Monica Kraft, MD, is Murray M. Rosenberg Professor of Medicine and System Chair for the Department of Medicine at Mount Sinai Health System and the Icahn School of Medicine at Mount Sinai.

Monica Kraft, MD

AI-Enhanced ECGs Detect Early COPD, With Monica Kraft, MD

Recent research suggests that routine electrocardiograms (ECGs) analyzed through artificial intelligence (AI) may offer a scalable approach to identifying early 

, Monica Kraft, MD, of the Icahn School of Medicine at Mount Sinai, discussed key findings from a new study that assessed the accuracy of an automated COPD diagnosis using deep learning on ECGs.

“A lot of patients, especially as the years go by, undergo cardiology evaluation and don't necessarily have the lungs evaluated similarly,” Kraft said. “[This tool] may allow us to bring other issues to light.”

Chronic obstructive pulmonary disease (COPD) remains a leading cause of morbidity and mortality worldwide. The non-specific symptoms and reliance on time- and resource-intensive testing often delay a diagnosis.

Kraft and colleagues analyzed ≥ 760,000 ECGs across multiple hospital systems and the UK Biobank using a convolutional neural network, demonstrating robust performance in detecting COPD (area under the curve, 0.80–0.82). Kraft emphasized the role of subtle P wave changes as early indicators of COPD.

“The lungs and the heart are very interrelated,” she explained. “The heart is very subject to pressure changes that go on in the lung. COPD starts in [the] smallest [lung branches]. They can be a millimeter in size. The air that we breathe can get into those little airways, but has trouble getting out because those small airways are closed off. And so, those little airways don't empty as well. They hold on to this air, and the lungs get a little bit bigger than they should. That size change…can be very subtle, [and] early on, alters the pressure between the lungs and the heart and also the lungs and the ribs, the chest wall and the diaphragm, all these relationships get altered. Those…subtle pressure changes, we think, actually have an impact on the P wave.”

The model performed well in patients with documented smoking histories, underscoring its relevance to populations at greater risk. Kraft noted potential future applications for other obstructive airway diseases, such as asthma, cystic fibrosis, and bronchiectasis, although further study is needed.

If applied across a health system, Kraft said the algorithm could flag possible COPD on routine ECGs. Cardiologists could note this, prompting primary care providers to initiate further evaluation or pulmonology referral. This could lead to earlier intervention and improved outcomes.

She cautioned that the model is not perfectly specific. Only about 70% of flagged ECGs corresponded to confirmed COPD. Kraft said that a prospective evaluation will be important to define disease severity and guide clinical response before widespread deployment.

Despite limitations, Kraft highlighted the promise of AI-enhanced ECGs, which are inexpensive and easily accessible. Early detection of COPD could slow lung function decline, reduce hospitalizations, and improve survival. This study supports AI-based ECG as a tool for early COPD recognition, complementing traditional diagnostic methods such as spirometry.

“We need to remember that we also are at the early stages of this approach,” Kraft said. “We can identify COPD early, but we need to now understand what is the extent of the disease that we're identifying. We know a little bit about that from the chart review, but what we want to do is bring those patients in live and do an evaluation so that we can describe the type of COPD that we are identifying. Those are all pieces that are in the works…before it becomes a decision support tool at the ECG level, we would need to sort all of that out.”

Part 1 of our interview with Kraft can be viewed 

A relevant disclosure for Kraft includes AstraZeneca Pharmaceuticals LP.

Vaid A, Sharma J, Jiang J, et al. Automated diagnosis of chronic obstructive pulmonary disease using deep learning applied to electrocardiograms. 

Videos

Deep learning applied to routine ECGs identifies subtle cardiac signals linked to early COPD, potentially enabling earlier diagnosis and intervention.

AI-ECG Model Detects COPD Months Before Diagnosis, With Monica Kraft, MD

A deep learning model applied to routine electrocardiograms (ECGs) identified 

chronic obstructive pulmonary disease (COPD)

 months before clinical diagnosis, according to a new study.

The approach, validated across multiple cohorts, suggests a potential role for artificial intelligence (AI)-enhanced ECG analysis as a scalable screening tool in routine clinical care.

