Fig. 3From: Deep learning radiomics-based prediction model of metachronous distant metastasis following curative resection for retroperitoneal leiomyosarcoma: a bicentric studyFeature selection for the development of the DLRS using the LASSO regression model with a vertical line generated at the log (λ) value by using ten-fold cross-validation (a, b); The five radiomics features and 16 deep learning features and their corresponding coefficients (c); DLRS, deep learning radiomics signature; LASSO, least absolute shrinkage and selection operatorBack to article page