Acute promyelocytic leukaemia (APL) is characterized by t(15;17)(q22;q21), resulting in a PML-RARA fusion that is the master driver of APL. A few cases that can't be identified with PML-RARA by using routine methods (karyotype analysis, FISH, and RT-PCR) involve abnormal promyelocytes that are fully in accordance with APL in morphology, cytochemistry, and immunophenotype. To explore the mechanisms involved in pathogenesis and recurrence of morphologically diagnosed APL, we performed comprehensive variant analysis from structural variants (SVs) to single-nucleotide variants (SNVs) in paediatric patients mophologically diagnosed as APL by next generation sequencing.SV analysis using targeted gene sequencing(TGS) in 120 DNA samples from both diagnosis and relapse stage identified 95 samples with RARA rearrangement (including 94 with PML-RARA and one with NPM-RARA) and two samples with KMT2A rearrangement. In the cases failed to be identified with any RARA rearrangement, transcriptome sequencing was applied in the available 13 RNA samples at diagnosis. One case each with CPSF6-RARG, NPM1-CCDC28A, and TBC1D15-RAB21 and two cases with a TBL1XR1-RARB fusion were discovered, which strongly suggested their contributions to leukemogenesis as driver alternations and APL phenotype is inextricably linked to rearrangement of RARA gene.SNV analysis in 75 primary APL samples with RARA rearrangement showed recurrent alternations in FLT3(27%), WT1(13%), USP9X(9%), NRAS(7%), and ARID1A(4%), with a strong potential for involvement in pathogenesis, and WT1as the only gene recurrently mutated in 4/5 samples(3/4 cases) in relapsed APL. WT1 (17.6%), NPM1 (11.8%), NRAS (11.8%), FLT3 (11.8%), and NSD1 (11.8%) were identified as recurrently mutated in 17 primary samples without RARA rearrangement and WT1(44.4%), NPM1 (33.3%), TP53 (33.3%), and RARA (33.3%)as recurrently mutated in 9 relapsed samples.The retrospective study showed the survival rates of APL without RARA rearrangement(8y-EFS=38.7%±17.8%) were much worse than APL with RARA rearrangement(8y-EFS=90.9%±3.9%), which is similar to AML (non-M3) patients(8y-EFS=57.7%±3.0%).Thus,NGS is necessary in cases failed to be identified with RARA rearrangement by karyotype analysis, FISH, and RT-PCR, which can afford accurate diagnosis and treatment guidance. APL phenotype may arise by abnormities of other members of the nuclear receptor superfamily involved in retinoid signaling(RARB or RARG) or even by mechanisms distinct from the formation of aberrant retinoid receptors. Differences in genetic alternations and survival indicated that APL that cannot be identified as having a RARA rearrangement are more reasonably classified as a subclass of AML other than APL, and an AML treatment regimen or individualized treatment should be considered according to the genetic abnormalities.Figue1. (A)Mutational profile between APL with and without RARA rearrangement. (B)Comparison of estimated probability of EFS and OS among APL with and without RARA rearrangement and AML(non-M3) group . “Typical APL” means APL with RARA rearrangement and “Atypical APL” means APL without RARA rearrangement. DisclosuresNo relevant conflicts of interest to declare.