Synapse
⌘+K
Synapse
PulseExploreClubsResearchersJournals
Instagram
HomeClubsExplore
July 23, 2026The International Journal of Cardiovascular Imaging

Low-output/uncoupled HF phenotype linked to ~115% higher 6-month mortality vs preserved-output/coupled phenotype.

View Full Paper
Ask AI
Bookmark
Share

Why the study?

Does machine learning phenotyping based on Doppler-derived energetics (CPO, LVSWI, VAC) improve risk stratification for 6-month mortality in ADHF patients with LVEF < 40% compared to conventional echocardiography?

Population

500 patients with acute decompensated heart failure (ADHF) and LVEF < 40%

Comparison

Machine learning phenotyping and augmented… vs Traditional logistic model based on conventional…

Design

Cohort

Follow-up

6 months

Key result

A low-output/uncoupled heart failure phenotype identified by machine learning was associated with higher 6-month mortality compared to a preserved-output/coupled phenotype (HR 2.15; 95% CI 1.08-4.30).

Authors

KRKapil RajendranAAAju AjayAAArun Jude Alphonse

Discussion

Loading...

Member takes

Overview

May stratify short-term HF mortality risk via ML; leaves open prospective validation before clinical use.

Key Points

  • To evaluate if machine learning phenotyping based on cardiac energetics improves risk stratification in acute decompensated heart failure.
  • Enrolled 500 ADHF patients with LVEF < 40%, assessing cardiac power output, stroke work index, and VAC.
  • Applied K-means clustering and compared phenotype outcomes using Kaplan–Meier curves and Cox regression.
  • Developed and validated a random forest model for 6-month mortality prediction.
  • Low-output phenotype had 21.4% mortality versus 11.8% in preserved-output (log-rank p = 0.03; HR 2.15, 95% CI 1.08–4.30).
  • Augmented logistic model AUC was 0.76, outperforming the traditional model with AUC 0.67.
  • Random forest classifier achieved AUC of 0.81, identifying LVSWI as the strongest predictor.

Study Design

Type

Cohort (n=500)

Structured PICO

Does machine learning phenotyping based on Doppler-derived energetics (CPO, LVSWI, VAC) improve risk stratification for 6-month mortality in ADHF patients with LVEF < 40% compared to conventional echocardiography?

P
Population
500 patients with acute decompensated heart failure and LVEF < 40%, followed for 6 months.
E
Exposure
Machine learning phenotyping and augmented logistic model incorporating Doppler-derived cardiac energetics (left-ventricular cardiac power output [CPO], stroke-work index [LVSWI], and ventriculo-arterial coupling [VAC])
C
Comparator
Traditional logistic model based on conventional echocardiography (EF, TAPSE, LVEDP, RVSP, RAP)
O
Outcome
6-month mortalityhard clinical

Main Result

Hazard Ratio: 2.15 (95% CI 1.08–4.3)

Absolute Event Rate: 21.4% vs 11.8%

p-value: p=0.03

Integrating Doppler-derived cardiac energetics (CPO, LVSWI, VAC) into routine echocardiography using machine learning significantly improves 6-month mortality risk stratification in patients with acute decompensated heart failure and reduced ejection fraction.

Cite This Study

Rajendran et al. (2026) conducted a cohort in Acute decompensated heart failure (ADHF) (n=500). Low-output/uncoupled phenotype (identified by machine learning) vs. Preserved-output/coupled phenotype was evaluated on 6-month mortality (HR 2.15, 95% CI 1.08-4.30, p=0.03). A low-output/uncoupled heart failure phenotype identified by machine learning was associated with higher 6-month mortality compared to a preserved-output/coupled phenotype (HR 2.15; 95% CI 1.08-4.30).

synapsesocial.com/papers/6a61b246faa9903c5116b549https://doi.org/10.1007/s10554-026-03777-8
View Full Paper
Ask AI
Bookmark
Share

Also Consider

Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context:

  1. 1Congestion–Perfusion Phenotypes and In-Hospital Mortality in Acute Heart Failure: Phenotype-Specific Prognostic Markers and Right-Heart Involvement2026
  2. 2Phenotypic stratification and prognostic value of cardiac magnetic resonance in non-dilated left ventricular cardiomyopathy2025 · 3 citations
  3. 3LightGBM-Based Classification of Heart Failure Phenotypes Using Morpho-Energy Features from High-Resolution ECG2026
  4. 4Differentiating HFmr/rEF from HFpEF using standard 12-lead ECG measurements: an interpretable machine learning study2026
  5. 5Predicting Left Ventricular Ejection Fraction Recovery After Percutaneous Coronary Intervention in Patients With Chronic Coronary Syndrome by Using Interpretable Machine Learning Models: Retrospective Study2025