Abstract

Abstract 18

In a significant proportion of older patients with AML intensive chemotherapy is not considered a viable option[1]. Such patients may receive low-dose Ara-C (LDAC), best supportive care (BSC) with hydroxyurea or an experimental agent but outcomes are poor. We have shown that LDAC is superior to BSC [2], but only in patients who enter CR (fewer than 20%).[2]. There is therefore scope to improve outcomes in these patients, and a number of possible treatments have been evaluated in the randomised UK NCRI AML16 trial, one of which is gemtuzumab ozogamicin (GO), which we have shown to benefit the majority of younger patients when given in conjunction with standard chemotherapy[3].

In AML16 novel agents or combinations are tested in untreated older patients with AML or high risk MDS (marrow blasts >10%) using a “pick a winner” design. The design allows unpromising treatments to be identified early (typically after 50 or 100 patients per arm): only those arms which show promise will continue to a trial with OS and DFS as endpoints. The aim is to at least double the remission rate from 15% to 30%, and thus improve overall survival.

We now report the results of LD Ara-C (20mg bd days 1–10 for 4 courses) versus LDAC combined with GO (at a fixed dose of 5mg on day 1 of each course for 4 courses at 6–8 week intervals).

Patient Details:

The comparison opened as part of the UK LRF AML14 trial and was carried forward to AML16 unchanged. Between June 2004 and June 2010, 495 patients were randomised, 249 to LDAC plus ATO, 246 to LDAC. The median age was 75 years (range 54–90); 83% of patients were aged over 70 years, and 61% were male. To be eligible patients’ LFTs had to be less than twice ULN. Treatment arms were balanced for age, gender, performance status, de novo/secondary AML/high risk MDS, presenting WBC and cytogenetic risk group. Follow-up is complete to 1st January 2010, with median follow up of 21 months. Survival and remission data is available on 412, 404 patients respectively.

Treatment Results:

The trial passed through both stopping hurdles based on CR/CRi rates and therefore continued to full accrual, with overall survival as primary outcome measure. The table shows the distribution of outcomes: there was no heterogeneity by recruitment period (AML14 vs AML16).

CR/CRi30-day mortality8 week mortality12 month survival from CR12 month OS
LD Ara-C (n±205) 31/200 (16%) 17% 27% 58% 27% 
LD Ara-C + GO (n=207) 61/204 (30%) 16% 28% 51% 28% 
OR/HR & 95% CI 0.44 (0.28–0.70)   1.81 (1.00–3.25) 1.00 (0.80–1.24) 
P-value 0.0005   0.05 1.0 
CR/CRi30-day mortality8 week mortality12 month survival from CR12 month OS
LD Ara-C (n±205) 31/200 (16%) 17% 27% 58% 27% 
LD Ara-C + GO (n=207) 61/204 (30%) 16% 28% 51% 28% 
OR/HR & 95% CI 0.44 (0.28–0.70)   1.81 (1.00–3.25) 1.00 (0.80–1.24) 
P-value 0.0005   0.05 1.0 

The causes of death (316) were:-

InfectionHaemorrhage/CVAResistant diseaseCardiacRelapseOther
LD Ara-C (n=157) 44 79 18 
LD Ara-C + GO (n=159) 34 58 23 31 
InfectionHaemorrhage/CVAResistant diseaseCardiacRelapseOther
LD Ara-C (n=157) 44 79 18 
LD Ara-C + GO (n=159) 34 58 23 31 

There were no significant interactions between treatment and any of the baseline variables on either remission or survival outcomes. Likewise there were no major toxicity implications, although resource usage tended to be higher in patients given GO.

Discussion: While the addition of GO significantly improves CR rate, achieving the 30% response originally sought, this does not translate into survival. This is predominantly due to an increase in relapse, indicating that if GO is to have a role in this setting, it will require effective treatment to maintain remission.

[1] Juliusson G et al. Blood 2009; 113: 4179 – 4187

[2] Burnett et al. Cancer 2007 109: 1114–1124

[3] Burnett et al. JCO 2010 to appear.

Disclosures:

Off Label Use: Mylotarg (gemtuzumab ozogamicin).

Author notes

*

Asterisk with author names denotes non-ASH members.