Case-finding of COPD and AI

We read with great interest the paper by Moreno Mendez et al.,¹ reporting near perfection prediction capability (AUC value exceeded 0.97) of GOLD 2023 post-BD airflow limitation (COPD) by using only pre-bronchodilation data in the absence of information on gender and age, variables identified via a machine-learning algorithm. Indeed, AI could be potentially useful as a disease screening method. However, we challenge their statement in Results: “… The positive correlation between pre- and post-BDT spirometry results, along with the more widespread use of forced spirometry without bronchodilator testing in primary care centers, influenced the use of pre-BDT results instead of post-BDT results, without significantly affecting the predictive power of the model based on the AUC of the different classifiers analyzed.”

It is hard to challenge that post-BDT spirometry predicts COPD, as airflow limitation is in any COPD definition. Similarly, it is difficult to defend the usefulness of pre-BDT spirometry, as all COPD case-finding studies performed to date aim to reduce the number of spirometries needed to identify those still undiagnosed. Initiatives like GesEPOC² and GOLD recommend that post-BDT spirometry is still fundamental for diagnosing COPD, although pre-BDT spirometry can be used later during follow-up. Indeed, bronchodilation is a poor discriminative test in COPD.³

Finally, we cannot find the yield of their case-finding study, the number of discarded spirometries because of low quality, or a flow chart of participation, elements that help ensure clarity and transparency in observational research, as encouraged by reporting standards like the EQUATOR STROBE guidelines.⁴ As a token, in our case-finding study of 10,071 adult smokers without prior respiratory diseases recruited by GPs from 48 primary care centres in Spain, COPD was confirmed using spirometry in only 1.9% of all participants.⁵

Declaration of generative AI and AI-assisted technologies in the writing process

None declared.

Funding

None declared.

Authors’ contributions

Both authors wrote this Letter and approved its final version.

Conflicts of interest

DCS declares no conflicts of interest. JBS has received pharmaceutical company grants from 2021 to 2025 from Chiesi, GSK, Linde and Novartis via Hospital Universitario de La Princesa. Participated in speaking activities, advisory committees, and consultancies from 2021 to 2025 sponsored by Air Liquide, Almirall, AstraZeneca, Boehringer Ingelheim, CHEST, Chiesi, CNPT, ERS, FTH, Gebro, Grifols, GSK, IHME, Laminar Pharma, Linde, Lipopharma, Menarini, Mundipharma, Novartis, OMS/WHO, Pfizer, ResApp, RiRL, ROVI, Sapio, SEPAR, Seqirus, WHO EUR, Takeda and Zambon. Finally, JBS declares never, directly or indirectly, received any funding from tobacco manufacturers or their affiliates.