A Diagnostic Paradigm Shift in Acute Myocardial Infarction

Abstract

Background

The current ST-segment elevation myocardial infarction (STEMI) and non-ST-segment elevation myocardial infarction (NSTEMI) paradigm have long been the cornerstone of myocardial infarction (MI) care but fail to identify many patients with acute coronary occlusion (ACO), delaying treatment and worsening outcomes.

Objectives

The new occlusive MI (OMI) and nonocclusive MI (NOMI) paradigm addresses these limitations by shifting from rigid electrocardiography (ECG)-driven thresholds to a pathophysiology-based strategy. It integrates clinical judgment, ECG subtleties, supportive artificial intelligence, biomarkers, and imaging when needed to actively detect ACO. The current trial investigates whether this approach improves outcomes.

Methods

The DIFOCCULT-3 (Time for a DIagnostic paradigm shift from ST-elevation/non-ST-elevation to OCClUsion/non-occLusion myocardial infarcTion? trial is a multicenter, modified cluster-randomized, open-label study designed to compare the OMI/NOMI paradigm with the STEMI/NSTEMI approach in 6,000 participants presenting with suspected MI. The OMI/NOMI approach integrates clinical gestalt, artificial intelligence–supported ECG analysis, and adjunctive diagnostic tools to actively identify ACO. The primary endpoint is the composite of all-cause mortality and all-cause rehospitalization at 1-year follow-up. Secondary endpoints include infarct size, left ventricular ejection fraction, wall motion score index, and accurate diagnosis of ACO.

Results

Enrollment and follow-up are ongoing; acute outcomes such as earlier identification of ACO and infarct size will be reported after enrollment, and late outcomes including mortality and rehospitalization will be available after 1-year follow-up.

Conclusions

The DIFOCCULT-3 study aims to determine whether the OMI/NOMI paradigm represents a significant advancement in MI care by improving outcomes through earlier and more accurate identification of ACO than the STEMI/NSTEMI framework (Time for a Diagnostic Paradigm Shift From STEMI/​NSTEMI to OMI/​NOMI [DIFOCCULT-3]; NCT06570759)

Central Illustration

Acute coronary occlusion (ACO) with insufficient collateral circulation places the patient at immediate risk of myocardial infarction (MI) unless emergent reperfusion is performed via thrombolysis or percutaneous coronary intervention (PCI). Currently, ST-segment elevation serves as the primary electrocardiographic (ECG) marker guiding emergent coronary revascularization decisions.^1,2 However, the existing ST-segment elevation myocardial infarction (STEMI) and non-ST-segment elevation myocardial infarction (NSTEMI) paradigm is increasingly recognized as suboptimal.3, 4, 5, 6

Emerging evidence suggests that current STEMI criteria fail to identify approximately one-third of patients with ACO, resulting in delayed or missed treatment opportunities.7, 8, 9, 10, 11, 12, 13, 14 This diagnostic gap stems from several limitations inherent to the STEMI/NSTEMI paradigm. First, the STEMI criteria were not explicitly designed to detect ACO; rather, they were established based on empirical thresholds without direct validation against ACO.15, 16, 17 Second, the emphasis on ST-segment with predefined cutoffs¹⁸ results in a low sensitivity^6,14,17,19 and often overlooks subtle ECG findings associated with ACO.^20,21 Third, ACO can occur in the absence of STE, but the term “STEMI” restricts the diagnosis to one small part (STE) of one test (the ECG).^22,23 These limitations collectively hinder accurate identification and timely management of ACO and result in worse outcomes in patients labeled as NSTEMI under the current guidelines.24, 25, 26

In response, a paradigm shift has been proposed,3, 4, 5^,22,27,28 advocating for the adoption of the occlusive myocardial infarction (OMI) and nonocclusive myocardial infarction (NOMI) framework. The OMI/NOMI paradigm integrates clinical assessment, ECG interpretation, and adjunct diagnostic modalities to actively seek evidence of ACO. This approach addresses the conceptual barrier posed by the “STEMI” label, which often delays intervention for NSTEMI patients despite guideline recommendations to expedite treatment for some high-risk subsets.²⁹ The OMI/NOMI paradigm aims to overcome these limitations by redefining diagnostic criteria and treatment strategies to prioritize timely intervention.

