Leading Pulmonologists from Mount Sinai Unveil Breakthrough Research on AI-Driven Respiratory Disease Diagnosis and Treatment at ATS 2025
At the upcoming American Thoracic Society (ATS) 2025 International Conference, scheduled to take place in San Francisco from May 18 to May 21, experts from the Mount Sinai Health System will present a series of pioneering studies delving into the intersection of artificial intelligence (AI) and respiratory medicine. These presentations spotlight radical advancements in the diagnosis and management of chronic respiratory disorders, including asthma, chronic obstructive pulmonary disease (COPD), obstructive sleep apnea (OSA), and lung cancer. The work from Mount Sinai’s top pulmonologists and biomedical engineers illustrates a future where AI-enhanced diagnostics and data-driven medicine transform patient outcomes in respiratory care.
A key focus will be the development of AI models capable of accurately predicting chronic respiratory failure in COPD patients through analysis of multiple blood biomarkers. Current prognostic tools have struggled to reliably anticipate the trajectory of COPD, a complex condition exhibiting considerable heterogeneity in disease progression and treatment response. Mount Sinai researchers demonstrate how integrating multi-biomarker panels via machine learning algorithms can surpass traditional clinical prediction models, offering personalized insights into disease severity, potential exacerbations, and the timing of respiratory failure. This dynamic approach could revolutionize COPD management by allowing clinicians to tailor interventions based on biological signatures rather than solely clinical assessment.
In a complementary domain, Mount Sinai investigators are pushing the boundaries of AI applications in sleep medicine. Obstructive sleep apnea, a prevalent yet underdiagnosed respiratory disorder, is known for its variable connection between clinical severity—typically measured by apnea-hypopnea index (AHI)—and symptoms such as excessive daytime sleepiness. Leveraging physiology-guided machine learning, researchers are integrating ventilation, hypoxia, arousal, and autonomic nervous system data to develop more robust predictive models for sleepiness. Such models hold immense promise for refining patient stratification, optimizing treatment choices, and elucidating the multifactorial pathophysiology underpinning OSA symptoms.
Further innovations in AI-assisted imaging are also being showcased. One breakthrough involves the creation of deep learning methods for identifying central and upper airway obstructions from pulmonary function test (PFT) flow-volume loops. Traditional visual inspection protocols suffer from low sensitivity, often missing subtle airway narrowing that correlates with clinical respiratory distress. By training AI systems on large datasets of flow-volume curves, Mount Sinai’s team has engineered a tool capable of automatically detecting signature plateau patterns indicative of airway obstruction, enabling earlier and more objective diagnosis. This technology could significantly enhance the clinical utility of PFTs and reduce diagnostic delays in patients presenting with dyspnea.
The integration of hybrid PET/MRI imaging with AI analytics forms another frontier in Mount Sinai’s research portfolio. Using 18F-fluoro-2-deoxy-D-glucose (18F-FDG) PET combined with MRI scans, researchers quantify both visceral and subcutaneous adipose tissue volumes and metabolic activity in patients with OSA before and after continuous positive airway pressure (CPAP) therapy. AI-driven segmentation accelerates the annotation process, facilitating nuanced examinations of how abdominal obesity contributes to cardiovascular disease risk in sleep apnea patients. These insights are crucial as visceral adiposity emerges as a key mediator linking OSA with adverse cardiometabolic outcomes, challenging previous assumptions about CPAP’s cardiovascular benefits.
Mount Sinai’s accomplishments extend into the realm of genomics and molecular biology, particularly in elucidating asthma’s heterogeneity. The ATLANTIS study, an extensive multi-center investigation, examines small airway disease and inflammatory pathways through transcriptomic analyses of nasal epithelial samples. Novel findings reveal associations between cellular senescence genes—specifically MMP2 and FN1—and asthma susceptibility and severity. Further computational approaches applied to bulk RNA sequencing data have identified circular RNAs (circRNAs) potentially involved in asthma pathogenesis, representing a promising avenue for biomarker discovery and targeted therapies aimed at discrete asthma endotypes.
