Related Experiment Video
Updated: Jun 30, 2026
Establishing a Competing Risk Regression Nomogram Model for Survival Data
Published on: October 23, 2020
Development and external validation of a multivariable nomogram for predicting severe immune checkpoint
Tao Luan1,2,3,4,5, Kangjing Ma1, Shuaiying Wang2
1Department of Pulmonary and Critical Care Medicine, The First Affiliated Hospital of Guangzhou Medical University, National Center for Respiratory Medicine, National Clinical Research Center for Respiratory Disease, State Key Laboratory of Respiratory Disease, Guangzhou Institute of Respiratory Health, Guangzhou, China.
Background:
Immune checkpoint inhibitors (ICIs) have improved outcomes in advanced lung cancer, but immune checkpoint inhibitor-associated myocarditis (ICIM), particularly severe ICIM, remains a rare yet potentially fulminant and fatal immune-related adverse event (irAE). We aimed to develop and externally validate a clinically accessible risk-prediction model based on routine baseline cardiac assessments to enable early identification and risk stratification of patients at high risk of severe ICIM around ICI initiation.
Methods:
This multicenter retrospective study included a training set, an internal validation set, and an external validation set. Patients with early-stage lung cancer, no ICI exposure, incomplete baseline cardiac evaluation [electrocardiogram (ECG)/echocardiography], missing predictor data, or insufficient follow-up were excluded; pre-existing cardiovascular comorbidities were not exclusionary and were recorded as baseline characteristics. The primary endpoint was severe ICIM [Common Terminology Criteria for Adverse Events (CTCAE) v5.0 grade ≥3]. Candidate predictors [high-sensitivity troponin, B-type natriuretic peptide (BNP), ECG, and echocardiography] were obtained from baseline/routine assessments at ICI initiation, prior to any ICIM event. Model Discrimination was assessed using receiver operating characteristic curve/area under the curve (ROC/AUC) with calibration plots and decision curve analysis (DCA) evaluating calibration and clinical utility.
Results:
A total of 834 patients were enrolled, with an overall mean age of 60 years [standard deviation (SD) =14]; by cohort, mean ages were approximately 59 years (training set), 61 years (internal validation set), and 61 (external validation set) years. Males accounted for 79.5% of the total cohort, with a slightly higher proportion in the internal validation set (83.2%) compared to the training (75.5%) and external validation (81.4%) sets. The prevalence of smoking was 61.0% and comparable across cohorts; 52.3% had comorbidities. Adenocarcinoma was the most common subtype (39.8%), followed by squamous cell carcinoma (35.4%), with similar distributions across cohorts. Based on these cohorts, four readily available predictors (troponin, BNP, ECG, echocardiography) were retained to construct the severe ICIM model and nomogram. The AUCs were 0.926 [95% confidence interval (CI): 0.880-0.971] in the training set, 0.885 (95% CI: 0.742-1.000) in the internal validation set, and 0.949 (95% CI: 0.884-1.000) in the external validation set; calibration plots and DCA demonstrated favorable agreement between predicted and actual outcomes, as well as significant net benefit across clinically relevant thresholds.
Conclusions:
This four-predictor severe ICIM risk model and nomogram, developed from multicenter real-world data and externally validated, supports early risk stratification and monitoring strategies for advanced lung cancer patients receiving ICIs. Prospective studies are warranted to confirm its generalizability and optimize implementation.