Recent studies have shown that patients with NSTEMI later identified as having ACO (NSTEMI but OMI) experience significant delays in catheterization and face adverse outcomes comparable to,^26,30 or even worse than,³¹ those of STEMI patients. The most direct comparison of the OMI/NOMI and STEMI/NSTEMI paradigms in an unselected emergency department population comes from the DIFOCCULT (DIagnostic accuracy oF electrocardiogram for acute coronary OCClUsion resuLTing in myocardial infarction) 1 study,¹² which provided retrospective evidence that the OMI/NOMI paradigm outperforms the STEMI/NSTEMI approach in diagnosing ACO and predicting long-term mortality.

Despite this evidence, it has yet to be determined whether adopting the OMI/NOMI paradigm in clinical practice, particularly through early intervention in OMI patients who do not meet STEMI criteria, translates into improved outcomes. The DIFOCCULT-3 trial aims to address this critical question, evaluating whether the OMI/NOMI paradigm results in better clinical outcomes than the STEMI/NSTEMI approach.

Materials and methods

Study design

The DIFOCCULT-3 Trial (NCT06570759) is a multicenter, modified cluster-randomized, open-label study designed to compare the outcomes of the new OMI/NOMI paradigm with the established STEMI/NSTEMI paradigm (Central Illustration). The trial aims to enroll at least 6,000 participants across 18 centers in Turkey, with enrollment expected to be completed by October 2025, follow-up concluding by October 2026, and full trial results available by December 2026.

Central Illustration.

The Evolution of MI Paradigm and Design of DIFOCCULT-3 Study

The evolution of MI management reflects not only advances in diagnostic tools and treatment options but also a deeper understanding of underlying pathophysiology. The old Q-wave vs non-Q-wave MI paradigm focused on diagnosing irreversible damage retrospectively with limited intervention. The STEMI/NSTEMI paradigm introduced urgent intervention but relied on rigid ECG cutoffs, using ST-segment elevation as a surrogate for ACO. This still leaves up to 25% to 30% of patients undetected and treated passively, as in the older paradigm. The new OMI/NOMI paradigm shifts focus directly to the underlying pathology, combining clinical gestalt, subtle ECG signs, AI-supported interpretation, and adjunctive imaging to proactively detect ACO earlier. The DIFOCCULT-3 trial compares these paradigms using cluster randomization across 18 centers and 6,000 patients, evaluating both acute outcomes (infarct size, time to intervention) and long-term clinical endpoints (mortality and rehospitalization). ACO = acute coronary occlusion; AI = artificial intelligence; Dx = diagnosis; ECG = electrocardiographic; MI = myocardial infarction; NOMI = nonocclusive MI; NSTEMI = non-ST-segment elevation myocardial infarction; OMI = occlusive MI; STEMI = ST-segment elevation myocardial infarction.

Study population and recruitment

Patients aged 18 years or older who are admitted to the emergency department with a clinical presentation consistent with acute coronary syndrome are screened for enrollment in the study. Exclusion criteria include rejection or withdrawal of consent, active pregnancy, or the administration of intravenous thrombolytics. All patients diagnosed as MI based on clinical, ECG, or troponin findings are enrolled in the study (Figure 1). Patients ruled out for MI using the European Society of Cardiology 0- to 2-hour algorithm¹ are excluded from the screening phase and are not included in the study. If a patient is enrolled based on a clinical or ECG diagnosis, they remain in the study even if subsequent troponin levels are negative (ie, not MI). However, if a patient is enrolled solely on the basis of troponin positivity, but the troponin elevation is later attributed to an alternative diagnosis, and angiography with the intent of PCI is aborted, that patient is excluded from the study.

Figure 1.

Patient Inclusion and MI Diagnostic Flow

This figure summarizes the inclusion criteria and diagnostic flow for patients in the DIFOCCULT-3 trial. Patients presenting with suspected acute coronary syndrome are screened for MI based on clinical presentation, ECG findings (clinician-only or AI-supported, according to the study arm), and high-sensitivity cardiac troponin levels. Patients ruled out for MI using the ESC 0- to 2-hour algorithm are excluded. Enrolled patients are categorized according to real-time diagnostic criteria and undergo follow-up for confirmation of final MI diagnosis. AI = artificial intelligence; CAG = coronary angiography; ECG = electrocardiogram; ESC = European Society of Cardiology; iv = intravenous; MI = myocardial infarction; Tn = high-sensitivity cardiac troponin.

Randomization and formation of the study arms

To directly compare the OMI/NOMI and STEMI/NSTEMI paradigms, we designed the study with 2 parallel diagnostic strategy arms. However, randomizing patients individually between these 2 approaches was not feasible in practice because interventionalists trained in the OMI paradigm might still recognize ACO in a patient assigned to the STEMI/NSTEMI arm whose ECG does not meet strict STEMI criteria. In such situations, they would be unlikely to disregard an OMI diagnosis and withhold timely reperfusion, which could unintentionally introduce significant cross-contamination between study arms. To address this, we adopted a modified cluster randomization design in which interventionalist teams, rather than individual patients, are randomized to either the STEMI/NSTEMI or OMI/NOMI approach for each duty day.

