MATERIALS AND METHODS
We retrospectively analyzed patients with refractory or relapsed NB treated at the GianninaGaslini Institute in Genoa (Italy) from 1996 to 2014. Those who had undergone treatment with Th-131I-MIBG followed closely by ASCT after HD-MAT with Busulfan and Melphalan after induction therapy were included in the analysis, while those who had received Th-131I-MIBG alone or in combination with a different schedule of HD-MAT were excluded.
Figure 1 summarizes treatment schedules and disease response evaluations.
All patients or their guardians signed a consent form allowing the use of their data for clinical research purposes. We followed the procedures in accordance with our institution’s ethical standards and according to Italian guidelines. The Liguria Regional Ethics committee (ComitatoEticoRegionale – Regione Liguria – IRCCS AOU San Martino – IST) approved the collection for retrospective studies (360REG2014) of data from medical records of patients admitted to our department.
All data were collected from our institutional database and retrospectively analyzed. The main clinical characteristics of patients and the types of treatment are summarized in Table 1 .
Criteria for diagnosis, staging and inclusion in the study
The diagnosis of NB was established by histological examination of tumor specimens or by bone marrow infiltration and/or elevated urinary excretion of catecholamine metabolites. Disease extension on diagnosis and response to treatment were evaluated by means of imaging (ultrasonography/CT scan/MRI) of the primary tumor,123I-MIBG scintigraphy and study of bone marrow on at least 2 aspirates and 2 trephine core biopsies. In the absence of MIBG uptake, skeletal involvement was evaluated by means of technetium 99m-MDP scintigraphy or, more recently, 18F-FDG PET/CT. However, in this specific case, the patient was not deemed eligible for the study. All patients were staged according to INSS criteria [10] . MYCN gene amplification was defined as a copy number of 10 or more.
To be included in the study, patients had to have a poorly responding HR-NB (MIBG-avid) after induction followed by further chemotherapy or, alternatively, a residual HR-NB (MIBG-avid) after a salvage treatment for relapse.
Treatments: Induction therapy, PBSC harvest and Th-131I-MIBG
Induction therapy was administered according to the Italian NB97 protocol [14] before 2002 and the European SIOPEN NBAR-01 protocol [15] in subsequent years. Patients with refractory disease after induction underwent salvage therapy with additional courses of chemotherapy (ICE and/or TVD and/or TEMIRI ).
Peripheral blood stem cells (PBSC) were collected only if bone marrow evaluation (at least 2 aspirates and 2 biopsies from different sites) demonstrated complete remission; collection was performed after stimulation with Granulocyte Colony-Stimulating Factor (G-CSF), administered subcutaneously at a dosage of 10 μg/kg/day. The minimum CD34+ cell-dose required in order to consider collection to be adequate was 3 x 106/kg of the recipient’s weight (optimal dose ≥ 4 x 106/kg).
Before 131I-MIBG administration, patients underwent thyroid blockade according to EANM procedure guidelines[16] and showed normal organ function and a value of WBC > 2000/uL and of platelets > 75000/uL. They were then admitted to the Nuclear Medicine Department, where Th131I-MIBG was administered in a single intravenous infusion over 2 hours (median dose 8.5mCi/Kg [IQR 6.5-12]). All patients remained in radiation protective isolation for 5-7 days after131I-MIBG administration.
High-dose BU-Mel and autologous stem cell transplantation
High-dose chemotherapy consisted of Busulfan, administered in 16 doses from day -7 to day -3 before ASCT; administration was oral until 2011 (cumulative dose 16 mg/kg) and intravenous in the following years (cumulative dose according to weight range: <9 Kg: 16 mg/Kg; 9-16 Kg: 19.2 mg/Kg; 16-23 Kg: 17.6 mg/Kg; 23-34 Kg: 15.2 mg/Kg; >34 Kg: 12.8 mg/Kg ). Melphalan was administered on day -1 in a single intravenous dose of 140 mg/m2. On day 0, patients underwent ASCT.
Disease response criteria and toxicity
Disease response was evaluated at the end of the induction phase, before and after Th-131I-MIBG and HD-BuMel, at the end of treatment and at the last follow-up examination.
Considering the disease response evaluated before Th-131I-MIBG and HD-MAT BuMel, according to the International Neuroblastoma Response Criteria (INRC) [10] , we defined 3 groups of response: i) good response (complete response [CR], very good partial response [VGPR]); ii)partial response (PR); iii) poor response (no response [NR], stable disease [SD] and progression of disease [PD]).
Toxicity was graded according to Common Terminology Criteria version 4.03 [11] , on considering mucosal, gastrointestinal, pulmonary, hepatic and renal toxicities and veno-occlusive disease/sinusoidal obstruction syndrome (VOD/SOS) according to the Baltimore [12] and Seattle criteria [13] . Only proven bacterial bloodstream infection or clinical sepsis and deep mycosis were regarded as significant infective complications. Engraftment was defined as the first of 3 consecutive days on which neutrophil and platelet counts reached 500/mm3 and 50,000/mm3, respectively.
Statistical analysis
Descriptive statistics are reported in terms of absolute frequencies and percentages for continuous data), and Pearson’s chi-square test was used to compare proportions.
The overall survival (OS) probability and the cumulative risk of relapse (CRR) were calculated by means of the Kaplan-Meier method, considering time from diagnosis and from therapy (Th-131I-MIBG and HD-MAT BuMel) to death from any cause (for OS) or to relapse (for CRR); if neither event occurred, data were censored on the date of the last follow-up examination. Differences among subgroups were assessed by means of the Log Rank test. Cox modeling was adopted for multivariate analysis, considering MYCN amplification, induction therapy, MIBG score after induction, HD-CT with Thiotepa and disease status before Th-131I-MIBG as variables.
Analyses were performed by means of STATA software (version 14.2, Stata Corp., College Station, TX, USA). Two-tailed probabilities are reported; a p-value of 0.05 was adopted to define nominal statistical significance.