The long-term outcome of 1390 children with acute lymphoblastic leukemia (ALL), treated in two successive clinical trials (Taiwan Pediatric Oncology Group (TPOG)-ALL-97 and TPOG-ALL-2002) between 1997 and 2007, is reported. The event-free survival improved significantly (P ¼ 0.0004) over this period, 69.3±1.9% in 1997-2001 to 77.4±1.7% in 2002-2007. A randomized trial in TPOG-97 testing L-asparaginase versus epidoxorubicin in combination with vincristine and prednisolone for remission induction in standard-risk (SR; low-risk) patients yielded similar outcomes. Another randomized trial, in TPOG-2002, showed that for SR patients, two reinduction courses did not improve long-term outcome over one course. Decreasing use of prophylactic cranial irradiation in the period 1997-2008 was not associated with increased rates of CNS relapse, prompting complete omission of prophylactic cranial irradiation from TPOG protocols, beginning in 2009. Decreased use of etoposide and cranial irradiation likely contributed to the low incidence of second cancers. High-risk B-lineage ALL, T-cell, CD10 negativity, t(9;22), infant, and higher leukocyte count were consistently adverse factors, whereas hyperdiploidy 450 was a consistently favorable factor. Higher leukocyte count and t(9;22) retained prognostic significance in both TPOG-97 and TPOG-2002 by multivariate analysis. Although long-term outcome in TPOG clinical trials is comparable with results being reported worldwide, the persistent strength of certain prognostic variables and the lower frequencies of favorable outcome predictors, such as ETV6-RUNX1 and hyperdiploidy 450, in Taiwanese children warrant renewed effort to cure a higher proportion of patients while preserving their quality of life.

This study used finite element (FE) analysis with the load-controlled method (LCM) and the displacement-controlled method (DCM) to examine motion differences at the implant level and adjacent levels between fusion and non-fusion implants. A validated three-dimensional intact (INT) L1-L5 FE model was used. At the L3-L4 level, the INT model was modified to surgery models, including the artificial disc replacement (ADR) of ProDisc II, and the anterior lumbar interbody fusion (ALIF) cage with pedicle screw fixation. The LCM imposed 10 Nm moments of four physiological motions and a 150 N preload at the top of L1. The DCM process was in accordance with the hybrid testing protocol. The average percentage changes in the range of motion (ROM) for whole non-operated levels were used to predict adjacent level effects (ALE%). At the implant level, the ALIF model showed similar stability with both control methods. The ADR model using the LCM had a higher ROM than the model using the DCM, especially in extension and torsion. At the adjacent levels, the ALIF model increased ALE% (at least 17 per cent) using the DCM compared with the LCM. The ADR model had an ALE% close to that of the INT model, using the LCM (average within 6 per cent), while the ALE% decreased when using the DCM. The study suggests that both control methods can be adopted to predict the fusion model at the implant level, and similar stabilization characteristics can be found. The LCM will emphasize the effects of the non-fusion implants. The DCM was more clinically relevant in evaluating the fusion model at the adjacent levels. In conclusion, both the LCM and the DCM should be considered in numerical simulations to obtain more realistic data in spinal implant biomechanics.