Stavropoulou, A; Szmul, A; Chandy, E; Veiga, C; Landau, D; McClelland, JR; (2021) A multichannel feature-based approach for longitudinal lung CT registration in the presence of radiation induced lung damage. Physics in Medicine & Biology 10.1088/1361-6560/ac1b1d. (In press).

Preview

Text Stavropoulou+et+al_2021_Phys._Med._Biol._10.1088_1361-6560_ac1b1d.pdf - Accepted Version Download (1MB) | Preview

Abstract

Quantifying parenchymal tissue changes in the lungs is imperative in furthering the study of radiation-induced lung damage (RILD). Registering lung images from different time-points is a key step of this process. Traditional intensity-based registration approaches fail this task due to the considerable anatomical changes that occur between timepoints. This work proposes a novel method to successfully register longitudinal pre- and post-radiotherapy (RT) lung CT scans that exhibit large changes due to RILD, by extracting consistent anatomical features from CT (lung boundaries, main airways, vessels) and using these features to optimise the registrations. Pre-RT and 12-month post-RT CT pairs from fifteen lung cancer patients were used for this study, all with varying degrees of RILD, ranging from mild parenchymal change to extensive consolidation and collapse. For each CT, signed distance transforms from segmentations of the lungs and main airways were generated, and the Frangi vesselness map was calculated. These were concatenated into multi-channel images and diffeomorphic multichannel registration was performed for each image pair using NiftyReg. Traditional intensity-based registrations were also performed for comparison purposes. For the evaluation, the pre- and post-registration landmark distance was calculated for all patients, using an average of 44 manually identified landmark pairs per patient. The mean (standard deviation) distance for all datasets decreased from 15.95 (8.09) mm pre-registration to 4.56 (5.70) mm post-registration, compared to 7.90 (8.97) mm for the intensity-based registrations. Qualitative improvements in image alignment were observed for all patient datasets. For four representative subjects, registrations were performed for 3 additional follow-up timepoints up to 48-months post-RT and similar accuracy was achieved. We have demonstrated that our novel multichannel registration method can successfully align longitudinal scans from RILD patients in the presence of large anatomical changes such as consolidation and atelectasis, outperforming the traditional registration approach both quantitatively and through thorough visual inspection.

Download activity - last month

Export as XMLJSONCSV

Download activity - last 12 months

Export as CSVXMLJSON

Downloads by country - last 12 months

Export as XMLJSONCSV

Archive Staff Only

View Item