Semi-quantification image methods for assessing severity of scleroderma-associated interstitial lung disease according to computed tomography data

Abstract

The aim of this study was to compare three semi-quantification scales for prospective assessment of scleroderma-associated interstitial lung disease (SS-ILD) severity. Methods. From 110 prospectively followed patients with SS-ILD, we selected 12 patients (mean age, 42 ± 13 years, 11 females) with obvious improvement (n = 6) or worsening (n = 6) of lung lesions on high resolution computed tomography (HRCT) during a year. The patients had diffuse (n = 7) or limited (n = 7) SS with mean length of the disease of 8.5 ± 6.7 years (range, 1 to 23 years). HRCT was done at baseline (inclusion in the study) and in a year. CT scans were quantitatively assessed by four radiologists including one experienced radiologist. A blinded analysis of HRCT scans was done using three scales: J.H.Warrick et al. (1991), A.U.Wells et al. (1997), and E.A.Kazerooni et al. (1997). The intraclass correlation coefficient (ICC) was calculated to evaluate the assessment reliability. T-test for independent samples was used to evaluate reproducibility of the assessments. Agreement between independent experts' opinions was evaluated using Kendall's rank correlation coefficient. Results. The measurements were significantly divergent between the radiologists, both for the baseline and the follow-up HRCT scans. ICCs for investigated radiological parameters were 0.56 to 0.76. The highest ICC (0.76) was obtained for A.U.Wells' scale. All scales used to assess HRCT scans had lower interoperator reproducibility. Conclusion. Combined use of currently available semi-quantification methods for follow-up assessment of HRCT in SS-ILD patients allowed thorough qualitative evaluation of lung lesions, but the reliability of the radiological parameters in detecting 1-year fibrosis progression in SS patients was low. The risk of significant interoperator bias limited the use of the radiological parameters in clinical trials of SS-ILD patients.