In each center, interventional cardiologists participating in the study were divided into 2 groups in a 1:1 ratio, STEMI/NSTEMI and OMI/NOMI, ensuring both have comparable levels of interventional experience (eg, years of training, angiography volume, and primary PCI counts in the preceding year). After these groups were formed, patients are randomized into the STEMI/NSTEMI or OMI/NOMI cohorts based on the team on duty, meaning the approach followed at each center on a given day will be determined by the assigned team. Duty schedules are determined randomly adhering to institutional policies. If some interventionalists at a particular center are not willing to participate in the study, patients presenting on nonparticipating interventionalists' duty days will be excluded, even if the intervention is performed by a participating interventionalist. This modification ensures balance in expertise, minimizes cross-contamination between study arms, and upholds ethical considerations by preventing trained OMI/NOMI interventionalists from withholding emergent reperfusion. Patients are assigned to treatment arms based on the interventionalist team on duty at the time of presentation, maintaining a pragmatic yet controlled study environment.

First responder pathways

All potential first responders within the networks of the participating centers, those who initially contact the patient and acquire the first ECG (referring physicians, emergency physicians, or cardiologists, depending on the center), were also informed about the study and provided with a dedicated artificial intelligence (AI)-powered smartphone application for ECG diagnosis.

On STEMI/NSTEMI days, all first responders and interventionalists are encouraged to follow the current guidelines^1,2 and use the established cutoffs for the diagnosis of STEMI: 1) new ST-segment elevation at the J-point in 2 contiguous leads with the cut-point: ≥1 mm in all leads other than leads V₂–V₃ where the following cut-points apply: ≥2 mm in men ≥40 years; ≥2.5 mm in men <40 years, or ≥1.5 mm in women regardless of age.¹⁸ The AI-powered smartphone application for ECG diagnosis is turned off and is not available to the first responders associated with that center. If first responders can independently identify OMI in patients who do not meet STEMI criteria, they can advocate for urgent intervention by the interventionalists. This approach allows the study to capture a snapshot of current real-life practices while ensuring physicians are not ethically compelled to adhere strictly to the STEMI/NSTEMI paradigm.

On OMI/NOMI days, physicians are encouraged to actively search for ACO, regardless of whether STEMI criteria are present on the initial ECG. A diagnosis of OMI can be based on clinical gestalt, ECG findings, and adjunct modalities.

Clinical gestalt includes hallmark presentations such as almost pathognomonic chest pain and ischemic arrhythmias, hemodynamic instability, or cardiac arrest following typical symptoms. Investigators in the OMI/NOMI arm are permitted to proceed directly to urgent catheterization when, in their clinical judgment, the presentation is highly suggestive of ACO due to these factors and when there is no alternative explanation.

ECG diagnosis, whether interpreted by physicians or aided by an AI-powered smartphone application, incorporates static or serial changes for ACO using the DIFOCCULT-1 study algorithm.^5,12,20 On OMI/NOMI days, the smartphone application is activated and available to all first responders associated with this center. This application assists diagnosis, but the final decision is left to the interventionalist on duty.

Adjunct modalities include bedside echocardiography demonstrating new or presumed new wall motion abnormalities in patients with ongoing or recurrent chest pain and significantly elevated initial troponin levels. For high-sensitive cardiac troponin (hs-cTn) T, it has been shown that a level exceeding 1,000 ng mL^-1 is highly specific for major epicardial coronary artery occlusion.³² Similarly, a hs-cTn I >200 times the upper limit of normal (ULN; e.g., Architect, Abbott Diagnostics: 5,000 ng/L; ADVIA Centaur, Siemens Healthcare: 5,000 ng/L; Access, Beckman Coulter: 2,400 ng/L) is defined as a marker for OMI in patients with ongoing or fluctuating chest pain. In patients diagnosed with OMI, immediate catheterization laboratory activation with the intent to perform PCI is pursued. In NOMI patients, initial medical stabilization is prioritized, followed by elective catheterization per the NSTEMI pathway unless high-risk features are identified.

AI-powered smartphone application

All first responders were equipped with a dedicated AI-powered smartphone application (Powerful Medical, Samorin, Slovakia), trained by expert ECG interpreters and validated in a large cohort.³³ The application was provided 1 month prior to the first patient enrollment, and each responder's log-in status was verified through an online system. First responders were linked exclusively to a single participating center, ensuring no overlap between centers.