Clinically rare but consequential phenomena are also under the microscope, such as spontaneous lung intercostal herniation (SLIH) and its complications. Mount Sinai clinicians present a case of SLIH triggered by severe coughing in a COPD patient, highlighting the need for increased awareness of such extraordinary presentations. Additionally, they investigate pulmonary embolism readmissions and complications, offering a data-driven perspective on predictors of acute care utilization in thromboembolic disease. These studies deepen the understanding of complex pulmonary conditions beyond common paradigms.
Beyond disease-specific inquiries, Mount Sinai’s research underscores the pervasive impact of environmental and behavioral factors on respiratory health. For instance, heavy marijuana use and vaping are implicated in rare but severe pulmonary complications such as the Macklin effect—air dissecting through bronchovascular sheaths post-alveolar rupture—with notable implications for asthma exacerbations. Similarly, explorations into sleep architecture irregularities in ICU patients link distinctive atypical N3 (AN3) sleep patterns to neurological outcomes like sepsis-associated encephalopathy, underscoring the interplay between critical care environments and respiratory neurology.
Mount Sinai’s clinicians and researchers are also addressing systemic issues in lung cancer screening and management. Given that millions of incidental lung nodules (ILNs) are identified annually—in a context where a substantial proportion of nodules are lost to follow-up—Mount Sinai has implemented an innovative incidental lung nodule program leveraging natural language processing for efficient patient tracking. Early data demonstrate improved follow-up rates and earlier lung cancer detection, highlighting the potential of informatics and AI to enhance cancer screening workflows and reduce healthcare disparities.
Underpinning much of this pioneering work is the strategic integration of AI, machine learning, and computational biology within Mount Sinai’s academic and clinical ecosystem. By fusing cutting-edge technology with clinical insight, Mount Sinai is driving a paradigm shift towards precision respiratory medicine. Their commitment to leveraging large datasets, advanced imaging modalities, and novel computational tools epitomizes the frontier of pulmonary research, aiming not only to improve diagnostic accuracy but also to unlock new pathways for therapeutics tailored to individual patient profiles.
Moreover, this confluence of AI and respiratory medicine resonates beyond academia. As respiratory diseases like COPD, asthma, and OSA continue to impose significant morbidity and mortality worldwide, innovations such as those emerging from Mount Sinai are poised to influence clinical guidelines and public health policies. By advancing early detection, refining phenotyping, and unraveling complex disease mechanisms, these discoveries promise to reduce health disparities and improve patient quality of life on a global scale.
The Mount Sinai Health System, with its robust infrastructure integrating hospitals, outpatient centers, research laboratories, and educational institutions, exemplifies a holistic approach to healthcare innovation. Its recognition by leading rankings, including U.S. News & World Report’s “Best Hospitals” Honor Roll and Newsweek’s “World’s Best Smart Hospitals,” validates its position at the vanguard of medical research and clinical excellence.
As the ATS 2025 conference unfolds, the respiratory community will closely watch Mount Sinai’s presentations for insights that will likely catalyze future research, clinical practices, and AI applications in the field. This convergence of data science, molecular biology, and clinical medicine heralds a new era in understanding and managing respiratory diseases, promising improved outcomes for millions affected worldwide.
—
Subject of Research: Applications of Artificial Intelligence and Machine Learning in the Diagnosis and Management of Respiratory Disorders including COPD, Asthma, Obstructive Sleep Apnea, and Lung Cancer
Article Title: Leading Pulmonologists from Mount Sinai Unveil Breakthrough Research on AI-Driven Respiratory Disease Diagnosis and Treatment at ATS 2025
News Publication Date: May 21, 2025
References: Research presented at the American Thoracic Society (ATS) 2025 International Conference
Image Credits: The Mount Sinai Health System
Tags: advancements in asthma and OSA diagnosisAI models for chronic respiratory failureAI-driven solutions for lung cancerATS 2025 conference presentationsdata-driven approaches in healthcareinnovative respiratory disease diagnosticsmachine learning in respiratory medicineMount Sinai Health System researchmulti-biomarker analysis for lung conditionspersonalized medicine in COPD treatmentpredicting COPD progression with AItransforming patient outcomes in respiratory care