Artificial Intelligence in Chronic Obstructive Pulmonary Disease Phenotyping: A Valuable Tool? Volume 65- Issue 4

Siafakas N1* and Koutonen S2

• ¹Department of Thoracic Medicine, Medical School, University of Crete, Crete, Professor Emeritus Of Medicine, University of Crete, Heraklion, Greece
• ²Department of Thoracic Medicine, Medical School, University of Crete, Crete, Msc, University of Crete, Medical School, Greece

Received: May 11, 2026; Published: May 18, 2026

*Corresponding author: Nikolaos Siafakas, Department of Computer Science, Room H307, University Campus Heraklion 70013, Crete, Greece

DOI: 10.26717/BJSTR.2026.65.010230

Abstract PDF

Chronic Obstructive Pulmonary Disease (COPD) is a common, preventable, and treatable disease that is characterized by persistent respiratory symptoms and airflow limitation. It is a leading cause of morbidity and mortality world-wide and presents a substantial financial burden in Health systems everywhere [1]. Furthermore, it is marked by considerable heterogeneity, which creates significant challenges in daily clinical practice [1,2]. To enhance the proper clinical management of COPD patients the concept of phenotypes had been developed to cluster individuals exhibiting similar clinical, radiological, functional and biological traits [3].

Phenotyping has advanced precise medicine, particularly in quiding treatment decisions to each individual patient [3]. However, because of the numerous types, their variability and especially their overlap this valuable process of phenotyping is regarded as quite intense and challenging [2,3]. Recently, the application of Artificial Intelligence (AI) in Medicine had demonstrated notable advancements in diagnosis, severity, prognosis, and efficacy of different treatment approaches. While approximately 300 studies have explored the use of AI in COPD, only a very small number focus on the topic of phenotypes. Some researchers employed radiological image-based AI methods to analyse the phenotypes of COPD. They mainly utilized data bases of CT-images and compered their findings with those of methods using clinical and lung function data. These studies highlighted the effectiveness of radiological AI techniques asserting that these methods could play a significant role in the phenotypic analysis [4,5]. Another approach focus on clinical and spirometric data noting that these databases are essential for AI driven phenotypic studies [6,7] A third approach employed specialized Deep Learning methods to predict disease exacerbations, overall prognosis and specific comorbidities such us sleep disorders and cardiovascular [5,8,9]. Other studies utilized the GENE-COPD data to investigate the phenotypic variability based on genetic findings [10]. Recently, cellular, multi-omic, and other biological biomarkers have been integrated into data for AI exploration of COPD phenotypes with very promising results [11]. Finally, Paoletti et al added clinical data, lung function tests, radiological CT-images and blood gases into an AI algorithm and improved phenotypic analysis [12]. This Editorial concludes that while the application of AI in the phenotypic analysis of COPD patients shows great promises it is still in its early stage. Currently it tends to rely on a limited number of parameters rather than adapting a holistic approach. Consequently, there is a need for a comprehensive AI system that integrates clinical, radiological, functional, biological, cellular, multi-omic, and genetic data to enhance personalized management in COPD., considering though the potential risks of AI in Medicine [13].

1. Celli BR, MacNee W (2004) ATS/ERS Task Force. Standards for the diagnosis and treatment of patients with COPD: a summary of the ATS/ERS position paper. Eur Respir J 23(6): 932-946.
2. Christenson SA (2023) COPD Phenotyping. Respir Care 68(7): 871-880.
3. Siafakas N, Corlateanu A, Fouka E (2017) Phenotyping Before Starting Treatment in COPD?. COPD 14(3): 367-374.
4. Almeida SD, Norajitra T, Lüth CT, Wald T, Weru V, et al. (2024). Capturing COPD heterogeneity: anomaly detection and parametric response mapping comparison for phenotyping on chest computed tomography. Front Med (Lausanne) 11: 1360706.
5. Yuan NF, Hasenstab K, Retson T, Conrad DJ, Lynch DA, et al. (2022) Unsupervised Learning Identifies Computed Tomographic Measurements as Primary Drivers of Progression, Exacerbation, and Mortality in Chronic Obstructive Pulmonary Disease. Ann Am Thorac Soc 19(12): 1993-2002.
6. Bodduluri S, Nakhmani A, Reinhardt JM, Wilson CG, McDonald ML, et al. (2020) Deep neural network analyses of spirometry for structural phenotyping of chronic obstructive pulmonary disease. JCI Insight 5(13).
7. Akgün KM, Sigel K, Cheung KH, Kidwai-Khan F, Bryant AK, et al. (2020) Extracting lung function measurements to enhance phenotyping of chronic obstructive pulmonary disease (COPD) in an electronic health record using automated tools. PLoS One 15(1).
8. Feng S, Zhang R, Zhang W, Yang Y, Song A, et al. (2025) Predicting Acute Exacerbation Phenotype in Chronic Obstructive Pulmonary Disease Patients Using VGG-16 Deep Learning. Respiration 104(1): 1-14.
9. Nikolaou V, Massaro S, Garn W, Fakhimi M, Stergioulas L, et al. (2021) The cardiovascular phenotype of Chronic Obstructive Pulmonary Disease (COPD): Applying machine learning to the prediction of cardiovascular comorbidities. Respir Med 186: 106528.
10. Castaldi PJ, Boueiz A, Yun J, Estepar RSJ, Ross JC, et al. (2020) COPDGene Investigators. Machine Learning Characterization of COPD Subtypes: Insights from the COPDGene Study. Chest 157(5): 1147-1157.
11. Gillenwater LA, Helmi S, Stene E, Pratte KA, Zhuang Y, et al. (2021) Multi-omics subtyping pipeline for chronic obstructive pulmonary disease. PLoS One 16(8).
12. Paoletti M, Camiciottoli G, Meoni E, Bigazzi F, Cestelli L, et al. (2009) Explorative data analysis techniques and unsupervised clustering methods to support clinical assessment of Chronic Obstructive Pulmonary Disease (COPD) phenotypes. J Biomed Inform 42(6): 1013-1021.
13. Siafakas N, Vasarmidi E (2024) Risks of artificial intelligence (AI) in medicine. Pneumon 37(3): 40.