4179Feasibility and validation of routine CMR-based phenotyping for the prediction of heart failure admission or death in patients with systolic dysfunction
Bibliographic record
Abstract
Abstract Background Standardized patient phenotyping using cardiovascular magnetic resonance (CMR) imaging has been shown to be of clinical value for prediction of adverse events in patients with heart failure and reduced ejection fraction (HFrEF). Studies have validated the prognostic capacity of function (LV, RV and LA) and replacement fibrosis burden in patients with ischemic and non-ischemic cardiomyopathy. The translation and validation of routine CMR-based phenotyping into clinical practice has yet to be demonstrated in prospective studies. Purpose This study was designed to explore feasibility and prognostic value of routine CMR-based patient phenotyping in a high-volume clinical referral center for patients with HFrEF. Methods One thousand three hundred and ninety-three consecutive patients with chronic HFrEF were prospectively recruited between January 2015 and June 2018. Chronic HFrEF was defined by LVEF≤50% by CMR, with no recent (within 90 days) acute myocardial infarction or myocarditis diagnosis. Patients with congenital heart disease and those without LGE CMR protocol were excluded. All patients underwent standardized CMR protocols with multi-chamber volumetric analysis and regional myocardial fibrosis coding. Pharmacy, ECG, laboratory and patient reported data was used for statistical modelling. A minimum three-month follow-up was mandated to identify the composite clinical outcome of heart failure hospitalization or death. Results The cohort had a median age of 61 years with 23% being female. The median follow-up was 737 days with 146 patients (10.5%) experiencing the composite outcome. Numerous imaging and non-imaging variables were significantly different between patients with and without the composite outcome, including: median LVEF (32% vs 39%, p<0.0001), RVEF (46% vs 51% p<0.0001), LV mass (77g/m2 vs. 65g/m2, p<0.0001), digoxin (19% vs. 9%, p<0.0001) and diuretic (63% vs 41%, p<0.0001) use. Presence of replacement fibrosis (HR=2.09, p=0.001), particularly midwall striae (HR=2.01, p<0.0001), diffuse (HR=3.88, p<0.0001) and RV insertion site fibrosis (HR=1.54, p=0.022) patterns, were significantly associated with the combined endpoint. A stepwise multivariable model was applied using all eligible variables and resulted in robust accuracy for prediction of the combined outcome with a concordance index of 0.751 (Figure 1). Conclusions This study demonstrates the feasibility and prognostic value of automated patient phenotyping that captures patient reported data, imaging, and administrative data for risk prediction modelling in HFrEF. The incremental application of machine learning is being explored. Acknowledgement/Funding J White: Early Investigator Award (Heart and Stroke Foundation of Alberta), Calgary Health Trust
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 | 0.003 | 0.008 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.000 |
| Bibliometrics | 0.001 | 0.000 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.001 | 0.000 |
| Open science | 0.001 | 0.001 |
| Research integrity | 0.000 | 0.000 |
| Insufficient payload (model declined to judge) | 0.002 | 0.001 |
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".