EUVIMEDEUVIMED entdecken
Crossref

Adaptive AI for Cardiovascular Event Adjudication: Cardiovascular Event Adjudication Across Different Definitions in the ODYSSEY OUTCOMES and EUCLID Trials

Sreekanth Vemulapalli iD, Karla Peña Guerra iD, Daniel Wojdyla iD, W Schuyler Jones iD, Kenneth W Mahaffey iD, Robert A Harrington iD, Philippe Gabriel Steg iD, Gregory G Schwartz iD, Manesh R Patel iD, Renato D Lopes iD, Ricardo Henao iD

Circulation · 2026

DOI: 10.1161/circulationaha.126.080072 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Circulation, 2026. Verfasst von Sreekanth Vemulapalli und 10 weiteren Autor:innen. Beteiligte Einrichtung: Duke Clinical Research Institute, Durham, NC (S.V., D.W., W.S.J., M.R.P., R.D.L., R.H.). Hauptthemen laut MEDLINE: Artificial Intelligence, Endpoint Determination, Cardiovascular Diseases. Die Arbeit wurde bisher 2-mal zitiert (Crossref) und nennt 13 Literaturangaben. Gefördert von: Duke University School of Medicine.

AbstractSchlagwörterBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

BACKGROUND: Clinical end point classification (CEC) is the gold standard for cardiovascular end point measurement in clinical trials but adds time and cost. We developed and validated an artificial intelligence (AI) algorithm (adaptive AI for CEC [ADAPT-CEC]) that adjudicates multiple cardiovascular end points and adapts to new definitions. METHODS: ADAPT-CEC was derived from myocardial infarction (MI), stroke, and heart failure from the ODYSSEY OUTCOMES (Evaluation of Cardiovascular Outcomes After an Acute Coronary Syndrome During Treatment With Alirocumab) trial and externally validated on MI, stroke, bleeding and cardiovascular (CV) death from the EUCLID (Examining Use of Ticagrelor in Peripheral Artery Disease) trial after adaptation with 20 EUCLID suspected events per end point. ADAPT-CEC was compared via F1 score with direct Generative Pretrained Transformer (GPT) 4o adjudication and a hybrid approach where the 30% of suspected events with the lowest AI prediction certainty used human adjudication. The EUCLID primary end point of CV death, MI, or stroke was re-estimated for all 3 adjudication strategies. RESULTS: Among 13 885 suspected EUCLID primary end point events, ADAPT-CEC, hybrid, and GPT 4o strategies correctly classified 86.4%, 95.6%, and 76.3% of all end points and 99.4%, 99.6%, and 99.8% of all non–end points compared with human adjudication, respectively. Hybrid adjudication F1 metrics were the highest (CV death, 0.94 [95% CI, 0.92–0.96]; MI, 0.80 [95% CI, 0.77–0.82]; stroke, 0.82 [95% CI, 0.78–0.86]; bleeding, 0.83 [95% CI, 0.82–0.85]). ADAPT-CEC F1 metrics were lower for CV death, MI, and stroke but similar to GPT 4o, while bleeding (0.78 [95% CI, 0.77–0.79]) was superior to GPT 4o. The EUCLID primary treatment effect was similar by human adjudication (hazard ratio [HR], 1.02 [95% CI, 0.93–1.13]); hybrid (HR, 1.04 [95% CI, 0.94–1.15]); ADAPT-CEC (HR, 0.98 [95% CI, 0.88–1.09]), and GPT 4.0 (HR, 1.06 [95% CI, 0.95–1.19]) adjudication. CONCLUSIONS: After brief adaptation, a single trial–derived AI algorithm can adjudicate similar (MI and stroke) and new end points (CV death and bleeding) in a second trial and replicate the EUCLID primary outcome treatment effect. A hybrid approach with humans adjudicating those suspected events with the lowest 30% of ADAPT-CEC prediction certainty was superior to ADAPT-CEC alone or GPT 4o alone and replicated the EUCLID primary outcome treatment effect. Prospective studies of adaptive AI adjudication are needed to determine future trial implementation.

Schlagwörter

MeSH-Schlagwörter (MEDLINE)

Artificial IntelligenceEndpoint DeterminationCardiovascular DiseasesHumansStrokeMyocardial InfarctionTicagrelorAlgorithmsAdaptive AlgorithmsHeart FailureTreatment OutcomeIntelligent Systems

Stichwörter der Autor:innen

artificial intelligenceclinical trialdeathhemorrhagemyocardial infarctionstroke

Substanzen

Ticagrelor

Autor:innen und Institutionen

  • Duke Clinical Research Institute, Durham, NC (S.V., D.W., W.S.J., M.R.P., R.D.L., R.H.).
  • Departments of Medicine (S.V., W.S.J., M.R.P., R.D.L.), Duke University School of Medicine, Durham, NC.
  • Department of Biostatistics and Bioinformatics, Duke University, Durham, NC (K.P.G., R.H.).
  • Stanford Center for Clinical Research, Stanford University School of Medicine, Palo Alto, CA (K.W.M.).
  • Weill Cornell Medicine, New York, NY (R.A.H.).
  • Université Paris-Cité, AP-HP and INSERMU_1148, Paris, France (P.G.S.).

ORCID-Profile: Sreekanth Vemulapalli, Karla Peña Guerra, Daniel Wojdyla, W Schuyler Jones, Kenneth W Mahaffey, Robert A Harrington, Philippe Gabriel Steg, Gregory G Schwartz, Manesh R Patel, Renato D Lopes, Ricardo Henao

Förderung

  • Duke University School of Medicine

Zitieren

APA

Vemulapalli, S., Peña Guerra, K., Wojdyla, D., Jones, W. S., Mahaffey, K. W., Harrington, R. A., Steg, P. G., Schwartz, G. G., Patel, M. R., Lopes, R. D., & Henao, R. (2026). Adaptive AI for Cardiovascular Event Adjudication: Cardiovascular Event Adjudication Across Different Definitions in the ODYSSEY OUTCOMES and EUCLID Trials. Circulation. https://doi.org/10.1161/circulationaha.126.080072

Vancouver

Vemulapalli S, Peña Guerra K, Wojdyla D, Jones WS, Mahaffey KW, Harrington RA, et al. Adaptive AI for Cardiovascular Event Adjudication: Cardiovascular Event Adjudication Across Different Definitions in the ODYSSEY OUTCOMES and EUCLID Trials. Circulation. 2026. doi:10.1161/circulationaha.126.080072

BibTeX
@article{Vemulapalli2026,
  author = {Vemulapalli, Sreekanth and Peña Guerra, Karla and Wojdyla, Daniel and Jones, W Schuyler and Mahaffey, Kenneth W and Harrington, Robert A and Steg, Philippe Gabriel and Schwartz, Gregory G and Patel, Manesh R and Lopes, Renato D and Henao, Ricardo},
  title = {Adaptive AI for Cardiovascular Event Adjudication: Cardiovascular Event Adjudication Across Different Definitions in the ODYSSEY OUTCOMES and EUCLID Trials},
  journal = {Circulation},
  year = {2026},
  doi = {10.1161/circulationaha.126.080072}
}
Kostenlos starten →