Clinical and Biological Features of Pediatric Acute Myeloid Leukemia with Primary Induction Failure in the Japanese Pediatric Leukemia/Lymphoma Study Group (JPLSG) AML-05 Study

INTRODUCTION: Treatment outcome of pediatric acute myeloid leukemia (AML) patients has improved over the past few decades due to intensive chemotherapy. Primary induction failure (PIF), defined as failure to achieve complete remission after the second induction chemotherapy, occurs in around 10% of children with newly diagnosed AML. Effective treatment strategy for these patients is not established and their poor prognoses have been reported elsewhere. Here we investigated the clinical and biological features of pediatric AML patients with PIF who were registered in the Japanese Pediatric Leukemia/Lymphoma Study Group (JPLSG) AML-05 Study (AML-05).

METHODS: We conducted a retrospective survey of patients who were registered in the AML-05. We analyzed probabilities of overall survival (OS) according to gender, age at diagnosis, white blood cell count (WBC) at diagnosis, FAB classification, etc. Genetic analysis of primary leukemic blasts was performed. We screened fusion genes and gene mutations as follows: RUNX1-RUNX1T1, CBFB-MYH11, NUP98-NSD1, DEK-NUP214, MLL-PTD, WT1, NPM1, and MLL-rearrangements, gene mutations of KIT and FLT3-ITD. Furthermore, we evaluated the clinical and biological background of these PIF patients to define their prognostic factors statistically. RESULTS: Among the 443 patients registered in the AML-05 between November 2006 and December 2010, 43 patients (9.7%) had PIF. The observation period from the date of PIF was 5-78 months (median, 20 months). The age at diagnosis was 0.1-17 years (median, 6.1 years). There was no significant difference in gender (28 males vs15 females). The 3-year OS rate was 19.0% (95% confidence interval 1.0%-38.0%). This was extremely poor compared to the patients who achieved complete remission (CR): 66.7% for high-, 73.2% for intermediate-, and 92.6% for low-risk groups, respectively. Disease recurrence is the major cause of death (n=19, 46.3%). The mean WBC at diagnosis of PIF patients was 96 x 10⁹/L: this is significantly higher than that of non-PIF patients (51 x 10⁹/L, P<0.01, Cox regression analysis). Morphologically, FAB classification M7 was more frequently detected in the PIF group compared with non- PIF group, although not statistically significant: 11 patients (25.6%) vs 43 patients (9.1%) (P=0.07). Unfavorable prognostic cytogenetics such as FLT3-ITD (n=16, 37.2%), NUP98-NSD1 (n=6, 15.4%), and complex karyotype (n=9, 20.9%) were detected more frequently in PIF patients, as well as MLL-PTD (n=7, 16.2%), c-KIT mutations (n=6, 14.0%), DEK-NUP214 (n=2, 4.7%), and monosomy 7 (n=6, 14%). There were no cases with WT1 or NPM1 mutations. Only 3 patients had core binding factor leukemia with RUNX1-RUNX1T1 or CBFB-MYH11 in the PIF patients. Thirty-seven PIF patients received stem cell transplantation (SCT). The patients without SCT did not achieve CR, and died of primary disease. The patients who had achieved CR before SCT showed significantly better OS (P=0.03). There was no significant difference in donor source and conditioning regimen. Younger age at diagnosis (<10 years of age, P=0.02) was associated with better outcome. High-dose cytarabine-containing re-induction chemotherapy seemed effective on univariate analysis, which indicates the type of re-induction chemotherapy to be prognostic. Finally, multivariate analysis showed only age at diagnosis (<10 years of age) to be a significant prognostic factor.

CONCLUSIONS: The prognosis of pediatric AML patients with PIF was extremely poor. In our survey, unfavorable molecular biological background of leukemia was associated with PIF. It is necessary to explore more effective treatment strategies based on leukemia biology for these patients in the future.