The application’s functionality varied based on the study arm determined by the team on duty. On OMI/NOMI days, the AI application is fully activated and accessible to all first responders associated with that center. When a user captures a photo of an ECG, the application digitalizes the image, interprets the data, and displays one of 2 messages: “OMI” or “Not-OMI.” First responders were instructed to promptly inform the interventionalist on duty for potential catheterization laboratory activation if result shows “OMI.” On STEMI/NSTEMI days, the AI-supported application is deactivated for that center. If a first responder attempts to capture a photo of an ECG, a warning message is displayed: “We are now following the standard STEMI/NSTEMI approach. Please continue your usual practice.”

A commercial version of the same smartphone application by the same company is also available on the market. During the study, if a network address is detected accessing both the commercial and study-specific applications, the commercial version is deactivated by the company, and a notification mail is sent explaining that the commercial smartphone application will not be available to users in Türkiye for the duration of the study. In addition, all ECGs stored in the study database will be cross-referenced with the commercial smartphone application’s ECG history. If any matches are identified, the corresponding patient will be excluded from the study.

Data collection

Study data are collected and managed using Research Electronic Data Capture (REDCap) tool hosted at a dedicated server.^34,35 REDCap is a secure, web-based software platform designed to support data capture for research studies, providing: 1) an intuitive interface for validated data capture; 2) audit trails for tracking data manipulation and export procedures; 3) automated export procedures for seamless data downloads to common statistical packages; and 4) procedures for data integration and interoperability with external sources. The researchers were instructed that the entry of collected data into the REDCap system within 3 days of acquisition is mandatory. A data monitoring board ensures the completeness, integrity, and accuracy of the entries, providing feedback to the data entry team and requesting explanations or modifications as needed. Collected baseline variables and their definitions are listed in the REDCap printout in the Supplemental Appendix Personal data are stored in encrypted form.

Interventions

Patients are managed in accordance with current guidelines.^1,2 All patients are followed up for 48 to 72 hours post-PCI, during which serial ECGs and troponin measurements are conducted. Troponin levels are measured at admission, hourly if needed for diagnostic purposes, every 6 hours until they peak following an MI diagnosis, and then daily. The peak troponin level within the 24- to 72-hour period (typically at 48 hours) is used as a surrogate marker for infarct size. The study centers use 4 different troponin kits, one hs-cTn T (Roche Elecsys, Roche Diagnostics) and three hs-cTn I kits (Architect, Abbott Diagnostics; ADVIA Centaur, Siemens Healthcare; and Access, Beckman Coulter). Comparisons of infarct size will be conducted using z-scores, calculated based on the mean and standard deviation.

Before discharge, an echocardiogram is performed to evaluate the wall motion score index (WMSI) and left ventricular ejection fraction (LVEF), following standard conventions.³⁶

Coronary angiography and interventions are performed following standard conventions. Each angiogram is reviewed by 2 interventionalists. In the event of a discrepancy between their interpretations, a third interpreter is consulted. The following points are assessed to determine the presence of an ACO: 1) the TIMI flow level in the infarct-related artery; 2) lesion characteristics, including degree of stenosis, irregular lesion borders, angiographic thrombus, contrast staining, the presence of well-developed collaterals to the distal vessel, and the ease of guidewire crossing if total occlusion. If necessary, the primary operator is contacted for clarification.

Definition of diagnostic groups for final analyses

As the study evaluates the STEMI/NSTEMI and OMI/NOMI paradigms, 4 diagnostic groups will naturally emerge based on final diagnoses. These groups will be analyzed using both intention-to-treat and per-protocol approaches. In intention-to-treat analyses, classification will reflect real-world clinical decision-making at enrollment. In per-protocol analyses, strict STEMI criteria and expert interpretation will be applied to ensure consistency in subgroup definitions.

Specifically, in intention-to-treat analyses, STEMI/NSTEMI status will be defined based on real-world clinical practices observed during the study, which may deviate from strict STEMI criteria. In the STEMI/NSTEMI arm, this labeling is performed at the time of enrollment based on investigators' clinical judgment. In the OMI/NOMI arm, all ECGs will be randomly allocated to STEMI/NSTEMI investigators with essential clinical information for retrospective evaluation, determining how these patients would have been classified under the STEMI/NSTEMI paradigm.

