OBJECTIVETo investigate the clinical value of clearance of leukemic cell during induction of remission therapy in children with precursor B cell acute lymphoblastic leukemia (BCP-ALL), and to assess the applicative value of different indexes.

METHODFrom April 2005 to April 2008, 206 children with de novo BCP-ALL were admitted. We firstly analyzed the effect of clearance of leukemic cells during induction of remission therapy on relapse-free survival (RFS). Four indexes were used to assess the clearance of leukemic cells including prednisone response on day 8 (d8-PR), percentage of lymphoblast in bone marrow on day 22 (d22-BM) and day 33 (d33-BM), and bone marrow (BM) minimal residual disease (MRD) detection on day 33 (d33-MRD). Then the sensitivity, specificity, positive predictive value and negative predictive value of the four indexes to assess their ability to predict relapse were analyzed. Finally, the consistency between two of the four indexes to explore the relationships among them were analyzed.

RESULTThere were significant differences between RFS of the sub-groups divided according to d8-PR, d22-BM, d33-BM, d33-MRD (P < 0.01); Cox proportional hazard model analysis showed that d33-MRD ≥ 10(-3) and positive BCR/ABL fusion gene were the independent prognostic factors. Sensitivity of d33-MRD was higher than that of morphology detection (d22-BM, d33-BM and d8-PR) in prediction of relapse, and positive predictive value of morphology detection was higher than that of d33-MRD. Sensitivity could be greatly increased by combination with clinical and biological characteristics. Consistency could not be found between d8-PR and d22-BM, d33-BM, d33-MRD, as well as between d22-BM, d33-BM, and d33-MRD. However, all cases of d22-BM, d33-BM M2/M3 were d33-MRD ≥ 10(-3), while the same phenomenon could not be found for patients with poor d8-PR.

CONCLUSIONClearance of leukemic cell during induction of remission therapy in children with BCP-ALL had important clinical value. Sensitivity of MRD detection after induction of remission therapy was higher than that of morphological analysis to predict relapse. Morphological analysis could only identify a few patients with very high risk of relapse and the sensitivity could be increased by combination with clinical biological characteristics. The simple prednisone response may contain some prognostic information that could not be covered by analysis of BM cells. It may be the best way to assess the clearance of leukemic cells to combine the prednisone response with MRD detection after induction of remission therapy.