The Hi’s and Lo’s of cytarabine in acute myeloid leukemia

Summary:

Cytarabine is the backbone of AML therapy, but the dose used during induction has remained controversial. Using an intermediate dose of cytarabine, compared to conventional dose, was shown to improve disease-free and overall survival in adult patients in China up to age 55, particularly in patients with intermediate cytogenetic risk.

In this issue of Clinical Cancer Research, Wei and colleagues re-examine the important and unanswered clinical question of cytarabine dose intensity in younger patients with de novo acute myeloid leukemia (AML).¹ To accomplish this, they conducted a large (591 patients) open-label, randomized trial comparing induction using conventional dose cytarabine (CDAC: 100 mg/m2/d for 7 days) and intermediate dose cytarabine (IDAC: 100mg/m2/d days 1–4 followed by 1g/m2 every 12h days 5–7), i.e., 700mg vs. 6400mg of cytarabine during the initial induction phase. Patients also received daunorubicin 40mg/m2 on days 1–3 and omacetaxine 2mg/m2 on days 1–7 (the standard induction regimen in China). In patients achieving remission, a second randomization was done prior to consolidation therapy, and half of patients received 3 cycles of high dose cytarabine (HiDAC) (3g/m2 every 12 hours for 3 days) and half received 2 cycles of IDAC (1.5g/m2 every 12h for 3 days—note this is different than the dose used during induction) combined with an anthracycline. Eligible patients could receive allogeneic hematopoietic cell transplantation (HCT). The primary endpoint was disease-free survival (DFS), and investigators and statisticians assessing outcomes were blinded to the cytarabine dose.

This was a large, well-done clinical trial that adds potentially practice-changing information to the field. Notably, the investigators demonstrated superior 3-year DFS (67% vs. 54%; Hazard Ratio [HR]=0.67; p=0.005) and survival (68% vs. 59%; HR=0.72; p=0.014) with IDAC-based induction compared to conventional dose, and this advantage was maintained after multivariable analysis and censoring subjects for HCT. To us, it is undeniable that a one-third decreased risk of death or relapse (primary endpoint) with IDAC in this population is clinically meaningful. Also, of note, the second randomization done post-remission had no impact on outcomes, suggesting that the standard repeated cycles of “HiDAC” consolidation is at least equivalent to multi-drug regimens, with the added benefit that it avoids excess anthracycline exposure. Additionally, in a sub-analysis, the survival benefit with IDAC-based induction was only seen in the patients who received HiDAC consolidation.

Looking more closely at the data, what is driving this survival benefit? Complete remission (CR) rates were higher with IDAC compared to CDAC (87% vs. 77%, p=0.004)—with the caveat that a small number of patients on the conventional dose arm were not treated for residual leukemia at day 14 (prior to a protocol amendment). In terms of safety, while the duration of neutropenia and thrombocytopenia were slightly longer in the IDAC group, there was no difference in induction deaths. In addition, rates of HCT between the two groups were the same. Collectively, these data suggest that IDAC-based induction provided most of its benefit early, with, numerically, a 10% better CR rate and ~10% improvement in survival. To what extent fewer relapses contributed to this survival difference in unclear (cumulative incidence of relapse [CIR] not given for induction cohorts). Of note, the advantage with IDAC-based induction also held true when looking at event-free survival (death, primary induction failure, and relapse).

When the authors examined patients by cytogenetic subgroup, the survival advantage was only seen in patients with intermediate risk disease. However, two-thirds of all patients on study had intermediate-risk cytogenetics, and there were very few adverse risk patients, making it difficult to interpret the sub-analysis in these groups. This skew in cytogenetic distribution may be due to selection bias implicit to clinical trial enrollment, or to patient demographics in China and differences in disease biology. Regardless, the more overall favorable cytogenetic distribution, along with the younger median age (36 years), may explain the impressive 3-year OS rate of ~64% for all patients (68% for IDAC; 59% for CDAC). It is an interesting question if the molecular genetic or epigenetic profile of patients in China differs from that of the west, or other more heterogeneous populations, and whether this could confer increased sensitivity to cytarabine. Importantly, the fact that a statistically significant and meaningful survival benefit was observed despite the overall good outcome, only re-enforces the clinical significance of using IDAC-based induction in this population, although long-term follow-up still needs to be reported as the data matures.

