Figure 1.

Radiomics features identified by the LASSO regression method. (a) The binomial deviation curve of the omics model with the parameter λ. The vertical axis represented the binomial deviation, and the horizontal axis represented log (λ) values; the parameter λ was adjusted to identify the optimized feature subset. The left vertical dotted line indicated log (λ) values corresponding to the optimized λ, and the number represented the number of selected features. (b) The changes in the radiomics model with λ. (c) Radiomics features screened based on the CT model and the corresponding coefficients.