42: Usage of long axial crystals for PET applications: the AX-PET demonstrator and beyond

Abstract

process applied to bootstrap data replicates was used to avoid underfitting/overfitting and to assess the final multivariable model. Results: Data of 247 patients were available (CONV:116, HYPO:131). 71/247 (28.7%) reported IPSS15end=1. The most predictive dosimetric tools (as assessed through ttest) were the absolute weekly delivered dose (DSHw and DVHw). DSHw and DVHw were alternatively inserted in the bootstrap variable selection flow, together with clinical risk factors. Due to the number of events, a logistic model containing 6 variables was accepted. The panel (a) of the figure displays the frequency of appearance of variables in the top 6 explanatory factors (with a cut-off on variable appearing at least 1/10 of times). On the basis of observed frequency a model including basal IPSS (median OR=1.22, median p=0.00001), use of antihypertensives (median OR=2.7, median p=0.01), absolute bladder surface receiving more than 10.5 Gy/week (s10.5w, median OR=1.16, median p=0.0001), and absolute bladder surface receiving more than 12.5 Gy/week (s12.5w, median OR=1.07, median p=0.005), was choosen. AUC of this model was 0.80 (panel (b) of the figure). Silmilar results were obtained when using weekly absolute DVH instead of DSH. Conclusions: Basal IPSS, use of hypertensive drugs, s10.5w/v10.5w and s12.5w/v12.5w are the main predictors of IPSS>=15 at the end of radiotherapy. Bootstrap variable selection technique gives the modeler more insight into the importance and stability of the different variables selected and allows development of more robust models.