An externally validated clinical-laboratory nomogram for myocardial involvement in adult idiopathic-inflammatory-myopathy patients

Abstract

Objectives: This study aimed at identifying clinical and laboratory risk factors for myocardial involvement (MI) in idiopathic inflammatory myopathies (IIMs) patients as well as constructing a risk-predicted nomogram for prediction and early identification of MI.

Methods: An IIMs cohort in southeastern China was constructed, including 504 adult IIMs patients who met the inclusion and exclusion criteria, and were hospitalized at four divisions of the First Affiliated Hospital, Zhejiang University School of Medicine from January 1st 2018 to April 30st 2022. After dividing patients into the training cohort and the validation cohort, risk factors for MI were identified through least absolute shrinkage and selection operator regression and multivariate logistic regression. A risk-predicted nomogram was established and validated internally and externally for discrimination, calibration and practicability.

Results: In this cohort, 17.7% of patients developed MI and the survival was significantly inferior to that of IIMs patients without MI (P < 0.001). In the training cohort, age > 55 years old (P < 0.001), disease activity > 10 points (P < 0.001), interleukin-17A (IL-17A) > 7.5 pg/ml (P < 0.001), lactic dehydrogenase (LDH) > 425 U/L (P < 0.001), anti-mitochondrial antibodies (AMAs, P = 0.017), and anti-MDA5 antibody (P = 0.037) were significantly correlated with development of MI. A nomogram was established by including the above values to predict MI and was found efficient in discrimination, calibration, and practicability through internal and external validation.

Conclusion: This study developed and validated a nomogram model to predict the risk of MI in adult IIMs patients, which can benefit the prediction and early identification of MI as well as timely intervention in these patients.

Keywords: Anti-MDA5 antibody; Idiopathic inflammatory myopathy; Interleukin-17A; Myocardial involvement.

Fig. 1

Comparisons of serum cytokines in IIMs patients with or without MI A. Comparison of serum IL-2 between MI and non-MI groups. B. Comparison of serum IL-4 between MI and non-MI groups. C. Comparison of serum IL-6 between MI and non-MI groups. D. Comparison of serum IL-10 between MI and non-MI groups. E. Comparison of serum TNF-α between MI and non-MI groups. F. Comparison of serum IFN-γ between MI and non-MI groups. G. Comparison of serum IL-17A between MI and non-MI groups. IIMs: Idiopathic inflammatory myopathies; MI: Myocardial involvement; IL: Interleukin; TNF: Tumor necrosis factor; IFN: Interferon

Fig. 2

Selection of clinical or laboratory factors using logistic regression with the LASSO regularization. A. LASSO coefficient profiles of the clinical or laboratory factors. B. Tuning parameter (λ) selection in the LASSO model using 10-fold cross-testing via minimum criteria. LASSO: least absolute shrinkage and selection operator

Fig. 3

A risk-predicted nomogram for prediction of MI in IIMs patients. MI: Myocardial involvement; IIMs: Idiopathic inflammatory myopathies

Fig. 4

Internal validation of the nomogram. A. The ROC curve of the nomogram in training cohort; B. The ROC curve of the nomogram with 500 bootstrap resamples in the training cohort; C. The calibration curve of probability of MI for IIMs patients in training cohort. The x-axis and y-axis represented the nomogram-predicted probability and the actual probability of MI, respectively. Perfect prediction would correspond to the slope of black dotted line that made a 45° angle. The blue line represented the whole validation cohort (n = 350), and the red line indicated bias correction by bootstrap resamples (B = 500 repetitions), revealing the perfect agreement between MI and the risk prediction by the nomogram; D. DCA of the nomograms for development of MI in training cohorts. The x-axis represented the threshold probability, and the y-axis represents the net benefit calculated by adding the true positives and subtracting the false positives. The horizontal green line along the x-axis represented the assumption of IIMs patients without MI, whereas the diagonal red line represented the assumption that all IIMs patients developed MI at a specific threshold probability. The blue line represented the nomogram; E. CIC of the nomgram for the development of MI in the training cohort. The red line (i.e., number of high-risk patients) represented the number of patients classified as positive by the model at each threshold probability, meanwhile the blue curve (i.e., number of high-risk patients with an outcome) represented the number of patients with true positives at each threshold probability. ROC: Receiver operating curve characteristics; MI: Myocardial involvement; IIMs: Idiopathic inflammatory myopathies; DCA Decision curve analysis: CIC: Clinical impact curve

Fig. 5

External validation of the risk-predicted nomogram. A. The ROC curve of the nomogram in validation cohort; B. The ROC curve of the nomogram with 500 bootstrap resamples in the validation cohort; C. The calibration curve of probability of MI for IIMs patients in validation cohort. The x-axis and y-axis represented the nomogram-predicted probability and the actual probability of MI, respectively. Perfect prediction would correspond to the slope of black dotted line that made a 45° angle. The blue line represented the whole validation cohort (n = 154), and the red line indicated bias correction by bootstrap resamples (B = 500 repetitions), revealing the perfect agreement between MI and the risk prediction by the nomogram; D. DCA of the nomograms for development of MI in validation cohorts. The x-axis represented the threshold probability, and the y-axis represents the net benefit calculated by adding the true positives and subtracting the false positives. The horizontal green line along the x-axis represented the assumption of IIMs patients without MI, whereas the diagonal red line represented the assumption that all IIMs patients developed MI at a specific threshold probability. The blue line represented the nomogram; E. CIC of the nomgram for the development of MI in the validation cohort. The red line (i.e., number of high-risk patients) represented the number of patients classified as positive by the model at each threshold probability, meanwhile the blue curve (i.e., number of high-risk patients with an outcome) represented the number of patients with true positives at each threshold probability. ROC: Receiver operating curve characteristics; MI: Myocardial involvement; IIMs: Idiopathic inflammatory myopathies; DCA Decision curve analysis: CIC: Clinical impact curve