To systematically review and evaluate the performance of prediction models for postoperative atrial fibrillation (POAF) in lung cancer patients, highlighting its clinical significance.
Key Findings:
Six studies were included, primarily using logistic regression for model development.
Common predictors included age, sex, cardiovascular comorbidities, and surgical factors.
Reported AUC values ranged from 0.72 to 0.89, with a pooled AUC of 0.79 (95% CI: 0.71–0.87), indicating good overall discrimination but varying clinical relevance.
Substantial heterogeneity was observed (I2 = 98.7%), but subgroup analysis with consistent outcome definitions showed reduced heterogeneity.
All studies had a high overall risk of bias, which limits their reliability.
Interpretation:
Current POAF prediction models for lung cancer patients demonstrate acceptable discriminative ability but are limited by methodological weaknesses, such as small sample sizes and lack of external validation.
Limitations:
High risk of bias in all included studies, affecting the reliability of findings.
Lack of external validation for the prediction models, limiting generalizability.
Methodological weaknesses restrict clinical applicability and necessitate further research.
Conclusion:
While the existing POAF prediction models show promise, their clinical utility is hampered by significant methodological limitations, underscoring the need for improved study designs.