, investigator Monica Kraft, MD, from the Icahn School of Medicine at Mount Sinai, emphasized the clinical importance of earlier recognition. She noted that identifying patients with early or subclinical COPD could create opportunities for timely intervention, particularly smoking cessation and initiation of therapies aimed at slowing disease progression.

“I think what's happened in the past is when we've done lung function screening [and] testing, oftentimes the lung function…in the clinic looks very normal,” Kraft said. “You can imagine [a] scenario where patients are smoking, they have their lung function done, it looks normal, and they think,’ Hey, this is great. I can keep smoking.’ Whereas, if we have a mechanism that demonstrates that there may actually be some lung damage going on because of what the EKG shows, then there may be an opportunity to really educate the patient that it's really important that they make a lifestyle change now in order to avoid future problems down the road.”

COPD remains a leading cause of morbidity and mortality worldwide, with delayed diagnosis representing a persistent challenge. Symptoms are often nonspecific in early disease, and confirmatory testing, primarily spirometry, can be underutilized due to logistical and resource constraints.

 Current guidelines do not recommend routine screening in asymptomatic individuals, further contributing to underdiagnosis.

In this study, Kraft and colleagues developed a convolutional neural network (CNN) model trained on > 760,000 ECGs from > 67,000 patients across the Mount Sinai Health System, with additional validation in an independent cohort from the UK Biobank.

 The model demonstrated consistent performance, with AUROC values of 0.80 in internal testing, 0.82 in external validation, and 0.75 in the UK Biobank cohort.

The model’s predictive signal appeared to precede clinical diagnosis. Analyses showed that ECG-based predictions could identify COPD-related changes approximately 6 to 15 months before formal diagnosis, suggesting the model is capturing early physiologic alterations rather than late-stage disease alone.¹ Kraft highlighted that this window may be clinically actionable for high-risk individuals, such as active smokers, where earlier counseling and intervention could alter disease trajectory.

The model leverages subtle ECG features associated with pulmonary vascular and cardiac changes seen in COPD. Explainability analyses indicated a focus on P-wave morphology, consistent with early right atrial and pulmonary vascular alterations.

 While such changes are typically recognized in advanced disease, the application of AI enables detection of patterns not readily discernible to clinicians interpreting individual ECGs.

The clinical implications of this approach lie in its potential integration into existing workflows. ECGs are widely available, low-cost, and routinely performed across inpatient and outpatient settings. As Kraft noted, this creates an opportunity for patients undergoing ECGs for unrelated indications to be flagged for further pulmonary evaluation, which is valuable in health systems with limited access to spirometry or populations less likely to undergo formal testing.

However, the model is not intended to replace spirometry, which remains the diagnostic standard for COPD. Rather, it may function as a triage or decision-support tool, identifying patients who warrant confirmatory testing.

Investigators highlighted several limitations, including reliance on ICD-coded diagnoses, declining performance over time, and modest correlations with spirometry, suggesting the model may reflect cumulative cardiopulmonary changes more than early airflow limitation alone.

The team noted that prospective validation is needed to determine whether earlier detection via AI-ECG translates into improved clinical outcomes. Kraft underscored this, noting that the next phase of research will involve identifying at-risk patients in real time, conducting comprehensive clinical evaluations, and assessing whether early interventions, such as smoking cessation or pharmacologic therapy, alter disease progression.

“We still have some work to do. The algorithm, just creating it and applying it, is step 1,” Kraft said. “Step 2 is now to see how it applies to patients moving forward so [not] quite in the hands of clinicians yet, but we hope it will be very soon.”

https://doi.org/10.1016/j.ebiom.2025.106066

World Health Organization. Chronic obstructive pulmonary disease (COPD). 

https://www.who.int/news-room/fact-sheets/detail/chronic-obstructive-pulmonary-disease-(copd)

US Preventive Services Task Force. Screening for chronic obstructive pulmonary disease: recommendation statement. 

Deep learning applied to routine ECGs identified COPD up to 15 months earlier, supporting earlier intervention and risk stratification strategies.


