Frequency and Prognostic Relevance of Gene Mutations in Pediatric AML Patients At First Relapse.

Abstract 2480

Relapsed pediatric AML patients respond poorly to conventional re-induction therapy and as a consequence long term survival rates of these patients are low (up to 36%). Accurate characterization and risk group stratification at first relapse may facilitate personalized, targeted re-induction therapies for these patients and ultimately improve outcome. Gene mutations provide putative targets in personalized treatment (e.g. FLT3/ITD, KIT or RAS mutations) and their incidence and prognostic relevance in pediatric AML at first relapse remains to be elucidated. Changes in mutational status occur during therapy; hence, assessment of mutations at first relapse is warranted to establish the accurate mutation status of the leukemic cells before the start of salvage treatment.

Mutational status at first relapse as well as clinical and karyotypic data were retrospectively analyzed in a large set (n = 240) of relapsed non-FAB M3, non-Down syndrome AML patients younger than 19 years. The majority of patients (88%) were uniformly treated at first relapse according to the ‘Relapsed AML 2001/01’ protocol of the International Berlin-Frankfurt-Münster (BFM) study group that involved two re-induction courses of chemotherapy with FLAG as standard treatment with or without liposomal daunorubicin (randomization), followed by allogeneic stem cell transplantation in most cases. Other patients received FLAG based (5%) or other high dose cytarabin based therapy (6%).

We screened the relapse samples for hotspot mutations in a selected panel of genes (FLT3, WT1, KIT, N-RAS, K-RAS, NPM1) relevant for AML and found one or more mutations in 139 out of 240 patients (57.9%). Gene mutations were mutually exclusive in 73 out of 139 patients carrying mutations (52.5%), while in 66 out of 139 patients (47.5%) two or more gene mutations were observed. FLT3/ITD mutations coincided with 50% of NPM1 and 53% of WT1 exon 7 mutated cases (p=.003 and p<.0001 respectively). The frequencies of mutations at first relapse are summarized in the Table below including their impact on event free and overall survival, according to uni-variate analysis. In multivariate analyses, we included mutations and other variables (e.g. FAB type, WBC) with a uni-variate P value below 0.2 to exclude confounding factors. From these analyses, three independent factors significantly increased the risk of a second relapse (RFS after first relapse diagnosis); WT1 mutations (HR=11.2, P<.0001), WT1 single nucleotide variants (HR 3.4, P=.003) and FAB type M7 (HR= 2.2, P=.027). Different independent factors were associated with dismal overall survival after first relapse including FLT3/ITD mutations (HR=2.2, P=.028) and high WBC (HR=3.2, P=.004) at first relapse.

In conclusion, mutations in the studied panel of genes in this large cohort of pediatric relapsed AML patients were frequent. Overall, we detected receptor tyrosine kinase mutations in more than 25% of the patients, indicating the large proportion of relapsed patients that is eligible for targeted receptor tyrosine kinase inhibitor treatment. The gene mutations studied here provide strong prognostic factors for dismal outcome after first relapse. The relevance of these markers in predicting outcome should be validated in a prospective setting. Accurate risk group stratification of relapsed AML patients should be based on clinical and molecular characterization performed at first relapse. In addition, mutations in drugable genes can identify patients that are eligible for personalized, targeted re-induction strategies.