fig8

Figure 8. Development and performance evaluation of the MRI-based radiomics signature. (A) Tuning parameter (λ) selection in the LASSO model used 10-fold cross-validation via minimum criteria. The optimal λ value of 0.096 with log (λ) = -2.339 was identified. (B) LASSO coefficient profiles of the 1,316 radiomic features. The dotted vertical line is drawn at the value selected using 10-fold cross-validation in (A), where optimal λ resulted in 7 nonzero coefficients. (C) Histogram showing the coefficients of the selected features in the radiomics signature. The radiomics score was calculated as a linear combination of the 7 selected features weighted by their respective coefficients. (D) ROC curve of the gene signature. (E) Waterfall plot for distribution of radiomics scores and m6A subtypes in all patients. LASSO: Least absolute shrinkage and selection operator; ROC: receiver operator characteristic curve; LoG: laplacian of gaussian; GLSZM: gray level size zone matrix; GLRLM: gray level run length matrix; GLCM: gray level cooccurence matrix; LRHGLE: long run high gray level emphasis; IDN: inverse difference normalized; LAHGLE: large area high gray level emphasis; IMC2: informal measure of correlation 2.