Development and validation of a predictive model of the impact of single nucleotide polymorphisms in the ICAM-1 gene on the risk of ischemic cardiomyopathy

Figures

• 

  Figure 1. Roadmap of screening and grouping the study cohorts.

• 

  Figure 2. Roadmap of establishing and validating the model.

• 

  Figure 3. Variables selected by performing the LASSO binary logistic regression model. (A) Tuning parameter (λ) selection in the LASSO model employed 10-fold cross-validation through minimum criteria. Moreover, the AUC curve was mapped vs. log(λ). With the minimum criteria and the 1-standard error of the minimum criteria (the 1-SE criteria), the dotted vertical lines were plotted. A λ value of 0.018 was selected (1-SE criteria) in accordance with 10-fold cross-validation. (B) LASSO coefficient profiles of the 20 features. Obviously, a coefficient profile plot was constructed against the log(λ) sequence. We also drew a vertical line at the value chosen based on 10-fold cross-validation, in which the optimal λ value led to 10 non-zero coefficients.

• 

  Figure 4. Nomogram for evaluating the risk of ICM. The nomogram was developed using several clinical and genomic variables, including age, smoking, diabetes, LDL-C, hemoglobin, NT-proBNP, ejection fraction, and rs5491 (figure generated by R software, https://www.r-project.org).

• 

  Figure 5. (A) Training group. (B) Validation group, and ROC to confirm the discrimination power of the nomogram.

• 

  Figure 6. Calibration plot (A) Training group. (B) Validation group of the nomogram (Ptraining group = 0.937, Pvalidation group = 0.910). The diagonal gold line indicates a good prediction by the ideal model. In addition, the diagonal 45° gold line represents a good calibration where the model's predictive ability has high accuracy in indicating the actual risk of ICM. The red line shows the performance of the nomogram, and a closer fit to the diagonal red line indicates a good predictive value.

• 

  Figure 7. Decision curve analysis for the nomogram. (A) Training group. (B) Validation group. Obviously, the threshold probability is shown by the x-axis. The threshold probability is the value at which the expected benefit of treatment can be equivalent to the value of avoiding treatment. The y-axis measures the net benefit, computed by subtracting the rate of all patients who are false positives from the proportion of those who are truly positives and weighted by the relative risk of forgoing treatment compared with the negative consequence of the unnecessary treatment.

• 

  Figure 8. Survival curve of patients undergoing ICM carrying the AA and AT genotypes in the SNP rs5491.

• 

  Figure 9. Diagram of ICAM-1 gene and location of rs5491.