Using transitional changes on hrct to assess the impact of treatment with cyclophosphamide or mycophenolate on systemic sclerosis-related interstitial lung disease from scleroderma lung study II

Introduction: Systemic sclerosis-related interstitial lung disease (SSc-ILD) is the leading cause of mortality in SSc. High-resolution computed tomography (HRCT) is increasingly used to measure SSc-ILD progression. This study examines whether changes in the quantitative extent of specific ILD patterns can be used to assess treatment response in the context of an SSc-ILD clinical trial. Material and Methods: We evaluated changes in the quantitative extent of specific lung patterns using volumetric HRCT scans obtained at baseline and 2 years in patients randomized to treatment with either cyclophos-phamide (CYC for one year followed by placebo in the second year) or mycophenolate (MMF for 2 years) in the Scleroderma Lung Study-II. The evaluated ILD patterns included lung fibrosis (LF), ground glass (GG), honeycombing (HC), and normal lung (NL). The paired baseline and 2-year HRCT scans were mapped in voxel by voxel using registration techniques and transitional ILD patterns were quantified. To calculate transitional ILD scores, the following steps were taken: (1) lung/lobe segmentation; (2) denoise; (3) grid-sampling; (4) texture feature calculation; (5) classification and book-keeping the location of the original image; (6) registration; (7) identification of the nearest voxels leading to mapping between the two scans; and (8) expressing a proportion of the voxels at baseline compared to those at 2-year. T-tests and mixed effect models were used to test transitional changes. Results: Forty-seven and 50 patients had baseline and follow-up scans in the CYC and MMF arms, respectively. Baseline characteristics were similar in two arms (mean 62 years old, 66% predicted forced vital capacity, and 13.8 modified Rodnan skin score). Mean probabilities of changing from one pattern to another in the whole lung in the CYC and MMF groups, respectively, were as follows: from LF to GG (15% and 12%), from LF to NL (25% and 22%), and from GG to NL (39% and 37%). Mean overall improvement in transitioning from GG to NL or LF to NL was significant for both treatments (all p<0.001), with no between-treatment differences. A significant correlation was found between transitional scores and changes in the skin scores (rho=-0.32; p=0.0016) Conclusions: Using voxel-by-voxel transitional scores on paired HRCT, we were able to quantitatively trace patterns of changes over 2 years and found significant changes in ILD patterns, including favorable transitions from GG to NL and LF to NL, in patients undergoing treatment for SSc-ILD. These findings demonstrate that quantitative radiographic measurements can be used to monitor treatment response in SSc-ILD trials.