Birgitta Versluijs

139 Outcome predictors in Allo-LS in children Results are expressed as hazard ratio (HR) and corresponding 95% confidence interval (CI). Variables associated with a p-value of < 0.05 were considered statistically signifi- cant. Variables showing significance in the total Allo-LS group were also analyzed in the IPS and BOS subgroups. Probabilities of Allo-LS free survival were calculated by using the Kaplan Meier estimate, we used the 2-sided log-rank test for univariate comparison. For the endpoints of non-relapse mortality and treatment failure, we used cumulative- incidence estimates and p-values were calculated using Gray’s. We did the statistical ana- lyses using SPSS version 21 and R version 3.2.4, using the packages survival and cmprsk. Results 361 patients underwent allogeneic HCT during study period, of whom 53 developed an allo-LS (14.7 %). Patient characteristics are shown in Table 1. In the group of Allo-LS patients busulfan-based conditioning regimen was used more often (p=0.02), and there were statistically more children with inborn errors of metabolism and less with bone marrow failure syndromes (p=0.04). Thirty patients were diagnosed with IPS (57 %) after a median time of 58 (range 16-140) days following HCT, and 23 (43%) patients developed BOS after a median time of 109 (16-331) days following HCT. Presenting symptoms were moderate to severe, with 43(81%) patients needing supplemental oxygen, and 22(42%) patients requiring mechanical ventilatory support. In 43/53 patients we have data on RV status at time of respiratory deterioration. In 23/43 patients one or more Respiratory Vi- ruses (RV) were detected (mostly rhinovirus), in the majority (65%) this RV was already found on pre-HCT screening. Chest HRCT scan was done in 44 patients and abnormal in all. The median HRCT composite score was 22.5 (range 8-50), the median HRCT allo- score was 11 (range 0-19). We have evaluable PFT at diagnosis in 20 patients (27 patients too young, 6 patients too dyspneic to perform PFT), 17 with BOS and 3 with IPS. Median FEV1% was 44% (SD 14.5), median FVC% was 55% (SD 15.1), median FEV1/FVC was 0.86 (SD 0.18), median RV/TLC was 51% (SD 13.9), median TLCO was 55 % (SD 10.1) of predicted values. First line treatment was started after a median time of 3 days (range 0-22) after first symptoms in patients with IPS, and after a median time of 10 days (range 3-82) after first symptoms in patients with BOS. Of the 53 patients 46 received first line therapy ac- cording to protocol (MP-pulse therapy). Five of the earliest patients received prednisone 2 mg/kg/day instead of MP-pulse because of clinicians’ decision, one started with mo- noclonal antibody therapy because of concomitant severe GvHD (overall grade 3) in gut and skin and one received upfront fludarabine because of strong objections to (the side effects of ) steroid therapy. 8