Fig. 2From: Preoperative CT-based deep learning radiomics model to predict lymph node metastasis and patient prognosis in bladder cancer: a two-center studyFeatures were selected by LASSO regression model. a The 12 features with non-zero coefficients and the roles of each feature that contributed to the model are shown. b Tuning parameter (λ) selection in the LASSO model. c The coefficients have been plotted vs. log(λ)Back to article page