MRI Quantification of Pulmonary Fibrosis in Scleroderma Patients

Overview

Assessment of pulmonary fibrosis is currently based on high-resolution CT (HRCT) and pulmonary function tests (PFT) such as forced vital capacity, (FVC) and carbon monoxide diffusion (DLCO). These techniques allow a semi-quantitative analysis of the pulmonary disease but are imperfect. The mains weaknesses are the lack of reproducibility, the limited sensitivity and for CT the resulting radiation dose. Recent advances in MRI sequences allow exploring the lung parenchyma with millimeter slice thickness. Development of computer-assisted post-processing such as elastic registration opens new perspectives in the functional study of the lung parenchyma, especially the analysis of its deformation during the respiratory cycle and therefore of its elasticity. Pulmonary involvement in scleroderma is present in 70 to 100% of patients and is the leading cause of death. Initial assessment of pulmonary involvement and follow-up are important for therapeutic decisions and patient prognosis. Quantitative analysis should be developed to reliably evaluate pulmonary fibrosis and increase the reproducibility. The purpose of our study is to evaluate the feasibility of quantifying pulmonary fibrosis by successively performing full inspiration then full expiration volumetric MR acquisitions using a VIBE - Volumetric Interpolated Breath-hold examination sequence. Post processing of the 2 volumes using elastic registration is performed to evaluate pulmonary deformation in the normal and fibrotic lung areas, hypothesizing that it would be different.

Scleroderma is a chronic connective tissue disease of poorly understood origin. The prevalence ranges between 100 and 260 cases per million inhabitants in Europe and the United States, with a female predominance (3/1). Pulmonary involvement is frequent, observed in 70 to 100% of the patients. It is the leading cause of death in scleroderma. It has mainly two forms: diffuse fibrotic interstitial pneumonitis, which occurs in the majority of cases in the form of non-specific interstitial pneumonia (76%) and pulmonary arterial hypertension (PAH). PAH in scleroderma may be secondary to pulmonary fibrosis or develop on its own, in patients with no parenchymal involvement. CT and pulmonary function tests play a very important role in the detection and follow-up of patients with lung disease associated with scleroderma. However, the analysis of HRCT, looking for ground glass opacities, reticulations, traction bronchiectasis, predominantly sub-pleural, basal and posterior remains semi-quantitative and shows great inter-observer variability, when based on visual assessment. Diagnosis and staging of chronic lung diseases, such as scleroderma, is a major challenge for both patient care and approval of new treatments. Magnetic resonance imaging (MRI) of lung disease may provide a non-invasive evaluation of lung fibrosis, bypassing radiation dose concerns of CT. The goal of the study is to develop a quantitative imaging biomarker based on MRI for the assessment of lung disease severity and monitoring disease progression. Fibrotic changes increase lung stiffness and induce a restrictive lung disease with decreased lung volumes on PFT. Thus, the investigators hypothesized that the assessment of lung elasticity would allow differentiating between fibrotic and healthy lung parenchyma. For each subject, they will examine an overall "elasticity index" (based on the total disease volume), as well as the spatial distribution of elastic properties. These properties will be calculated using a deformable registration between expiratory and inspiratory MR images. For the representation of lung tissue expansion they will examine several surrogates based on the determinant of the Jacobian of the deformation field. The objective is to evaluate a cohort of 30 patients, divided into 3 groups according to the GAP criteria: * 10 patients with mild pulmonary fibrosis (VFC\> 75% theoretical and DLCO / VA\> 55%) * 10 patients with moderate pulmonary fibrosis (VFC 50-75% and DLCO 36-55%) * 10 patients with severe pulmonary fibrosis (VFC \<50% theoretical or DLCO / VA \<35%). They will be compared to a control group of 10 healthy subjects. Inclusion and non-inclusion criteria are developed in a next section. All patients will be recruited in Cochin Hospital, during their annual follow-up. The procedure will be performed at the end of a cardiac MRI, performed as part of patient standard follow-up. A 12-channel surface antenna will be installed for the acquisition of the cardiac and pulmonary MRI images. VIBE (Volumetric interpolated breath-hold examination) T1 sequences, of 17 second duration (TR: 2.73 ms TE: 0.05 ms, Tilt angle: 5.5) will be acquired following a deep inspiration then a deep expiration. Elastic registration of the 2 sequences will be performed using an in-house software, developed at CentraleSupelec (Center for Visual Computing). The registration will allow determining a global lung elasticity index and separating healthy from fibrotic lung areas. Marks will be manually placed on pre-defined anatomical targets on inspiratory and then expiratory MR images in order to check the quality of automated registration. This will be done by two independent observers, and the distance between marks placed by the 2 observers will be compared to the results of automated registration. Clinical data for all patients of the cohort will be collected, including clinical history, sex, age, association to other systemic disease, positivity of biological markers, and ongoing treatment. These data will be retrieved from the patient medical charts at Cochin Hospital. MRI images will also be used to calculate the expiration and inspiration lung volumes in order to compare them to the volumes acquired during PFT. The data will be prospectively acquired, with only one evaluation for each patient. Patient data collection and image acquisition will start after the study approval by the ethics committee (Comité de Protection des Personnes, CPP) and the National Commission on Informatics and Civil Liberties (CNIL).. The strength of the Jacobians (deformation forces) when performing elastic registration will be compared to the FVC, by using Spearman correlation coefficient.