Improved Response To 5-Azacitidine In Patients With Primary Compared To Secondary Aml, Particularly If Npm1 Mutations Are Present

Abstract Abstract 3562 5-azacitidine (5-aza) is a widely-used agent for treatment of myelodysplastic syndromes (MDS). In contrast, data on its use in AML are far more limited and, to date, little is known on the clinical or biological markers predicting the response of AML to this drug. We retrospectively analysed a series of patients with AML who received 5-aza due to old age, poor condition or refractoriness to standard treatment, and investigated the potential association of a variety of genetic markers with response and survival. PATIENTS AND METHODS: From 2006 to 2011, 38 patients (23 males, 15 females) diagnosed as AML according to the WHO criteria1 received treatment with 5-aza (75 mg/m2 SQ daily, 5–7 days) at our institution as on-label or compassionate use program. Twenty-four patients (median age 68 years, range 37–85) had primary AML (pAML). Most of them (n=18) had failed to prior standard AML treatment: one line in 12 cases, 2 lines in 5 (including stem cell transplant [SCT] in 2 cases), and 3 lines in one. Fourteen additional patients (median age 68.3; range 32–88) had secondary AML (sAML); among them 8 previously treated: 6 patients had received one line of standard treatment before 5-aza, 1 patient 2 lines, and 1 patient 3 different lines. Twenty-two pAML and 14 sAML cases were screened for the presence of karyotype abnormalities. Screening for gene mutations [FLT3 internal tandem duplication (FLT3/ITD, nucleophosmin-1 gene mutation (NPM1), MLL partial tandem duplication, CCAAT/enhancer binding protein alpha mutation (CEBPA), AML1-ETO and CBFB-MYH11] was performed prior to starting 5-aza whenever possible. The response to treatment was usually assessed between the 4th and 6th courses, and classified according to WHO criteria. RESULTS: Patients received a median 7 courses (range 1–19) of standard-dose 5-aza. Within the pAML group, 13 patients responded [8 complete responses (CR) and 5 partial responses (PR)], and 4 patients with sAML responded (2 CR and 2 PR). Two patients in the pAML group and 1 patient with sAML subsequently received an SCT. After a median 218 days (10–792) follow-up in the pAML group, and 246 (19–995) in the sAML group, 66.7% patients with pAML and 28.6% with sAML were alive. Overall, no significant differences were seen in the rate of patients achieving CR or PR between the pAML and the sAML groups; despite this, the estimate 2-year survival was significantly higher in patients with pAML (p=0.01). Prior exposure to chemotherapy or other forms of treatment had no impact on response or survival. An abnormal karyotype (present in 12 pAML and 8 sAML patients) did not influence the response to treatment. All mutation analysis yielded negative results in sAML cases. In pAML cases, 3/20 were positive for FLT3/ITD, 5/19 for NPM1 and 1/19 for MLL. Most interestingly, NPM1 mutations in the absence of FLT3/ITD, were significantly (p=0.05) associated with improved response rate (5/5) in patients with pAML with no deaths being observed in this group. DISCUSION: Our results show a higher probability of survival after 5-aza treatment in patients with pAML compared to sAML. The presence of NPM1 mutations appeared to identify a subset of pAML patients that, despite having failed first line treatment, may be particularly sensitive to this therapy. If confirmed in larger series, this finding would be of the utmost interest for offering these patients a low toxicity rescue therapy or a maintenance strategy with his agent. Disclosures: No relevant conflicts of interest to declare.