AI-derived inflammatory-sarcopenic and metabolically dysregulated phenotypes were independently associated with impaired VO2 max (≤20 mL/kg/min) with ORs 2.67 and 2.29 respectively versus balanced phenotype in patients with heart failure.
Observational (n=505)
Yes
Does AI-driven clustering of multimodal clinical variables identify distinct phenotypes associated with impaired cardiorespiratory fitness in patients with heart failure?
AI-driven clustering of multimodal clinical data successfully identifies distinct heart failure phenotypes that strongly predict cardiorespiratory fitness impairment, offering a framework for early functional risk stratification.
Effect estimate: OR inflammatory-sarcopenic phenotype 2.67 (95% CI 1.35 to 5.27), OR metabolically dysregulated phenotype 2.29 (95% CI 1.30 to 4.03) vs balanced phenotype
p-value: p<0.01
Background Patients with heart failure (HF) frequently suffer from undetected declines in cardiorespiratory fitness (CRF), which significantly increases their risk of poor outcomes. However, current clinical practice lacks effective tools for early CRF risk stratification. Methods We conducted an artificial intelligence (AI)-driven unsupervised clustering analysis based on 15 multimodal clinical variables—including metabolic, inflammatory and body composition indicators—in 505 patients with HF. The associations between clustering-derived phenotypes and CRF impairment (maximal oxygen uptake (VO 2 max) ≤20 mL/kg/min) were evaluated using multivariable logistic regression and five supervised machine learning models. SHapley Additive exPlanations analysis was applied for model interpretability. External validation was performed in an independent cohort of 201 patients. Results Three distinct phenotypes were identified: balanced, inflammatory-sarcopenic and metabolically dysregulated. Compared with the balanced phenotype, both non-balanced phenotypes showed significantly higher odds of impaired VO ₂ max. In the derivation cohort test set, random forest (area under the curve (AUC)=0.75; 95% CI 0.62 to 0.87) and XGBoost (AUC=0.74; 95% CI 0.62 to 0.87) demonstrated the best discriminative performance. In the external validation cohort, the highest discrimination was observed for Naive Bayes (AUC=0.75; 95% CI 0.67 to 0.83), followed by random forest (AUC=0.74; 95% CI 0.58 to 0.91). Conclusion By integrating multimodal clinical data with AI-driven clustering and machine learning, this study identified novel CRF risk phenotypes in patients with HF and established a highly interpretable and generalisable risk stratification model. These findings offer a valuable framework for early functional assessment and pave the way for precision rehabilitation strategies in HF management.
Qiu et al. (Thu,) conducted a observational in Adult patients aged 18-90 years with heart failure diagnosed based on 2021 ESC guidelines, hospitalized with available cardiopulmonary exercise testing including VO2 max, body composition analysis, and relevant laboratory parameters, excluding acute decompensated HF or hemodynamic instability, malignancy, severe hepatic or renal dysfunction, active infection, missing key variables, pregnancy or lactation (n=505). AI-driven clustering and machine learning phenotyping for CRF risk assessment vs. Balanced phenotype and standard clinical variables was evaluated on Impaired cardiorespiratory fitness defined as maximal oxygen uptake (VO2 max) ≤20 mL/kg/min (OR inflammatory-sarcopenic phenotype 2.67 (95% CI 1.35 to 5.27), OR metabolically dysregulated phenotype 2.29 (95% CI 1.30 to 4.03) vs balanced phenotype, p=p<0.01). AI-derived inflammatory-sarcopenic and metabolically dysregulated phenotypes were independently associated with impaired VO2 max (≤20 mL/kg/min) with ORs 2.67 and 2.29 respectively versus balanced phenotype in patients with heart failure.