Defining CT subtypes in chronic obstructive pulmonary disease: real world daily practice does not meet guidelines.

Abstract

Aims: To evaluate the quality and inter-rater reliability of CT-definable chronic obstructive pulmonary disease (COPD) subtype reporting in CT chest reports in a real-world setting, and assess concordance with Fleischner Society guidelines.

Methods: We undertook a retrospective review of 100 randomly selected CT chest scans containing the terms 'emphysema' or 'COPD'. Existing reports were evaluated for the description of emphysema phenotype, severity, and location, as well as the presence of associated findings, benchmarked against the Fleischner Society guidelines for CT reporting. The scans were then read independently by two consultant thoracic radiologists and two radiology specialty residents, blinded to the original reports and each other's assessments. Inter-rater variability was assessed using Light's Kappa for categorical variables and intraclass correlation coefficient (ICC) for ordinal variables.

Results: Emphysema phenotype was described in 51 % of the pre-existing reports, with centrilobular emphysema being the most frequently reported subtype. Only 26 % of reports included all three key descriptors of phenotype, severity and location. Inter-rater agreement was fair for emphysema phenotype (κ = 0.371) and moderate for the grading of paraseptal emphysema (ICC = 0.733), but was more variable for associated features such as large airways disease (κ = 0.0646) and bronchiectasis (κ = 0.0996).

Conclusion: This study shows variability in the quality of CT reporting for COPD in a real-world setting, with frequent omissions of key descriptors and marked inter-rater variability. These findings highlight the need for standardisation in CT reporting, particularly in the context of increasing reliance on imaging for COPD diagnosis and management.