Abstract 11595: Applying Machine Learning on Myocardial Deformation Parameters for Identifying Subjects in High Risk of Heart Failure: The Copenhagen City Heart Study
Bibliographic record
Abstract
Background Recently, speckle tracking echocardiography (STE) and tissue doppler imaging (TDI) has gained increasing traction as a non-invasive tissue characterisation method within cardiology. But until now many patterns from the strain and TDI curves remain uninvestigated. In this work we applied supervised machine learning (ML) to identify unknown pathophysiological drivers of heart failure (HF) in the general population. Methods A total of 2383 subjects from the general population (Mean age 55.9 years ± 17.3, male 42%) underwent STE, TDI, and clinical examination. We applied an ensemble decision tree and a logistic regression model to investigate unknown speckle tracking parameters and to predict the endpoint: Occurrence of HF within five years. The ML models were evaluated with 20-fold cross-validation to provide a clear split between training and validation data. Results The median follow up-time was 5.41 years (ICR 4.49 - 6.28). 88 subjects (3.7%) developed HF within 5 years.We identified 4 new echocardiographic characteristics that improved the prediction of HF: 1. Peak systolic strain rate (SRs), 2. Strain at AVC, 3. Accumulated systolic strain/HR, 4. TDI Peak LV Diastolic Acceleration(Figure 1 and Figure 2). The model combining both novel strain- and TDI parameters as well as conventional echo parameters and clinical features performed significantly better than a model based on clinical features alone (AUC echocardiographic and clinical parameters vs. clinical parameters = 0.88/0.84, p = 0.04). An ensemble decision tree model predicted the outcome reasonably well with a precision of 24% at a sensitivity of 50%. Conclusion Adding novel echocardiographic parameters to a clinical ML model for prediction of HF improved the prognostic performance significantly. We identified 4 new parameters relevant for prediction of HF and found that peak systolic strain rate (SRs) was the most important predictor.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.005 | 0.005 |
| Meta-epidemiology (narrow) | 0.001 | 0.000 |
| Meta-epidemiology (broad) | 0.001 | 0.001 |
| Bibliometrics | 0.001 | 0.001 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.001 | 0.000 |
| Open science | 0.001 | 0.001 |
| Research integrity | 0.001 | 0.001 |
| Insufficient payload (model declined to judge) | 0.001 | 0.000 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".