Systemic sclerosis (SSc) is characterized by microvascular injury, inflammation and immune system activation, and an abnormal tissue deposition of extracellular matrix [1]. Inflammation and vascular injury, leading to vasculopathy, are the early cornerstones, thereafter, resulting in fibrosis of internal organs and their functional failure [2,3,4]. These events occur also in the lung, resulting in interstitial lung disease (ILD) [5, 6]. In SSc, ILD is a severe organ complication [7] and chest computed tomography (CT) is the most suitable, rapid, reliable technique for its evaluation and follow-up [8]. Moreover, the most recent CT scans can guarantee high spatial resolution also with low-dose acquisitions in a single breath-hold [9]. However, low-dose single CT examinations may suffer from a “cumulative dose”, especially in the case of young patients, during the long-standing follow-up. It is well known that CT has some limitations, due to the risk of contrast agent administration and to the lack of real “functional” evidence, like ventilation and perfusion. In fact, dynamic or perfusional CT acquisition is possible but burdened by higher dose exposure like pneumo-CT and perfusional studies.

In order to decrease radiation burden during prolonged repetitive follow-up in young patients, lung ultrasound (LUS) was recently proposed both for screening and follow-up in ILD-SSc patients. LUS has shown high sensitivity for ILD detection by assessing the presence of B lines, vertically extending hyperechoic artifacts originating from the pleura (described as “comet tails”), that represents an indirect sign of ILD [10,11,12]. LUS has also been used to assess the severity of ILD by correlating the number of B lines with the extent of ILD on HRCT [11] and with the change in diffusing capacity for carbon monoxide [13]. Despite LUS being a non-invasive, low-cost, easily learned technique, the lack of methodological standardization, the incomplete field of view, and the low specificity limit its use in clinical practice. Moreover, LUS cannot detect ground glass opacities (GGO) and this is an important limit, in particular because GGO results as one of the earlier sign of ILD in SSc patients. According to the Fleischner Society glossary, GGO is defined as “hazy increased opacity, with preservation of bronchial and vascular margins” [14]. Usually, GGO is characterized by areas of hazy increased attenuation of the lung with preservation of bronchial and vascular margins and can represent either interstitial or alveolar processes [15]. On this background, it should be underlined that GGO represents an important limit at CT examination, as well. In fact, CT can detect very well GGO but cannot differentiate if it is due to an inflammatory or fibrotic condition.

Actually, from a histopathologic standpoint, GGO may be the result of partial filling of alveolar spaces or thickening of the alveolar walls/septal interstitium, or a combination of both. This pattern may be generated either by early inflammatory or late fibrotic changes of the lung parenchyma. For this reason, in the clinical setting, the interpretation of the GGO may be tricky, because it is not possible to definitively define GGO as mainly inflammatory or fibrotic. This fact may significantly limit the clinician in the choice of the therapeutic strategy (immunosuppressants vs antifibrotics).

Therefore, despite that CT is the golden standard for ILD evaluation, still today GGO remains a clinical challenge. On CT, partially matching histopathologic features, GGO may be due to the reduction of air in the alveolar airspaces, partial filling of the alveolar airspaces, thickening of the parenchymal interstitium and alveolar walls, a relative increase in perfusion, or a combination of these factors. In order to fill up this gap, MRI may be seen as a powerful imaging tool, a radiation-free technique which may provide a discriminatory insight about the inflammatory or fibrotic nature of GGO.

Previously, MRI of the pulmonary parenchyma has been considered as a contradictory evaluation due to the extremely low proton density of lung tissue, to the magnetic susceptibility of sharp air/parenchyma interfaces and respiratory motion artifacts.

Recently, the new ultra-rapid MR sequences, such as the ultrashort echo-time (UTE), may overcome these limitations, counteracting lung parenchyma signal decay T2 star related [14,15,16]. Even if the geometric resolution of CT remains greatly superior to that of MRI, data from comparative studies have shown that, in ILD-GGO, UTE MRI provides comparable results to CT. Moreover, the speed of this kind of sequence, compared with the usual longer MRI acquisitions, may increase patient’s tolerability to the exam and reduced costs. UTE are acquired during free breathing conditions with fully automatic respiratory synchronization and isotropic voxels, in a scan time of about 8 min, and allow for both GGO detection and extent evaluation in every plan, by the multiplanar reconstructions permitted by isotropic voxels, with low incidence of major artifacts.

