Deep learning to predict progression independent of relapse activity at a first demyelinating event.

Best threshold to predict PIRA When the different interval-specific cumulative probabilities of reaching a first PIRA event were individually assessed as potential classifiers through ROC curves, we found that the probability at the first interval was the most accurate one: the AUC was greater for the first interval (0.6940 [95% CI 0.61642 to 0.77164]) than for the others (AUC between 0.6454 and 0.6925), although the difference did not reach statistical significance ( P = 0.0676) ( Fig. 4 ).

also in 111,027 other papers