As-Needed Rescue Treatment with Albuterol–Budesonide for Mild Asthma, with Monica Kraft, MD

BATURA phase 3 trial results are providing new evidence supporting albuterol/budesonide (Airsupra)’s benefit as standard of care for as-needed rescue treatment in patients with 

the American Thoracic Society (ATS) International Conference 2025

 and highlight a 47% reduced risk of severe exacerbations in mild asthma with albuterol/budesonide compared with albuterol alone.

“Our guidelines are a little antiquated when it comes to mild asthma here in the US,” Monica Kraft, MD, Murray M. Rosenberg Professor of Medicine and System Chair for the Department of Medicine at Mount Sinai Health System and the Icahn School of Medicine at Mount Sinai, explained to 

. “I think we have now appreciated that mild asthma really isn't always mild, that many patients who have mild asthma have significant asthma exacerbations. I think we've misperceived it over the years, that mild asthma is just something not to worry about.”

Kraft goes on to describe a recent shift toward prioritizing exacerbation reduction and prevention, drawing comparisons to how doctors treat blood pressure long-term to prevent strokes and cholesterol to prevent cardiac issues. She additionally describes Airsupra as "a great addition to our asthma armamentarium."

Airsupra, formerly known as PT027, is a first-in-class SABA/ICS rescue treatment for asthma in the US, to be taken as needed. It is an inhaled, fixed-dose combination rescue medication containing albuterol and budesonide and has been developed in a pressurised metered-dose inhaler using AstraZeneca’s Aerosphere delivery technology.

BATURA was a phase 3b, US, randomised, double-blind, parallel-group, event-driven trial, comparing the efficacy and safety of using inhaled albuterol/budesonide (180mcg/160mcg) as an as-needed rescue medication in response to symptoms compared to as-needed albuterol (180mcg) for up to 12 months.

A total of 2516 participants underwent randomization, of whom 1797 (71.4%) completed the trial. Of 2421 participants in the full analysis population (1209 assigned to the albuterol–budesonide group and 1212 to the albuterol group), 97.2% were ≥ 18 years of age and 74.4% used a SABA alone at baseline.

The trial was stopped for efficacy at a prespecified interim analysis. A severe exacerbation occurred in 5.1% of the participants in the albuterol–budesonide group and in 9.1% of those in the albuterol group in the on-treatment efficacy population (hazard ratio [HR], 0.53; 95% confidence interval [CI], 0.39-0.73) and in 5.3% and 9.4%, respectively, in the intention-to-treat population (HR, 0.54; 95% CI, 0.40-0.73) (

The annualized rate of severe asthma exacerbations was lower with albuterol–budesonide than with albuterol (0.15 vs 0.32; rate ratio, 0.47; 95% CI, 0.34-0.64), as was the mean annualized total dose of systemic glucocorticoids (23.2 vs 61.9 mg per year). Adverse events were similar in the treatment groups.

“The BATURA study showed us that even those patients with what we would consider very mild asthma can benefit from early anti-inflammatory therapy and ongoing anti-inflammatory therapy,” Kraft said. “I think we want to start thinking about that in terms of prevention, but also treating mild asthma appropriately, because mild asthma isn't always mild.”

Editors’ note: Kraft has relevant disclosures with AstraZeneca, Chiesi, Genentech, GlaxoSmithKline, Kinaset Therapeutics, Regeneron, and Sanofi.

LaForce C, Albers F, Danilewicz A, et al. As-Needed Albuterol–Budesonide in Mild Asthma. 

https://www.nejm.org/doi/full/10.1056/NEJMoa2504544

AstraZeneca. Statistically significant and clinically meaningful BATURA Phase III trial results provide new evidence for Airsupra as standard of care for as-needed rescue treatment in asthma. May 19, 2025. Accessed May 19, 2025. 

https://www.astrazeneca.com/media-centre/press-releases/2025/statistically-significant-clinically-meaningful-batura-phase-iii-trial-provide-evidence-airsupra-standard-care-as-needed-rescue-treatment-asthma.html

Phase 3 BATURA data reveal a 47% reduction in the risk of severe exacerbations in mild asthma with albuterol/budesonide (Airsupra) versus albuterol alone.

Monica Kraft, MD

Articles