TRV-2026-2934Version 1 · Published

Saved · see the current record

Why this version exists

Certified into the record

The exact text we saved

This is the version exactly as it was saved, character for character. The SHA-256 hash below is a 64-character code worked out from that text. Run the same text through SHA-256 and you get the same code; change a single letter and the code comes out completely different. Each version also stores the previous version’s hash, which chains them together so no version can be swapped out quietly.

TRUVACE RECORD VERSION
record: TRV-2026-2934
version: 1
kind: certified
reason: Certified into the record
timestamp: 2026-11-15T12:01:52.506020Z
status: published
lens: g_space
sector: health
headline: Artificial intelligence-enhanced electrocardiography for detection and prediction of dilated cardiomyopathy
dek: Background and aims A major challenge in dilated cardiomyopathy (DCM) care is the identification and management of individuals at risk before disease manifestation. An artificial intelligence (AI) model applied to electrocardiography (ECG) was developed to detect and predict the echocardiographic DCM phenotype (left ventricular dilatation with systolic dysfunction) and was evaluated for its ability to stratify the risk of incident DCM in familial cohorts. Methods A deep learning model was trained to detect the e…
gain_title: In a multi-cohort study, an AI model reading ECGs diagnosed the dilated cardiomyopathy phenotype with AUCs of .87-.97 and predicted future disease about as well as echocardiography (C-index .80-.87 versus .80-.90).
problem_title: (none)
trace_subject: (none)
gain_reading: In a multi-cohort study, an AI model reading ECGs diagnosed the dilated cardiomyopathy phenotype with AUCs of .87-.97 and predicted future disease about as well as echocardiography (C-index .80-.87 versus .80-.90).
gain_evidence: The AI-ECG model achieved areas under the receiver operating characteristic curve of .87-.97 for diagnosing the prevalent DCM phenotype, outperforming traditional ECG parameters. | AI-ECG was comparable to echocardiography for predicting future DCM (C-index .80-.87 and .80-.90, respectively).
problem_reading: (none)
problem_evidence: (none)
quick_read: Researchers trained a deep learning model on more than 330,000 ECG-echocardiogram pairs from Beth Israel Deaconess Medical Center in Boston to spot the echocardiographic signature of dilated cardiomyopathy, a heart muscle disease marked by an enlarged left ventricle that pumps poorly. It diagnosed the current phenotype with areas under the curve of .87 to .97 across outside cohorts and beat models built on traditional ECG features.

The model also predicted who would develop the disease later, with accuracy comparable to echocardiography. In two groups of genetically predisposed relatives found through cascade screening, 5-year negative predictive values were 95.5% in Antwerp and 91.5% in Maastricht, so a low score largely ruled out near-term disease.

Positive predictive values were lower, at 42.4% and 39.4%. High-risk scores were linked to a 5.1 to 6.4-fold higher risk of incident DCM. The authors say the low-cost tool could reduce surveillance burden and allow earlier intervention.
limitation: Positive predictive values in the familial cohorts were only 42.4% and 39.4%, so many people flagged high-risk did not develop disease; the tool's strength is ruling out short- to mid-term disease.
tag: Evidence-backed gain
key_points: Training used more than 330,000 ECG-echocardiogram pairs from Beth Israel Deaconess Medical Center in Boston. | External validation came from the UK Biobank, Zhongshan Hospital and Antwerp University Hospital. | Diagnostic AUCs for the prevalent DCM phenotype ranged from .87 to .97, outperforming traditional ECG parameters. | For future DCM, AI-ECG had a C-index of .80-.87, against .80-.90 for echocardiography. | In the familial cohorts, Antwerp (n = 166) and Maastricht (n = 177) each had 18% DCM-positive individuals. | Five-year negative predictive values were 95.5% in Antwerp and 91.5% in Maastricht. | Positive predictive values were 42.4% and 39.4%, and high-risk scores carried a 5.1 to 6.4-fold greater risk of incident DCM.
rundown: The model was built to address a core problem in dilated cardiomyopathy care: finding and managing at-risk people before disease appears. It was trained to detect the echocardiographic phenotype (left ventricular dilatation with systolic dysfunction) and to predict its future onset from a standard ECG.

Performance was compared against models using traditional ECG features and echocardiographic traits. Across external cohorts, the AI-ECG model outperformed traditional ECG parameters for diagnosis and matched echocardiography for predicting incident disease.

In familial testing, genetically predisposed people were identified through cascade screening at cardiogenetics clinics in Antwerp (n = 166) and Maastricht (n = 177); 18% in each group had DCM. High negative predictive values mean the tool best supports excluding short- to mid-term disease, which the authors say could lighten surveillance for relatives.
sources:
- peer_reviewed | European Heart Journal | https://doi.org/10.1093/eurheartj/ehag917 | 2026-11-13
prev: 0000000000000000000000000000000000000000000000000000000000000000
SHA-256 hash
6bd5f858389418f4053b8038bacce4254efb20b7bb0d01662d6b24f41445f1a1
Previous version’s hash
0000000000000000000000000000000000000000000000000000000000000000
Check this record hasn’t been altered
How to check it yourself, without trusting this page

Download any version from /api/record/TRV-2026-2934 and hash its canonical field yourself, for example with shasum -a 256. You should get its content_hash. Each version’s text also ends with prev:and the hash of the version before it (version 1 points to 64 zeros). If anyone had changed a single character of any version after it was saved, those hashes wouldn’t line up.