Background Currently, interstitial lung disease (ILD) related to systemic sclerosis (SSc) is assessed with pulmonary function tests (PFTs) and chest Computed Tomography (CT). FVC <70% and DLco <70% were proposed as parameters to define remarkable ILD . Recently, quantitative assessment of ILD extention on chest CT was suggested as reliable parameter for disease assessment . Quantitative score is based on voxel-wise analysis of lung density, summarized in the following quantiatative CT (QCT) parameters: kurtosis (Kurt), skewness (Skew), mean lung attenuation (MLA), standard deviation (Sdev) and fibrosis ratio (FR).
Objectives To test the correlation between QCT parameter and PFTs and report the most accurate QCT for ILD assessment. Furthermore, to descibe QCT parameters in patients with FVC and DLco above or below the suggested threshold.
Methods 226 SSc patients fullfilling ACR/EULAR diagnosis criteria undewent chest CT and PFTs in six different hospitals. All CTs were processed with an open-source DICOM-viewer (OsiriX)  that provided FR and the other QCT parameters (i.e. Kurt, Skew, MLA, Sdev) both related to normal lung parenchyma (nQCT) and to total lung (tQCT). Spearman rank test was used to verify the correlations between QCT parameters and PFTs data. The Mann-Whitney test was used to assess differences between patients with FVC and DLco above or below the suggested threshold. QCT parameters discriminative performances were verified using ROC analysis. A p-value <0.05 was considered significant.
Results Among QCT parameters, nKurt showed the best correlation with FVC (r =0,535; p<0,0001) and DLco (r =0,394; p<0,0001). Notably, nKurt was significantly lower in patients with FVC and DLco <70% (p<0.00001). ROC analysis showed that nKurt =1,69 can discriminate very well patients with FVC <70% (sensibility 67,0%, specificity 81,0%). Similarly nKurt =5,25 distiguishes subjects with DLco <70% (sensibility 80,8%, specificity 53,6%).
Conclusions QCT parameters correlate with PFTs as reported in literature. In particular, nKurt showed the strongest correlation with FVC and DLco. Furthermore, the proposed thresholds of nKurt could be very useful in clinical practice allowing furhter characterization of ILD associated with SSc.
Khanna D, et al. Arthritis Rheum 2005;52(2):592–600.
Ariani A, et al. Rheumatol Int 2013.
Rosset A, et al. J Digit Imaging 2004;17(3):205–216.
Disclosure of Interest None declared