Artificial Intelligence for Comprehensive Diagnostics, Examinations in Acute Coronary Syndromes

In the world of modern medicine, in which technological advancements are driving transformative changes across health care sectors, one innovative tool has emerged to provide hope for improving outcomes in the realm of acute coronary syndromes (ACS). The Algorithm CoDE-ACS, an acronym for Comprehensive Diagnostics and Examinations in Acute Coronary Syndromes, stands as a protentional improvement in data-driven approaches and collaborative efforts to enhance the diagnosis and management of ACS.¹

Image credit: sudok1 - stock.adobe.com

ACS, encompassing conditions such as unstable angina and myocardial infarction, remain one of the leading causes of mortality and morbidity.² Timely and accurate diagnosis, followed by appropriate management, is critical to ensuring optimal patient outcomes. The Algorithm CoDE-ACS is designed to streamline and guide medical professionals through this complexity, resulting in more precise and effective interventions.

In their study, researchers introduced an innovative clinical decision support system that employs machine learning (ML) models to predict an individual's likelihood of experiencing a myocardial infarction. This system generates scores known as Collaboration for the Diagnosis and Evaluation of Acute Coronary Syndrome (CoDE-ACS), drawing on data from the High-Sensitivity Troponin in the Evaluation of Patients with Suspected Acute Coronary Syndrome (High-STEACS) trial population.¹

The CoDE-ACS system incorporates cardiac troponin concentrations obtained either upon initial presentation or through serial testing, integrating them with clinical characteristics to generate a continuous score ranging from 0 to 100. This score serves as an indicator of an individual's prospective risk of developing acute myocardial infarction.

The CoDE-ACS system underwent training in a sequential manner on groups of consecutive patients, differentiating those with and without myocardial injury at presentation. For the purpose of comparison, the researchers also examined traditional diagnostic pathways that rely on sex-specific 99th percentile cardiac troponin I thresholds. These conventional pathways offer predictions regarding the likelihood of myocardial infarction or the potential for ruling it out in an individual.¹

In this study, the research team initially conducted training of the CoDE-ACS system using machine learning (ML) models on data from derivation cohorts comprising 10,038 patients with potential myocardial infarction. The average age within the study cohort was 70 years, with 48% of the participants being female.

All individuals included had experienced adjudicated type 1, 4b, or 4c myocardial infarctions without ST-segment elevation during their initial hospital admission. Comparing patients without myocardial injury at presentation to those with myocardial injury, the CoDE-ACS score exhibited diagnostic performance in accordance with the predefined criteria, with scores below 3 indicating absence of injury and scores of 61 or higher signifying injury.

Among patients without a history of myocardial infarction, the CoDE-ACS displayed a robust negative predictive value of 99.5 and sensitivity of 90.2. Conversely, patients with a history of myocardial infarction exhibited a positive predictive value of 80.1 and specificity of 83.4. These CoDE-ACS scores demonstrated consistent efficacy across all subgroups.¹

The effectiveness of CoDE-ACS in discerning myocardial infarction was evident through its impressive area under the curve (AUC) value of 0.953. Moreover, CoDE-ACS identified 61% of patients at presentation as having a low likelihood of experiencing myocardial infarction.

This figure was notably higher than the 27% identified through the utilization of fixed cardiac troponin thresholds that maintained a comparable negative predictive value. Furthermore, CoDE-ACS scores facilitated the identification of fewer patients at elevated risk of acute myocardial infarction, leading to enhanced positive predictive value.

Patients assigned a low probability of myocardial infarction were linked to minimal post-discharge mortality risk. Remarkably, fewer than 1 in 300 of these individuals experienced cardiac death within a year after symptom onset. Conversely, patients exhibiting an intermediate or high probability of myocardial infarction faced escalated risk of cardiac death within 30 days and 1-year post-presentation, respectively.¹

This AI tool could help bridge the gap in disease diagnosis. The algorithm is effective, regardless of gender, age, and pre-existing health problems. In the case of cardiovascular diseases, prevention and early diagnosis are very important. They allow to prepare a report in a short time, which is ready for interpretation by the physician and making a final diagnosis. This, in turn, empowers them to initiate targeted therapies promptly, minimizing the potential for adverse events and preventing unnecessary interventions.