When we look at the body of literature, several studies have examined higher doses of cytarabine with induction, with varying study designs. In the late 1990s, 2 randomized studies compared HiDAC- to CDAC-based induction for de novo AML; one gave HiDAC 3g/m2 × 8 doses with daunorubicin and etoposide, and the other HiDAC 2g/m2 × 12 doses with daunorubicin. Despite no improvement in survival, both studies showed improvements in the CIR and relapse-free survival (RFS) in the HiDAC cohorts. In the mid-2000s, HOVON-SAKK compared HiDAC (2–3g/m2) versus a more intermediate dose of cytarabine (1g/m2) for AML induction (included de novo and secondary AML).² In this study, each group received 2 planned induction courses. The intermediate-dose group received idarubicin plus CDAC first, followed by a second induction with amsacrine plus cytarabine 1g/m2 (total cytarabine dose during induction 13.4g/m2). The high-dose group received idarubicin with cytarabine 1g/m2 first, followed by amsacrine plus cytarabine 2g/m2 (total cytarabine dose 26g/m2). Overall, this study found no significant difference in CR, relapse, EFS, or survival between the intermediate and higher dose groups. More recently, the EORTC and GIMEMA groups examined combination daunorubicin, etoposide, and cytarabine (100mg/m2 10d vs. 3g/m2 4d) induction in de novo and secondary AML patients.³ They reported statistically significant improvements in CR rate, EFS, and survival (at 6 years) for patients aged 15–45 who received HiDAC as a component of induction, as well as improved outcomes in patients with adverse cytogenetics, FLT3-ITD mutations, and secondary AML. In 2014, Wei et al conducted a meta-analysis of cytarabine dose in AML induction that assessed 6 different studies.⁴ Overall, there was no statistically significant difference in CR rate or survival between induction regimens containing HiDAC or CDAC; however, there was a statistically significant difference in RFS, favoring HiDAC. In a sub-group analysis, this RFS benefit was only recapitulated in favorable risk patients. Importantly, all of the aforementioned studies showed increased toxicity with higher doses of cytarabine, including conjunctivitis, skin reactions, and GI complications (count recovery times were comparable).

The broader question this study raises is the worldwide applicability of its findings. This is a near impossible question to answer. In China, where there are more cases of AML than in the United States and European Union combined, the results of this large randomized clinical trial clearly support the use of IDAC with induction for younger patients with de novo AML, particularly those with intermediate risk. Potential confounders that may limit generalizability include the use of omacetaxine and a daunorubicin dose of 40mg/m2 during induction. In many countries, higher doses of daunorubicin are used, and 90mg/m2 vs. 45mg/m2 has known survival benefit when given with CDAC. Another point to consider is that patients on the CDAC arm were allowed to receive a second induction based on the D14 bone marrow biopsy (although sensitivity analysis showed no difference in outcomes). The following questions could be raised and have not been (and may never be) answered: if cytarabine dose intensification were delayed until re-induction (if needed), would outcomes be any different than using IDAC upfront (and could such sequencing minimize potential excess toxicity in patients already sensitive to CDAC)? Along those lines, what is the role of HiDAC-based salvage if IDAC is used during induction? Lastly, as these are largely intermediate risk patients, some of who had a FLT3-ITD mutation (11%), what is the role of midostaurin if IDAC is used, both in terms of efficacy and safety? What about gemtuzumab ozogamicin (GO)?

