Clinical Study Of Prognostic Variables In 89 Childhood Acute Myeloid Leukemia

Background and Aims:To study the rationship between factors before and after treatment (biological factors、genetical factors、early treatment response) and the prognosis of acute myeloid leukemia.Methods:Retrospectively study eighty-nine acute myeloid leukemia children newly diagnosed and treated at least one course from January 1,2011 to December 31,2014,except acute promyelocytic leukemia.Pre-treatment factors--biological factors(gender、age、 white blood cell、platelet、hemoglobin、enlarged liver、enlarged spleen、 MDS、CSNL、risk group) and genetic factors (AML1-ETO、FAB、 cytogenetics) are analysed; pro-treatment factors--early treatment response(bone marrow blast cell accout at 48h after first induction course、 no respons to first induction、no response to second induction minimum residual disease positive after first induction、minimum residual disease positive after second induction)are analysed.The probabitity of event free survival was estimated by Kaplan-Meier analysis and the distribution of EFS were compared using log-rank test; Chi-square analysis or Fisher exact test was used to compare difference in distribution of presenting biological factors; A cox proportion hazard model was used to indentify independent prognostic factors.Results:①:Univariate analysis indicate that bone marrow blast cell accout at 48h after first induction course、male、high risk are associated with low risk complete remission (P<0.05).②Univariate analysis indicate that high risk、M7、no response to first induction、no response to second induction、minimum residual disease positive after first induction、 minimum residual disease positive after second induction are associated with low event free survival (P<0.05).③Multivariate analysis suggest that no response to first induction、no response to second induction minimum residual disease positive after second induction WBC≥100x10^9/L are independent unfavourable factors.Conclusions:Combining risk factors both pre-treatment and pro-treatment can predict prognosis, to provide a basis for treatment.