CoDE-ACS Pathway is an innovative clinical decision support system built on machine learning models. This pathway integrates diverse data, including testing time, serial cardiac troponin measurements at flexible intervals, and time from symptom onset.

Moreover, the algorithm CoDE-ACS shows the significance of interdisciplinary collaboration. By fostering communication between cardiologists, emergency physicians, radiologists, laboratory personnel, nursing, and pharmacy, the tool promotes a holistic approach to patient care. This collaborative framework not only enhances diagnostic accuracy but also facilitates the seamless transition between phases of care, such as the acute phase and subsequent post-ACS management.

As we navigate the ever-evolving landscape of health care, the algorithm CoDE-ACS exemplifies the potential of algorithms and data-driven solutions to revolutionize patient care. By merging cutting-edge technology with clinical acumen, it epitomizes the convergence of medicine and artificial intelligence for the greater good. As more institutions and health care professionals adopt and adapt the CoDE-ACS algorithm, we can anticipate improved outcomes, reduced mortality rates, and a brighter future for patients confronting the challenges of ACS.

Due to the growing number of people requiring cardiac diagnostics, artificial intelligence will also contribute to increasing the benefits for patients. These innovations may extend prophylaxis and treatment to a wider group of patients, precise and early detection of heart rhythm abnormalities and increase the time for physician-patient contact which may contribute to increase in satisfaction with health care services.

The algorithm CoDE-ACS stands as an exemplar of innovation in modern medicine, bridging the gap between data analytics and clinical practice. Through its comprehensive diagnostic approach, interdisciplinary collaboration, and adaptability, it holds the promise of transforming ACS management and, by extension, the lives of countless individuals worldwide.

As we continue to explore the capabilities of this groundbreaking tool, the algorithm CoDE-ACS serves as a beacon of hope for improved patient care, highlighting the boundless possibilities that lie at the intersection of technology and health care.

In the context of the CoDE-ACS implementation and utilization, pharmacists can play a crucial and multifaceted role in various stages of its implementation and utilization. Pharmacists can actively participate in patient assessment and data collection, ensuring accurate and comprehensive information for the CoDE-ACS algorithm.

They can review patient medication histories and other relevant clinical data to contribute to a holistic understanding of the patient's condition. Given their expertise in pharmacology and clinical knowledge, pharmacists can interpret cardiac troponin concentration results, which are essential inputs for the CoDE-ACS algorithm. They can collaborate with the clinical team to assess the significance of troponin levels in the context of the patient's overall health status.

Pharmacists can actively engage in the implementation of the CoDE-ACS pathway within the hospital or health care system. Pharmacists can contribute to patient follow-up by monitoring patients who have been assessed using the CoDE-ACS pathway. This includes tracking outcomes and collaborating with the health care team to ensure that patients receive appropriate care based on their CoDE-ACS scores.

Pharmacists may play an integral role in various aspects of the CoDE-ACS pathway implementation, leveraging their clinical expertise, medication knowledge, and collaboration skills to enhance patient care and optimize the diagnostic process for ACS.

References

1. Doudesis D, Lee K, Boeddinghaus J, et al. Machine learning for diagnosis of myocardial infarction using cardiac troponin concentrations. Nature Medicine. 2023; 29(5), 1201-1210. doi:10.1038/s41591-023-02325-4.
2. Tsao CW, Aday AW, Almarzooq ZI, Beaton AZ, Bittencourt MS, Boehme AK, et al. Heart Disease and Stroke Statistics—2023 Update: A Report From the American Heart Association. Circulation. 2023;147:e93–e621.