Fig. 1

Selection of cartilage morphology parameters features using LASSO binary logistic regression nomogram.Tuning parameter (λ) selection in the LASSO nomogram was used in the 10-fold cross-validation via minimum criteria and the 1 standard error of the minimum criteria (the 1-SE criteria). The binomial deviance was plotted versus log(λ). (A) Dotted vertical lines were drawn at the optimal λ values, the λ value of 0.0379, with log(λ)–3.272, was the chosen via minimum criteria at baseline. (B ) LASSO coefficient profiles of the 92 cartilage parameters. The dotted vertical line was plotted at the λ value of 0.0379, resulting in 6 parameters at baseline. (C) A λ value of 0.0427, with log (λ)–3.154, was chosen via the minimum criteria at 12 M. (D) LASSO coefficient profiles of the 92 cartilage parameters. The dotted vertical line was plotted at the λ value of 0.0427 via minimum criteria, resulting in 6 parameters at 12 M. (E) A λ value of 0.0384 with log (λ)–3.26 was chosen via minimum criteria at 24 M. (F) LASSO coefficient profiles of the 92 change cartilage parameters over 24 M.The dotted vertical line was plotted at the λ value of 0.1068 via minimum criteria, resulting in 5 parameters at 24 M