A feature-based affine registration method for capturing background lung tissue deformation for ground glass nodule tracking

Abstract

Apparent changes in lung nodule size assessed via simple image-based measurements from computed tomography (CT) images may be compromised by the effect of the background lung tissue deformation on the nodule, leading to erroneous nodule tracking. We propose a feature-based affine registration method and study its performance vis-a-vis several other registration methods. We implement and test each registration method using a lung- and a lesion-centred region of interest on 10 patient CT datasets featuring 12 nodules. We evaluate each registration method according to the target registration error (TRE) computed across 30–50 homologous fiducial landmarks selected by expert radiologists. Our results show that the proposed feature-based affine lesion-centred registration yielded a 1.11.2 mm TRE, while a Symmetric Normalisation deformable registration yielded a 1.21.2 mm TRE, with a baseline least-square fit of the validation fiducial landmarks of 1.51.2 mm TRE. The proposed feature-based affine registration is computationally efficient, eliminates the need for nodule segmentation, and reduces the susceptibility of artificial deformations. We also conducted a pilot pre-clinical study that showed the proposed featurebased lesion-centred affine registration effectively compensates for the background lung tissue deformation and serves as a reliable baseline registration method prior to assessing lung nodule changes due to disease.