Outcome
The median follow-up duration of all living patients was 9.2 years (range = 5.5–14.0 years). During this period, 20 patients relapsed. Of those 20 patients, 1 developed secondary myelodysplastic syndrome 13 years after the initial treatment for high-risk neuroblastoma. The 5-year EFS and overall survival (OS) rates from treatment initiation were 41.5% ± 7.7% and 56.1% ± 7.8%, respectively (Figure 1 ). The outcomes were not significantly different between the three institutions. Although 15 patients needed dose reduction of thiotepa and melphalan according to their Ccr results, the outcomes were not significantly different (Supporting Information Figure S1 ). The 5-year EFS and OS of patients who showed CR/VGPR/PR before HDC were 50.0% ± 8.6% and 67.6% ± 8.0%, respectively. The 5-year EFS rate of patients with a good remission status (CR/VGPR) before HDC was significantly superior compared to patients with a poor remission status (78.6% ± 11.0% vs. 22.2% ± 8.0%; P = 0.00041) (Figure 2A ). Similarly, the 5-year OS rate of patients with a good remission status (CR/VGPR) before HDC was significantly superior compared to patients with a poor remission status (92.9% ± 6.9% vs. 37.0% ± 9.3%; P = 0.00019).
Table 1 summarizes prognostic effects according to univariate analysis. Patients with bone metastasis showed a significantly lower 5-year EFS compared to patients without bone metastasis (32.3% ± 8.4% vs. 70.0% ± 14.5%; P = 0.021). Eight patients who underwent tandem SCT showed significantly poorer prognosis compared to those who didn’t undergo tandem SCT (P = 0.012). Interestingly, the 5-year EFS rate of MYCN -amplified high-risk neuroblastoma patients was significantly superior compared to MYCN -non-amplified high-risk neuroblastoma patients (60.9% ± 10.2% vs. 16.7% ± 8.8%; P = 0.000065) (Figure 2B ). Similarly, the 5-year OS rate ofMYCN -amplified high-risk neuroblastoma patients was significantly superior compared to MYCN -non-amplified high-risk neuroblastoma patients (73.9% ± 9.2% vs. 33.3% ± 11.1%; P = 0.00018) (Figure 2B ). Even when the outcome was analyzed in each group, which showed a good response to induction chemotherapy (CR/VGPR) and poor response, MYCN -amplification was associated with good prognoses (P = 0.094 in CR/VGPR patients, P = 0.019 in non-CR/VGPR patients) (Supporting Figure S2 ). Of note,MYCN -amplified high-risk neuroblastoma patients who showed good response to induction chemotherapy showed good long-term outcomes (5-year EFS = 81.8%). Among four significant prognostic factors (INRC before HDC, MYCN amplification, bone metastasis, and tandem SCT), backward stepwise selection eliminated bone marrow metastasis and tandem SCT in multivariate analysis. Finally, the strongest prognostic factor for EFS was MYCN amplification (hazard ratio = 0.29; 95% confidence interval = 0.12–0.66). Similarly, MYCN amplification and INRC before HDC were also chosen as significant prognostic factors for OS in multivariable analysis. Table 2 describes the details of multivariate analysis.