Cytarabine has served as the backbone of AML therapy for decades. The use of HiDAC during consolidation has established DFS and OS benefit versus CDAC in younger patients and is standard of care.⁵ Subsequent efforts to demonstrate further survival benefit by using higher doses of cytarabine during induction have shown mixed results. One problem is that it is nearly impossible to compare studies, as the dose of cytarabine, induction drug partner(s), number of planned inductions, consolidation strategy, and patient eligibility factors (such as age and de novo vs. secondary AML) vary considerably. The current study and others also raise the question of population differences in AML biology that may affect treatment-specific outcomes, and how to compare studies across diverse groups. Regardless of these considerations, Wei and co-workers have reminded us that cytarabine is the most active drug we have in AML therapy, and that the dose matters, even during induction. This is the first study to show a survival benefit with cytarabine intensification during induction in adult patients ages 15–55. Only one other study has shown a survival benefit in adult AML, but with higher doses of cytarabine and only for ages 15–45.³ Wei et al have importantly demonstrated that using an intermediate dose of cytarabine in the 6-gram range (not HiDAC, see Table 1), is sufficient to achieve a significant survival advantage compared to CDAC. Based on this, an intermediate-dose level of cytarabine with induction should be considered in younger adults, particularly intermediate risk, FLT3-negative patients. Given that the current study enrolled mostly intermediate risk patients, it is harder to assess the impact of IDAC on favorable or adverse risk disease. Although a meta-analysis has shown RFS benefit with HiDAC-based induction in favorable risk patients, one has to consider whether this outweighs the benefit of adding GO to standard 7+3. Adverse risk AML is a more complicated topic and ideally would be treated on a clinical trial, and the existing evidence appears less convincing for using higher doses of cytarabine with induction in these patients. Lastly, there was also an important negative finding on this study: deviating from standard HiDAC consolidation therapy was not supported.

Table 1.

Summary of randomized trials comparing cytarabine dose during induction in acute myeloid leukemia

