Feasibility of computed tomography-derived surgical margin assessment in an ex vivo sublobar lung resection model

Objectives: Computed tomography (CT) imaging of a sublobar resection specimen may inform intraoperative surgical margin assessment. However, consistency with final pathological margins has not been previously evaluated. In this study, we investigated the concordance between surgical margin measurements by CT versus pathology measurements using an ex vivo sublobar lung resection model.

Methods: Pig lung wedge samples containing agarose pseudotumours were harvested. CT images were acquired following specimen inflation. The specimen was bisected along the same plane observed by CT for accurate comparison with pathological surgical margin measurement. The bisected samples were then fixed in formalin before preparing haematoxylin & eosin slides. Surgical margin length at four distinct stages (CT, gross pre-formalin fixation, gross post-formalin fixation and pathology) were measured and compared.

Results: A total of 50 lung specimens were analysed. After specimen processing, Surgical margin length decreased in 94% (47/50) and increased in 6% (3/50) of samples. Mean surgical margin lengths were as follows: CT 14.0 mm (range: 4.5-28.3 mm), gross pre-formalin fixation 13.0 mm (range: 4.0-25.0 mm), gross post-formalin fixation 12.1 mm (range: 2.5-26.0 mm) and pathology 10.9 mm (range: 1.0-23.4 mm). There was an average -23.8% (range: +11 to -82%) change in surgical margin length from CT to final pathology (P < 0.001).

Conclusions: While CT-based surgical margin measurement is feasible, we observed an average 23.8% discordance when compared to final pathology measurement. Surgeons must be aware that the CT-derived surgical margin generally overestimates the pathology-derived surgical margin.