Inter-reader variability when applying the 2013 Fleischner guidelines for potential solitary subsolid lung nodules

Abstract

Background: In 2013, the Fleischner Society published recommendations for managing subsolid pulmonary nodules. Inter-reader variability has not yet been defined and has potential implications for the ease and reproducibility of applying the guidelines to clinical practice.

Purpose: To evaluate inter-reader variability when applying the 2013 Fleischner guidelines for potential solitary subsolid lung nodules.

Material and methods: Potential nodules were identified through a systematic retrospective review of CT studies that reported a ground-glass lesion. Three radiologists decided whether these lesions fit criteria of a subsolid nodule and thus merit application of the Fleischner Society guidelines, determined if a solid component was present, and measured each component in two dimensions. Final management recommendations were based on these intermediate decisions. Inter-reader variability for management was calculated and Fleiss' kappa was used to determine significance. Logistic regression and Fisher's exact test determined whether management was contingent on each intermediate decision.

Results: Forty-four nodules with mean diameter of 9.4 mm were evaluated by three radiologists. Final management recommendations were in agreement for 93 out of 132 cases (70.4%, kappa = 0.56). Inter-reader variability in management recommendation was contingent on disagreement over whether a pulmonary lesion fit criteria of a subsolid nodule for 24 cases (P < 0.01), whether there was a solid component for 10 cases (P = 0.01), and whether the measurement met the threshold of 5 mm for five cases (P = 0.12).

Conclusion: There is moderate inter-reader variability when applying the 2013 Fleischner Society management recommendations. Significant contributors of variability include whether the potential lesions fit subsolid nodule criteria and whether a solid component is present. Measurement variability does not significantly affect the final management decisions.

Keywords: CT – spiral; decision analysis; lung; thorax.