Authors	Study Subjects	Eligibility Criteria	Induction Cycles	Induction Regimens	Cumulative Cytarabine Dose in Planned Induction(s)	Consolidation	Outcomes
Bishop et al 1996	301	de novo AML, ages 15–60	1 cycle if CR, up to 3 identical induction cycles	DNR 50mg/m2 D1–3 + VP16 75mg/m2 D1–7 + cytarabine 100mg/m2 D1–7 DNR 50mg/m2 D1–3 + VP16 75mg/m2 D1–7 + cytarabine 3g/m2 q12h D1,3,5,7	700mg/m2 24g/m2	2 cycles of DNR 50mg/m2 D1–2 + VP16 75mg/m2 D1–5 + cytarabine 100mg/m2 D1–5	Improved CIR, RFS
Weick et al, 1996 *	723	de novo or secondary AML, ages 15–64	1 cycle	DNR 45mg/m2 D5–7 + cytarabine 200mg/m2 × D1–7 DNR 45mg/m2 D7–9 + cytarabine 2g/m2 q12h D1–6	1.4mg/m2 24g/m2	CR: Randomized to 2 cycles of DNR 45mg/m2 D6–7 + cytarabine 200mg/m2 D1–7 vs 1 cycle of DNR 45mg/m2 D6–7+ cytarabine 2g/m2 q12h D1–5 CR: Nonrandomly assigned to DNR 45mg/m2 D6–7 + cytarabine 2g/m2 D1–5	Improved CIR, RFS
Buchner et al, 2006 **	1770	de novo or secondary AML, ages 16–85 (median 60)	2 cycles	C1: DNR 60mg/m2 D3–5 + TG 100mg/m2 q12h D3–9 + cytarabine 100mg/m2 D1–8; C2: MTZ 10mg/m2 D3–5 + cytarabine 3g/m2 q12h D1–3 C1: MTZ 10mg/m2 D3–5 + cytarabine 3g/m2 q12h D1–3; C2: MTZ 10mg/m2 D3–5 + cytarabine 3g/m2 q12h D1–3 Note: For patients ≥60, higher dose cytarabine was dosed at 1g/m2 q12h D1–3	18.8g/m2 36g/m2	CR: 1 course of DNR 60mg/m2 D3–5 + TG 100mg/m2 q12h D3–9 + cytarabine 100mg/m2 D1–8 After consolidation chemotherapy, randomized to maintenance or autoSCT	No significant difference
Buchner et al, 1999 **	725	de novo AML, ages 16–60 (median 44)	2 cycles	C1: DNR 60mg/m2 D3–5 + TG 100mg/m2 q12h D3–9 + cytarabine 100mg/m2 D1–8; C2: C1: DNR 60mg/m2 D3–5 + TG 100mg/m2 q12h D3–9 + cytarabine 100mg/m2 D1–8 C1: DNR 60mg/m2 D3–5 + TG 100mg/m2 q12h D3–9 + cytarabine 100mg/m2 D1–8; C2: MTZ 10mg/m2 D3–5 + cytarabine 3g/m2 q12h D1–3	1.6g/m2 18.8g/m2	CR: 1 course of DNR 60mg/m2 D3–5 + TG 100mg/m2 q12h D3–9 + cytarabine 100mg/m2 D1–8 Consolidation followed by maintenance AlloSCT if <50 w/ sibling donor	No significant difference Subgroup analysis poor-risk (>40% residual blasts, adverse cytogenetics, elevated LDH): Improved CR, EFS, and OS
Buchner et al, 2009	1284	de novo AML, ages 16–85	2 cycles	C1: DNR 60mg/m2 D3–5 + TG 100mg/m2 q12h D3–9 + cytarabine 100mg/m2 D1–8; C2: MTZ 10mg/m2 D3–5 + cytarabine 3g/m2 q12h D1–3 C1: MTZ 10mg/m2 D3–5 + cytarabine 3g/m2 q12h D1–3; C2: MTZ 10mg/m2 D3–5 + cytarabine 3g/m2 q12h D1–3 Note: For patients ≥60, higher dose cytarabine was dosed at 1g/m2 q12h D1–3	18.8g/m2 36g/m2	CR: 1 course of DNR 60mg/m2 D3–5 + TG 100mg/m2 q12h D3–9 + cytarabine 100mg/m2 D1–8 Random assignment to autoSCT or maintenance	No significant difference
Lowenberg et al, 2011	858	de novo AML or RAEB (IPSS≥1.5), ages 18–60 (median 49)	2 cycles	C1: IDA 12mg/m2 D5–7 + cytarabine 200mg/m2 D1–7; C2:amsacrine 120mg/m2 D3,5,7 + cytarabine 1g/m2 q12h D1–6 C1: IDA 12mg/m2 D5–7 + cytarabine 1g/m2 q12h D1–5; C2: amsacrine 120mg/m2 D3,5,7 + cytarabine 2g/m2 q12h D1,2,4,6	13.4g/m2 26g/m2	If CR after 2nd induction, risk adapted strategy of HiDAC vs alloSCT vs autoSCT	No significant difference
Willemze et al, 2013	1942	de novo or secondary AML, ages 15–60 (median 45)	1 cycle if CR, 2nd identical induction if PR	DNR 50mg/m2 D1,3,5 + VP16 50mg/m2 D1–5 + cytarabine 100mg/m2 D1–10 DNR 50mg/m2 D1,3,5 + VP16 50mg/m2 D1–5 + cytarabine 3g/m2 q12h D1,3,5	1g/m2 18g/m2	If CR, one cycle of DNR 50mg/m2 D4–6 + cytarabine 500mg/m2 q12h D1–5 AlloSCT if suitable donor, otherwise autoSCT	Improved CR, EFS, and OS (ages <46)
Wei et al, 2020	591	de novo AML, ages 15–55 (median 36)	Up to 2 induction cycles in conventional cytarabine arm if CR not accomplished after C1	DNR 40mg/m2 D1–3 + omacetaxine 2mg/m2 D1–7 + cytarabine 100mg/m2 D1–7 DNR 40mg/m2 D1–3 + omacetaxine 2mg/m2 D1–7 + cytarabine 1g/m2 q12h D5–7	700mg/m2 6g/m2	If CR, randomized to 1st consolidation with DNR 40mg/m2 D1–3 + cytarabine 3g/m2 q12h D1–3 vs DNR 40mg/m2 D1–3 + cytarabine 1.5mg/m2 q12h D1–3 C2: MTZ 6mg/m2 D1–3	Improved CR, DFS, EFS, OS

^*Initially HiDAC was dosed 3g/m2 in induction and consolidation for pts<50, however this was reduced to 2g/m2 In 1988 (DMC). In all consolidation arms, DNR was reduced to 30mg/m2 for ages 50–64.

^**Maintenance therapy after consolidation consisting of monthly cytarabine 100mg/m2 q12h SQ D1–5 paired with alternating DNR 45mg/m2 D3–4, TG 100mg/m2 q12h D1–5, and cyclophosphamide 1g/m2 D3 for 3 year duration