A CT-Based Radiomics Approach for the Differential Diagnosis of Sarcomatoid and Clear Cell Renal Cell Carcinoma

Figure 3

Radiomics feature selection by using the LASSO regression method. The optimal was selected using a tenfold crossvalidation via the minimum of average mean square error. To the CMP features, and (a) and to the NP features, and (c). (b) and (d), respectively, showed the coefficient profiles along the full path of possible values in the CMP and the NP feature selection. In addition, dashed vertical lines were drawn at the optimal based on the minimum of average mean square error in (a–d).

(a)

(b)

(c)

(d)