Chen C, Zhou J, Yu H, et al. - Researchers used both random survival forest (RSF) and non-negative matrix factorization (NMF) analyses to assess the crucial predictors for all-cause mortality of acquired long QT syndrome (aLQTS) patients. They initially entered clinical features and manually recorded electrocardiographic (ECG) parameters into the RSF model. Subsequently, they entered latent variables, detected using NMF, into the RSF as additional variables. In this study with 327 aLQTS patients, 16 predictive factors with positive variable importance values were identified by the RSF model. In this patient population with aLQTS, all-cause mortality was shown to be predicted by cancer, potassium and calcium as well as by ECG indicators including JTc and QRS. A significantly improved mortality prediction was afforded by the present RSF-NMF model.