Similarly, in the OMI/NOMI arm, the classification is performed during enrollment using clinical gestalt, AI-supported ECG analysis, and adjunct diagnostic tools. In the STEMI/NSTEMI arm, all ECGs will be retrospectively reviewed by OMI/NOMI investigators using AI to determine how these patients would have been classified under the OMI/NOMI paradigm. The OMI diagnosis also includes clinical variables, such as clinical gestalt and very high first troponin levels. However, clinical gestalt cannot be acted upon retrospectively (eg, bedside echocardiography or serial ECGs). Nevertheless, if a patient is recorded with ongoing chest pain and a very high first troponin level (based on center-specific and troponin kit-specific values), this will be included in the OMI (+) NSTEMI subgroup in the STEMI/NSTEMI arm, even if the ECG is not interpreted as OMI by the AI-powered application.

For per-protocol analyses, expert interpretation and strict STEMI criteria will be applied for ECG classification to ensure consistency and rigor in subgroup evaluations. This combination of cross-comparison between paradigms and the evaluation of real-world clinical practice vs expert-evaluated strict guideline definitions will provide valuable insights into the effectiveness and practicality of both diagnostic paradigms in clinical settings.

ACO adjudication

Due to the dynamic nature of infarct-related artery occlusion, where the artery may spontaneously recanalize before angiography or a total occlusion may represent a chronic condition, an adjudication process for ACO was deemed necessary. We defined ACO as meeting any of the following criteria: 1) TIMI flow grade 0 to 1 or TIMI flow grade 2 if the lesion demonstrates acute culprit features (>90% stenosis, thrombus, ulceration, or contrast staining) will be considered ACO only in patients with clear evidence of troponin elevation consistent with MI (a peak troponin level ≥5 times the ULN plus at least a 20% rise within the first 24 hours);¹ or 2) highly elevated troponin levels with a peak troponin T level >1,000 ng/mL or a troponin I level >200 times the average ULN³² without an obvious alternative diagnosis or with supporting evidence, such as classical ECG evolution or a culprit-looking lesion on angiography in a coronary territory consistent with ECG or new or presumably new wall motion abnormality on echocardiography in a consistent coronary territory; or 3) cardiac arrest occurring before troponin levels have been documented with any supporting clinical evidence of possible ACO.

Expert ECG review

All ECGs will be reviewed and coded by 2 experts blinded to clinical outcomes but provided with brief clinical information about the patient, including age, sex, medical history, and the presence and duration of pain at the time of ECG acquisition. In case of a discrepancy between their interpretations, a third expert will be consulted. Interobserver variability will be calculated, and a final consensus diagnosis will be established. The diagnostic accuracy of ECG findings, as determined by expert opinion and the AI-powered application, will then be evaluated against the composite ACO endpoint.

Patients with ECGs clearly indicative of STEMI, without alternative or confounding diagnoses, will form a calibration cohort. This group includes patients who receive identical treatment in both arms and will serve to evaluate the consistency of intervention practices between the STEMI/NSTEMI and OMI/NOMI approaches. Although interventionists in both arms were chosen to have comparable expertise in coronary intervention, this cohort provides an additional benchmark to ensure consistency. Any differences in outcomes within this similarly treated group, attributable to interventionist practices rather than treatment protocols, will be identified and calibrated using the calculated factor from this cohort, if necessary.

Study outcomes

The primary composite endpoint is all-cause mortality and all-cause re-hospitalization during follow-up across the entire cohort. However, as the major difference in outcomes is expected in patients who are diagnosed as NSTEMI under the standard STEMI/NSTEMI paradigm but as OMI under the new OMI/NOMI paradigm, the study adopts a dual primary analysis design, placing equal emphasis on a predefined evaluation of the OMI (+) NSTEMI.

The primary source of outcome data will be the Turkish national electronic database (e-nabız), which provides comprehensive, real-time updates on all deaths and hospitalizations nationwide. To ensure the completeness and accuracy of data, direct phone contact with participants or their families will be conducted as a secondary measure. All collected outcomes will be reviewed by an independent outcome adjudication board blinded to the study arms.

The secondary comparisons will be done for the adjudicated ACO, the infarct size as defined by 24- to 72-hour peak troponin, WMSI, LVEF, and in-hospital cardiopulmonary resuscitation, intubation, and mortality. Outcomes will be analyzed both with intention-to-treat and per protocol approaches.

Acute outcome analyses will be reported after completion of enrollment to inform procedural and diagnostic performance, while long-term clinical outcome analyses will be reported separately after completion of the full 1-year follow-up.