Usually, free breathing scans are obtained at functional residual volume, and therefore it is easily understandable that high-resolution breath-hold scans that match the inspiratory level and resolution of CT are highly desired. These techniques have been also combined to fast acquisition strategies, such as Compressed-Sensing (CS) Volumetric-Interpolated-Breath-hold Examination (VIBE), to obtain breath-hold imaging. CS acquisitions allow the collection of the entire lung volume in breath-hold acquisitions of about 10 s. However, breath-hold scans duration may be increased up to 20 s, to improve both robustness towards respiratory motion and temporal resolution. Though, it was experienced that 20 s of apnea may be poorly tolerated by the great part of the non-early stage SSc-ILD patients and are scarcely reliable for ILD evaluation. At the moment, research groups are testing UTE and CS-VIBE sequences in SSc-ILD and it seems the better results are obtainable by free-breath UTE acquisitions [17, 18].

One of the first papers to investigate the role of MRI in assessing ILD on SSc patients was written by Pinal-Fernandez et al. [19] used a 2D BH half-Fourier single-shot TSE sequence (HASTE) in eighteen SSc patients. This sequence showed high sensitivity to detect SSc-ILD and was correlated with lung functional parameters (forced vital capacity, and diffusing capacity of the lung for carbon monoxide) and also with HRCT. However, MRI examination’s extent were consistently lower than HRCT and, thus, not directly comparable. These data confirmed those by Ohno et al. [20] who showed that mean T2 values are different when comparing connective tissue disease ILD and healthy subjects, with significant correlation with diffusing capacity of the lung for carbon monoxide and severity of ILD on HRCT. Later, also Gargani et al. [21] confirmed the potential role of lung MRI in the detection of ILD in SSc patients, analyzing STIR and T1 mapping sequences (before and after gadolinium injection) in 30 SSc patients. They also evaluated MR signal intensity in STIR and T1-weighted sequences in anatomical sections corresponding to normal lung parenchyma and ILD at CT scans reporting a significant difference between normal and pathological ILD lung areas. Moreover, STIR values are also able to predict worsening of ILD over time, independently of HRCT appearance. Miller et al. [22] tried to differentiate GGO with active pulmonary disease from non-active pulmonary disease, with only fibrotic lesions in 24 SSc patients, indeed they documented only a high agreement between MRI and CT.

Today, the main clinical question is whether MRI may differentiate inflammation from fibrosis in the context of GGO. Theoretically, at MRI, a predominantly inflammatory GGO rich in fluids (exudate) and vessels might give a different imaging than fibrosis, which is characterized by tissue rich in collagen and poor in fluid and vessels. In fact, MRI is very sensitive either in assessing the aqueous content or in recognizing fibrosis. For example, diffusion-weighted images (DWI) may give information on fluid presence and its microscopic motion (influenced by surrounding tissue components), as well as on microcirculation [23]. As far as we know, no data in the literature address this problem even if several attempts (also by us — Figs. 1 and 2) are ongoing.

Fig. 1
figure 1

Early inflammatory interstitial lung disease in SSc. CT (A) and MR (B, UTE acquisition; C, DwI acquisition; D, ADC map) studies. Early inflammatory involvement (black arrows and oval shaped area) with tiny GGO aspects, well demonstrated by CT (A). It can be seen that conspicuity of the involvement at UTE (B) acquisition is even better than that at CT. In the same area, DwI (C) shows medium intensity signal, which corresponds to an elevated value (around 2000 · 10-6 mm2/s) on the ADC map (D). This can agree with a high liquid presence in the flogistic lesions

Full size image
Fig. 2
figure 2

Fibrotic interstitial lung disease in SSc. CT (A) and MR (B, UTE acquisition; C, DwI acquisition; D, ADC map) studies. Fibrotic involvement (oval shaped area) with some GGO aspects, demonstrated very well by CT (A) and quite well by UTE (B) acquisition. In the same area, DwI (C) shows medium intensity signal, which corresponds to a quite low value (1200–1300· 10-6 mm2/s) on the ADC map (D). This can agree with a low liquid presence in the fibrotic tissue

Full size image

The administration of the contrast agent could enhance GGO assessment as well as the presence of vessels, but several problems await a solution together with the resting state (RS), the low signal-to-noise ratio (SNR), the artifacts inherent in the presence of air, and the neighboring heartbeat. In the clinical setting, MRI might also detect not only an inflammatory GGO but more likely a “mixed” GGO, where inflammation and fibrosis overlap in different percentage according to the disease activity and evolution. The problem of identifying patients with a progressing phenotype is today of paramount importance in SSc-ILD [5]. Before a large-scale prospective study is designed, a retrospective study on cases defined as predominantly fibrotic or predominantly inflammatory is needed. The comparison with the results of the biopsy is still unlikely, due to its rare use.

In conclusion, while CT remains the gold standard for ILD diagnosis and assessment in the next future MRI may solve the problem of the interpretation of GGO, thus unraveling inflammation and allowing a targeted treatment and a follow-up with an x-ray sparing approach.