|ECR 2019 / C-2418|
|HRCT features in patients with undifferentiated connective tissue disease|
We reviewed CT scans obtained in 22 patients with a clinically and serologically diagnosis of UCTD; they underwent an HRCT examination because of their respiratory symptoms and impairment of pulmonary function test results. All of them showed a parenchymal lung involvement, particularly the finding of interstitial lung disease with different patterns.
It is commonly known that the most of ILD associated with CTD is characterized by the radiologic and histopathologic pattern of NSIP, especially in systemic sclerosis, Sjӧgren’s syndrome and dermatomyositis/polymyositis. Otherwise, the UIP pattern has been more frequently reported with rheumatoid arthritis. 
In our study, NSIP CT pattern was observed in less than half of the patients, while most of them showed UIP CT pattern; it thus suggests that the predominant CT pattern in UCTD patients is UIP pattern. Our results support those of Kim et al., who concluded that in patients with ILD and UCTD the UIP pattern was as frequent as or more frequent than NSIP pattern, on the basis of radiologic and/or pathologic findings. 
Furthermore, ViJ et al. reported that 62% of patients with autoimmune-featured interstitial pneumonia had a typical UIP pattern on CT images. 
As found in our study, CT UIP pattern is characterized by the presence of reticular opacities and traction bronchiectasis with a subpleural and basal predominance; honeycombing in the posterior lung zones is critical for making a definite diagnosis.  (Fig.5)
Our results are discordant from those reported by previous studies, which described the NSIP pattern as the predominant pattern in UCTD patients. For instance, Kinder et al. showed that the majority (88%) of patients classified as idiopathic NSIP (histopathologic pattern) met the criteria for UCTD; it points out that this pattern is peculiar of this disease.  Several authors also described the common occurrence of UCTD in NSIP biopsy-proven. [12, 13]
On the other hand, the diagnostic criteria for UCTD are different, and Kinder et al. used broader and less specific criteria compared with a stricter and accepted definition of UCTD according to Mosca et al. Thus, it might result in an overestimation of UCTD. 
In the current study, the most common CT findings of UCTD–related lung disease were traction bronchiectasis and reticulation related to fibrosis. These abnormalities were observed numerically equal in each CT pattern, with an overall prevalence of 91% and 86% respectively. (Fig. 6)
These results are in contrast with Kinder et al., who found that patients with UCTD-ILD were significantly more likely to have GGO on HRCT.  It probably correlates with the predominant CT NSIP pattern described by Kinder et al.
Other common findings were airways abnormalities, including parietal thickening and calcification, reported in 15 (68%) of 22 patients. Honeycombing was frequently seen in 16 (73%) patients; it was observed in all UIP CT pattern with a bilateral basal predominance, and in one patient with NSIP CT pattern with an apical distribution. (Fig.7)
As previously reported in association with CTD-ILD, dilatation of the pulmonary artery and lymph nodes enlargement were common accessory thoracic findings.
Concerning the extent of parenchymal abnormalities, we observed that during a mean 5 year follow up period the finding of honeycombing had the highest progression rate in 13 (76,5%) patients. It is well known that histopathological UIP patients had progressive nature irrespective of UCTD diagnosis. 
In our study, the interobserver agreement for determination of all CT findings was good to excellent, except for a moderate agreement for evaluation of GGO. It is probably due to difficulty in discrimination between “pure” ground-glass opacity caused by acute inflammation and hazy increased lung attenuation superimposed on a fine reticular pattern, related to fibrosis.  Moreover, interobserver agreement is dependent on observer experience. (Fig.8)
A limitation of our study is that pathologic findings from surgical lung biopsy were not available. However, the diagnosis of usual interstitial pneumonia (UIP) can be made with HRCT in patients with concordant clinical presentation without the need for tissue confirmation. A typical UIP pattern on HRCT is highly accurate for the presence of UIP pattern on surgical lung biopsy. [15, 16]
In conclusion, our current study findings suggest that UIP pattern is frequently seen in patients with UCTD and the most common parenchymal abnormalities are traction bronchiectasis and reticulation.
As previously reported in several studies, accurate identification of pulmonary involvement in UCTD patients has important therapeutic and prognostic implication because of the better prognosis of patients with CTD-ILD than that of idiopathic pulmonary fibrosis (IPF). 
Thus, a multinstitutional prospective study with a larger number of patients should be conducted in the future. Further studies should confirm our findings.
Thematically related posters
ECR 2019 / C-2288
Usual interstitial pneumonia pattern: differential diagnosis and treatment possibilities