Statistical plan

For the overall cohort, we estimated that enrolling 3,185 participants would provide 95% statistical power to detect a hazard ratio (HR) of 0.87, corresponding to a 13% relative risk reduction in the combined primary endpoint for the OMI/NOMI approach compared to the STEMI/NSTEMI approach.^12,25,37,38 This estimate was derived using the standard Schoenfeld approximation for Cox proportional hazards models for time-to-event endpoints. To account for the hierarchical nature of our study design and the structured nature of PCI-based STEMI/NSTEMI treatment pathways, we selected an intracluster correlation coefficient of 0.015 and an average cluster size of 50 patients per interventionalist team.³⁹ Applying a design effect correction of 1.74, this adjustment increased our required sample size from 3,185 to 5,526 participants. To maintain feasibility and account for potential dropouts, the final target enrollment was rounded to 6,000 patients.

For the OMI (+) NSTEMI subgroup analysis, assuming a HR between 0.60 and 0.80, the required sample size before clustering ranges from 65 to 174 participants.^12,25,40 To account for clustering effects, we applied the same intracluster correlation coefficient and design effect correction, increasing the required sample size to 96 to 303 participants. Based on the expected 1:1.5 to 2 STEMI/NSTEMI ratio and a 30% OMI prevalence in the NSTEMI cohort, this corresponds to a total of fewer than 2,000 patients. Since both primary analyses are embedded within the same trial framework, the final target enrollment of 6,000 patients remains sufficient, ensuring robust statistical power across varying effect sizes.

To assess the diagnostic accuracy of the STEMI/NSTEMI vs OMI/NOMI classification paradigms, a sample size of 2,351 participants would provide 95% statistical power to detect a 5% difference in the area under the receiver operating characteristic curve (from 0.750 to 0.700) in identifying ACO, using the standard Hanley & McNeil method for area under the receiver operating characteristic curve sample size estimation.¹² This sample size is also expected to provide sufficient power to detect at least a 10% relative improvement in infarct size, LVEF, and WMSI among STEMI (−) OMI (+) patients undergoing early revascularization in the OMI arm compared to those receiving standard-timing revascularization in the STEMI/NSTEMI arm, assuming a 30% OMI prevalence in the NSTEMI cohort.

The primary analysis will utilize time-to-event methods, including Kaplan-Meier survival curves for the composite primary endpoint. Comparisons between groups will be performed using the log-rank test, and Cox proportional hazards regression will be used for multivariable modeling. Baseline characteristics with a P value of ≤0.05 in univariate analysis will be considered for inclusion, with a stepwise procedure applied to select the final covariates. A pathophysiological model including baseline Global Registry of Acute Coronary Syndrome risk score,⁴¹ pain to balloon time, and discharge treatment will also be incorporated. Participants will be censored if they are lost to follow-up, withdraw consent, or do not experience the composite endpoint by the end of the 1-year follow-up period. Censored data will be incorporated into Kaplan-Meier survival curves, with participants contributing to the risk set until the time of censoring. Cox proportional hazards regression will account for censored data, enabling unbiased estimation of HRs.

To address potential variability in outcomes due to interventionist or center-related factors, we will incorporate a random-effects (frailty) term into the Cox model. The interventionalist team is the unit of randomization and intervention, while the primary outcome is measured at the patient level. The calibration cohort will be used to estimate variability attributable to interventionist practices, and the random-effect variance calculated from this cohort will inform the frailty term in the full Cox model. The final model will adjust for patient-level covariates, include random effects for interventionalist teams or centers, and apply calibration adjustments based on the calibration cohort to ensure that differences in outcomes due to interventionist-related variability are properly accounted for.

We anticipate minimal missing data for the composite endpoint, as death and hospitalization events are covered in the electronic national database (e-nabız). For any missing baseline covariates, multiple imputation by chained equations will be used, with predictive mean matching for continuous variables, and logistic or multinomial regression for categorical variables. Imputed datasets will be combined using Rubin’s rules.

Other predefined secondary analyses include: 1) comparison of OMI/NOMI and STEMI/NSTEMI approaches in in-hospital mortality and infarct size by peak troponin, LVEF, and WMSI; 2) a receiver operating characteristics curve analysis for OMI components for ACO; 3) time-related benefit analysis in all ECG and MI subtypes; 4) the degree of the adoption of OMI approach in the early and late phase of the study.

Baseline characteristics will be summarized using standard descriptive statistics. Statistical analyses will be performed using SPSS (version 29.0.2.0, SPSS Inc).

