiving CIA, characterized, however, by more favorable prognostic features, mainly more prolonged duration of CR1. In addition, in this subgroup duration of response was significantly longer compared with CI. In either group, responses occurred after 1 course of therapy, and were observed at all dose levels; of note, 30% of patients could be referred for a SCT. Dose-limiting toxicity (DLT) included hyperbilirubinemia and elevated liver enzymes for CI-treated patients, whereas in those receiving CIA, mucositis and diarrhea were major factors accounting for DLT. Myelosuppression was ubiquitous, as expected; however, it was prolonged in only 13% of patients treated with CI and 5% of those given CIA. Nonhematologic toxicities, in particular hand-foot syndrome, did occur frequently, although it was generally reversible. Other drug-related side effects included skin rash (mainly in the CIA subgroup) and facial flushing. In general, toxicity was no different compared with that recorded in previous studies based on combination of clofarabine with Ara-C, and it is worthy of note that 30% of patients resulted as eligible for allogeneic SCT. In this setting, it would be of interest to have more information on the toxicity of allogeneic SCT in this group, especially at the hepatic level, given that clofarabine was associated in both arms with significant liver toxicity.

Data from this trial clearly account for the activity of clofarabine in high-risk AML. Obviously, no definite conclusions can be drawn regarding the different response rates observed by using the 2 schemes, as the aim of the study was not to make a comparison. However, the 48% response rate reported for the CIA combination is considerable as compared with results achieved with single-agent clofarabine (30%-35% CR2), and, as the authors report, the combination is currently under investigation in randomized studies versus clofarabine + Ara-C in relapsed or refractory acute leukemia.

The study has 2 main points of interest: first, patients were stratified according to CR1 duration and previous exposure to Ara-C; second, 2 different combinations of clofarabine with Ara-C and/or idarubicin were explored. On the contrary, the possibility of exploring the CI combination in patients never exposed to Ara-C seems to us of more limited clinical interest, because in the daily practice virtually all AML patients judged as eligible for conventional chemotherapy are given Ara-C. Nonetheless, data can be useful for the minority of patients who are initially recruited into experimental therapies based on new drugs such as hypomethylating agents or other biological response modifiers that are in turn able to induce CR. In the study by Faderl and coworkers, pretreatment with these agents was allowed, as well as with monoclonal antibodies such as GO. However, the percentage of patients receiving these agents was not reported, nor was the percentage of patients older than 70 years indicated; in theory, these 2 factors could have affected the different response rate observed in the 2 arms.

Although clofarabine acts as a myelosuppressive agent, its toxicity profile makes the drug potentially useful for patients excluded from intensive chemotherapy at diagnosis who have relapsed after CR was achieved with investigational therapy. A major hope is that it can replace Ara-C in induction treatment in newly diagnosed patients with the aim of increasing CR rate and duration. With this in view, a study from Medical Research Council is now comparing clofarabine versus Ara-C ± GO ± anthracylines in the elderly fit AML patient population. If results will be more favorable with clofarabine, a new standard therapeutic approach could be defined after more than 20 years. Additional trials are currently exploring the potential utility of the drug in young-adult AML patients in combination with anthracyclines and Ara-C. Data from unfit patients, including those with poor cytogenetics, are particularly exciting and need to be confirmed in larger studies in which the definition of “unfit” should be well established and reproducible among the different studies. In this regard, different comorbidity scores based on clinical and biological characteristics at diagnosis have been proposed and could be adopted. Experimental trials combining clofarabine with biologic response modifiers represent a further field of investigation, either in frail patients or in older fit patients with poor cytogenetics.

A recent analysis by Sekeres and coworkers has clearly shown that in an AML elderly patient population, a delay of a few days to acquire karyotypic information before starting induction chemotherapy had no impact on CR achievement and survival.9 Therefore, in the near future it would be desirable to stratify patients not only by comorbidity, performance status, or other clinical features, but also by well-defined cytogenetic and molecular findings. In high-risk MDS, phase 2 studies with clofarabine as a single agent have provided interesting results6; therefore, in this setting, either in newly diagnosed patients or in those progressing after hypomethylating agents, the only drugs currently registered for high-risk MDS, further exploration with clofarabine singly or in combination is warranted.

Finally, in contrast to other purine analogues, clofarabine demonstrates enhanced oral bioavailability. Data on the efficacy and tolerability of the oral formulation are currently limited, but different studies are ongoing in hematologic malignancies and solid tumors. The possibility of using oral clofarabine as maintenance therapy to suppress the minimal residual disease could also be taken into account.

In conclusion, although relapse still represents the most relevant obstacle to achieving a cure in AML, it offers an opportunity to investigate new therapeutic strategies, mainly in patients who are expected to gain negligible advantage from conventional salvage chemotherapy. Elderly patients with unfavorable prognostic features at presentation represent another ideal target for new agents. The development of well-designed randomized phase 2 trials, based on multiple outcomes, would allow several therapies to be evaluated, providing new and exciting knowledge.