Perfect diagnostic accuracy in a small derivation cohort and outdated classification criteria: Methodological concerns regarding miR-21 as a biomarker for chronic obstructive pulmonary disease-associated pulmonary hypertension

Dear Editor,

We read with great interest the case-control study by Eid et al., which investigated the diagnostic utility of circulating microRNA-21 (miR-21) and NT-proBNP for pulmonary hypertension (PH) in COPD patients, published in Lung India.[1] The rigorous laboratory methodology for miR-21 quantification and the translational relevance of the research question are appreciated. We wish to highlight two substantive methodological concerns that affect the reliability of the reported findings.

The most notable result—area under the curve (AUC) =1.000 with 100% sensitivity and 100% specificity for miR-21—warrants careful examination. A perfect biomarker for a clinically diverse condition such as COPD-associated PH is biologically unlikely, and in the diagnostic biomarker literature, an AUC of 1.000 in small derivation samples is a well-recognised indicator of spectrum bias rather than true discriminatory power.[2] No internal validation procedure—such as bootstrapping, cross-validation or split-sample analysis—was performed. In the absence of validation, the reported performance metrics cannot be considered reliable, and the conclusion that miR-21 “outperforms” NT-proBNP requires substantiation in an independent cohort, as mandated by Transparent Reporting of a Multivariable Prediction Model for Individual Prognosis or Diagnosis (TRIPOD) reporting standards.[3]

The mechanism behind this near-perfect separation appears to be the diagnostic threshold used for PH. Patients were classified as COPD-PH based on pulmonary artery systolic pressure (PASP) >50 mmHg or tricuspid regurgitation velocity >3.4 m/s via echocardiography. However, the current ESC/ERS 2022 guidelines define PH as a mean pulmonary artery pressure >20 mmHg measured by right heart catheterisation (RHC).[4] By limiting the COPD-PH group to patients with severe echocardiographic PH (PASP >50 mmHg, roughly corresponding to mPAP 35–40 mmHg), the authors create a stark bimodal phenotypic distribution. This deliberate selection of severe-end cases artificially maximises between-group differences in miR-21 expression and inflates discriminatory performance. The clinically critical diagnostic challenge—detecting mild-to-moderate PH in COPD (mPAP 21–35 mmHg)—is entirely unaddressed. Furthermore, echocardiographic PASP is prone to systematic overestimation in COPD due to hyperinflation and suboptimal acoustic windows, and no RHC was performed to confirm group allocation.[5]

We commend the authors for investigating miR-21 in this understudied PH subtype and encourage future studies incorporating RHC-confirmed diagnosis using current ESC/ERS thresholds, prospective enrolment, and rigorous internal and external validation to provide reliable evidence for clinical translation.

Conflicts of interest

There are no conflicts of interest.

Funding Statement

Nil.