Safety monitoring and reporting

Study REDCap forms necessitate in-hospital adverse events to be actively collected to monitor and report any in-hospital adverse events. An independent safety monitoring board has been established to oversee the safety and progress of the trial. The board convened via teleconference during the pretrial period, upon enrollment of 25% of the participant sample size, and will continue to meet after each subsequent 25% enrollment milestone. The primary objective of the board is to monitor enrollment milestones and the safety of the interventions. A three-point combined safety endpoint will be monitored: 1) in-hospital cardiopulmonary resuscitation; 2) in-hospital intubation; and 3) in-hospital mortality. If a statistically significant increase in the three-point combined safety endpoint is observed in the OMI/NOMI arm, after adjusting for the baseline Global Registry of Acute Coronary Syndrome risk score, during the enrollment of any 25% of the participant sample size, the safety monitoring board will make a recommendation regarding the potential discontinuation of the study.

Study integrity

The study is an investigator-initiated trial conducted under the auspices of the Turkish Society of Cardiology. The Turkish Society of Cardiology supports the investigator team in developing the trial design and organizing the participating centers. The steering committee oversees the processes of recruitment, consent and assent, follow-up, and ensures the validity and integrity of data acquisition. The trial has been approved by the Ethical Board of Marmara University (09.2021.523). Any change in protocol or centers will be addressed by this board. The study will be conducted in accordance with Good Clinical Practice guidelines.

We acknowledge that the implementation of the new OMI/NOMI paradigm, even in a randomized controlled trial, may pose challenges due to conceptual, logistical, and individual or institution-level barriers. To promote adherence, all interventionists in the OMI/NOMI arm, we participated in a comprehensive training program, including lectures and case simulations, focusing on the OMI/NOMI framework, subtle ECG findings, and the integration of AI-powered diagnostic tools prior to the study’s commencement. In addition, the REDCap system was designed to issue warning messages if patient management deviated from the OMI protocol. During the trial, the data-monitoring board oversees adherence, conducts regular audits, and provides feedback on protocol compliance. Any deviations are meticulously documented and analyzed to identify barriers, with adaptive feedback mechanisms employed to address these issues promptly. Investigators also receive ongoing case reviews and targeted discussions to reinforce protocol fidelity. Furthermore, per-protocol and sensitivity analyses will complement the primary intention-to-treat analysis to evaluate the impact of protocol adherence on outcomes. These measures collectively ensure the robustness of the study findings while offering valuable insights into the feasibility of implementing the OMI/NOMI paradigm in real-world clinical practice.

Trial status

Trial enrollment commenced on October 1, 2024, with 5,665 subjects, which have been enrolled by July 1, 2025. This equates to 94.4% of the trial’s samples size, which is ahead of the trial’s enrollment timeline (Figure 2).

Figure 2.

Study Timeline and the Current Status of Patient Enrollment

Enrollment began on October 1, 2024 and has reached 94.4% of the target sample size (N = 5,665) as of July 1, 2025. The solid blue line represents actual cumulative enrollment, and the dashed orange line indicates the expected recruitment pace. The sidebar highlights the proportion of target enrollment completed to date. Final enrollment is expected to close shortly, with follow-up extending through 1 year for primary clinical outcomes.

Discussion

Since its inception,⁴² OMI concept has aimed to redefine the diagnosis and management of MI, advocating for a more pathophysiological approach. Despite mounting evidence supporting its potential benefits over the years, its widespread adoption has encountered several obstacles, following a process reminiscent of Kübler-Ross's stages of reaction to challenge: denial, resistance, and bargaining, with acceptance yet to be achieved.⁴³

Denial phase arose from the perception that no ACOs were being missed. Patients without ST-elevation on their initial ECG but later found to have an occlusion during angiography were not considered to have “missed” diagnoses, as they were classified as NSTEMIs by definition. Moreover, if the artery spontaneously reperfused by the time of angiography, these cases simply went unnoticed as NSTEMI with worse outcomes, further obscuring the need for a paradigm shift. This created a ‘false-negative paradox’, where clinicians accepted the delayed identification of non-STEMI occlusions without questioning its consequences.^4,23 However, studies have shown that non-STEMI occlusions have nearly double the mortality of their nonocclusive NSTEMI counterparts.^12,24, 25, 26

The resistance phase has been understandably prolonged due to the fact that the STEMI paradigm has been the cornerstone of MI care for 35 years, deeply entrenched as the central dogma of cardiology.^1,2,15 The complexity of incorporating additional clinical or ECG clues has led some skeptics to question the need for a new naming system or diagnostic framework. Common arguments included that: 1) current practice already accounts for patients with chest pain by proceeding directly to angiography; or 2) adopting a new framework could lead to an increase in false-positive diagnoses. However, the first assertion is dangerously flawed: only a minority of patients presenting with refractory chest pain in the emergency department are ultimately diagnosed with ACO,⁴⁴ and studies show that the majority of patients with ongoing pain are not referred for catheterization.²⁹ The second concern is also unfounded, as evidence from expert-driven and AI-supported ECG algorithms demonstrates that these subtle findings are as reliable as traditional STEMI criteria and an OMI-based approach to ECG interpretation outperforms the STEMI paradigm.^12,13,26,33

The bargaining phase emerged as evidence accumulated regarding missed ACOs and the real possibility of their prospective identification. These cases were often reclassified as “STEMI equivalents,” preserving the traditional terminology.¹ Many distinct ECG patterns have been identified as indicators of ACO.45, 46, 47, 48, 49, 50, 51, 52, 53 However, some even do not include ST-segment elevation,^52,53 while others exhibit levels of ST-elevation indistinguishable from stable variants,46, 47, 48, 49, 50, 51 requiring careful differentiation.²⁰ Attempting to retain the “STEMI” label by appending 'equivalents' reflects a flawed insistence on rigidly adhering to a diagnostic test rather than defining the disease based on its underlying pathology. This approach has a major limitation: it restricts the diagnosis of ACO solely to ECG findings, despite the possibility of ACO occurring without any identifiable ECG changes. Another form of bargaining arises among pragmatists who accept the possibility of diagnosing ACO in the absence of STEMI criteria but question the feasibility of teaching first responders to recognize the subtle ECG findings indicative of OMI. This represents a helpful step forward, as history has shown that necessity drives innovation – once the need for a new diagnostic paradigm is acknowledged, solutions often follow. The advent of AI has largely addressed this concern, making the widespread application of the OMI paradigm feasible.³³

Thus, the OMI/NOMI paradigm has evolved into a comprehensive diagnostic and therapeutic framework. We now stand on the brink of acceptance, with the remaining step being the validation of its clinical impact through a prospective randomized controlled trial. The DIFOCCULT-3 trial is designed to fulfill this need, evaluating whether the OMI/NOMI approach results in improved outcomes compared to the STEMI/NSTEMI paradigm.

Study Limitations

This study has several limitations inherent to its design. First, although modified cluster randomization by the interventionalist team mitigates ethical concerns and cross-contamination, it introduces potential center- and operator-level variability. While this was addressed using calibration cohorts and frailty models, residual confounding may persist. The assumed intracluster correlation coefficient may have been underestimated, which could reduce statistical power if interventionalist-level variation is larger than anticipated. Second, despite efforts to standardize training and data collection, variations in the application of both the STEMI/NSTEMI and the OMI/NOMI diagnostic framework across centers may impact the consistency of diagnoses. Third, OMI diagnosis includes a degree of subjectivity, particularly when using clinical gestalt and adjunctive diagnostic tools; however, this reflects real-world practice and is complemented by expert and AI-assisted review. We acknowledge that reliance on clinical gestalt especially introduces subjectivity and may vary by operator experience, but the frequency and impact of these decisions will be reported and analyzed separately. Fourth, secondary outcome measures such as infarct size or LVEF are not assessed by a core laboratory, which may introduce heterogeneity in measurement techniques. Fifth, the absence of a centralized angiographic core lab may introduce center-level variation in angiographic interpretations. Lastly, as the trial is conducted entirely in Türkiye, the generalizability of findings to healthcare systems with different infrastructures and STEMI/NSTEMI management protocols may be limited.

Conclusions

The DIFOCCULT-3 study aims to determine whether the OMI/NOMI paradigm leads to a meaningful shift in the modern management of acute MI.

Perspectives.

COMPETENCY IN MEDICAL KNOWLEDGE: The DIFOCCULT-3 trial underscores that a significant proportion of acute coronary occlusions are not detected by standard STEMI criteria. Clinicians should recognize that subtle ECG findings, clinical gestalt, and emerging AI tools can support earlier identification of occlusive MI, potentially reducing treatment delays and improving outcomes in patients currently labeled as NSTEMI.

TRANSLATIONAL OUTLOOK: If the OMI/NOMI paradigm proves superior, adoption into practice will require updates to training, diagnostic pathways, and guidelines. Key barriers include variability in ECG interpretation skills, integration of AI tools into clinical workflows, and system-level adjustments to support faster decision-making. Further studies should assess how to best implement these tools across diverse practice settings and measure their impact on long-term outcomes and healthcare resources.

Funding support and author disclosures

Drs Meyers and Smith hold stocks from Powerful Medical. All other authors have reported that they have no relationships relevant to the contents of this paper to disclose.