摘要
Magnetic resonance imaging (MRI) of the lung was first attempted in the 1980s, and W. Richard Webb published the first comprehensive review of this topic in the first issue of the Journal of Thoracic Imaging 25 years ago.1 In that review, the basic knowledge of clinical applications and the specific challenges of MRI of the thorax were presented. The low field strength of 0.35 T available at that time was particularly favorable for MRI of the lungs. Webb focused on the visualization of mediastinal diseases, such as vascular lesions and masses, including the pulmonary hila, and abnormalities in the lung, such as nodules, mucous plugs, vascular lesions and consolidations. Although the potential of MRI was promising, the shortcomings were obvious, especially when compared with computed tomography (CT). In the following years, a few researchers strived to improve the performance and quality of MRI of the lungs; however, several general difficulties could not be overcome, including: low proton density equaling low signal, susceptibility artifacts secondary to multiple air-tissue interfaces, and motion artifacts because of respiration, and cardiac and vascular pulsation. Articles on MRI of the lung comprised descriptions of MRI findings of various pulmonary and mediastinal diseases, but did not generate a sufficient knowledge base, clinical impact, or benefits for patients. In addition, the advent of high-resolution CT and spiral CT increased the lead of CT over MRI, and MRI of the lung did not reach the clinical arena. Subsequently, several novel technological developments fostered new attempts to improve the quality and broaden the application of pulmonary MRI.2 These developments are described in the following paragraphs. First, paramagnetic contrast agents were introduced and opened the field of magnetic resonance angiography (MRA) in general and MRA of the thoracic aorta and the pulmonary arteries in particular.3 With the following improvement in gradient strength and the introduction of parallel imaging techniques, fast and ultrafast slice and volume acquisitions became feasible. Today, MRA of the pulmonary arteries with high spatial resolution, which is applied for diagnosis of suspected pulmonary embolism or pulmonary hypertension (Fig. 1), and multiphasic MRA with high temporal resolution for the depiction of vascular territories and pulmonary perfusion, are feasible.4 High-quality studies have shown the clinical impact of contrast-enhanced MRA in the diagnosis of acute pulmonary embolism, and it now serves as a second-line diagnostic modality when there are contraindications to CT or in instances in which CT is inconclusive. MRA has even been shown to be useful as a first-line imaging modality in the diagnosis of pulmonary hypertension secondary to recurrent thromboembolic disease.5 In patients with chronic pulmonary emboli, MRA shows dilation, abrupt caliber changes and cutoffs, thromboembolic wall thickening, and webs.6 MRA also provides conclusive results regarding delineation or vascular infiltration of mediastinal and lung tumors. Multiphasic MRA is also accepted as a primary imaging modality for the assessment of congenital or acquired pulmonary vascular anomalies, especially in the pediatric population, for the evaluation of abnormalities such as anomalous pulmonary venous return and arteriovenous malformations, among others.7 Perfusion MRI is a useful adjunct in the diagnosis of a variety of lung diseases and their consequences on pulmonary function and gas exchange.8 The image quality and the volumetric data set with high spatial and temporal resolution provided by MRI challenges the traditional nuclear medicine techniques to assess pulmonary perfusion.9 Similar to perfusion scintigraphy, perfusion MRI has successfully been applied to preoperatively calculate the postoperative forced expiratory volume in one second (FEV1) for patients scheduled to undergo lung resection. Perfusion MRI can also be targeted at the characterization of masses to assess malignancy and angiogenic activity.FIGURE 1.: A 20-year-old patient suffering from pulmonary arterial hypertension. The coronal acquired magnetic resonance angiography (20 s breath-hold) shows an absence of subpleural pulmonary artery branches (A). The transverse reformation image (B) shows the markedly enlarged main pulmonary artery and the abrupt caliber reduction from the central to the peripheral branches. The perfusion dataset (maximum intensity projection) is very sensitive for the detection of the inhomogeneous distribution of perfusion as well as visualization of the small patchy perfusion defects, characteristic for pulmonary arterial hypertension (C).Second, breath-hold imaging as well as prospective and retrospective gating allowed for a rather robust acquisition of high-resolution images in T1-weighting precontrast and postcontrast, T2-weighting, short tau inversion recovery (STIR), etc. The variety of different sequences allows for a detailed analysis of the internal structure and contents of cystic and solid mediastinal tumors. “Structural” MRI has clear advantages over CT and has been accepted as the imaging modality of choice for the delineation and characterization of mediastinal tumors.10 Together with direct acquisition in coronal or sagittal orientations, the excellent soft tissue contrast of MRI proved superior to CT in the evaluation of resectability of lung cancer, especially with regard to infiltration of the lung apex (Pancoast tumor), chest wall, and mediastinum (Fig. 2). In the meantime, state-of-the-art imaging of lung cancer has shifted from CT of the chest to whole-body positron-emission tomography/CT, bringing together morphological and metabolic information. Recently, MRI has also become a whole-body imaging modality with improved sensitivity and specificity because of diffusion-weighting.11 For the staging of lung cancer, the specific advantages of MRI over positron-emission tomography/CT in the detection of lesions with a high degree of conspicuity lead to an additional clinical benefit because of the typical pattern of metastatic spread of lung cancer: brain, bone, and liver. “Structural” MRI also detects the different manifestations of cystic fibrosis, including bronchiectasis, bronchial wall thickening, mucus plugs, small airway inflammation, and destruction (Fig. 3) as well as CT, even in free-breathing infants.12FIGURE 2.: A 42-year-old male patient with a central bronchial carcinoma invading the mediastinum. (A, B) T2-weighted turbo-spin-echo sequences with fat-saturation in coronal and transversal orientation. The tumor bulk is clearly visualized and the close contact to the trachea, ascending and descending aorta is shown. After contrast media application (C, D) the tumor enhanced in the periphery with a central necrosis (T1-weighted gradient-echo sequence with fat-saturation).FIGURE 3.: A 7-year-old girl with complete destruction of the middle lobe due to cystic fibrosis as shown in coronal and transverse T2-weighted images (A, B). The left lower lung shows typical bronchial wall thickening and bronchiectasis (arrow). (C, D) A 14-year-old girl with cystic fibrosis. T1-weighted postcontrast coronal and transverse images showing mucus plugging in the right upper lobe. The bronchial wall enhances significantly (arrow), whereas the mucus shows no enhancement.Third, dedicated MRI techniques were developed to directly visualize pulmonary ventilation.13 These methods include different tracer gases for imaging ventilation in a steady-state equilibrium phase using pure oxygen or fluorinated gases or during or after a single breath using the sophisticated hyperpolarized noble gases 3Helium and 129Xenon.14–17 Dedicated protocols provided much more than just images of ventilation, but a broad spectrum. Soon, researchers observed that these tracer gases could also be used to probe pulmonary function and microstructure. Diffusion-weighted MRI allows measurements of alveolar size, measurements of T1-relaxation or chemical shift for assessment of gas exchange, uptake, and alveolar partial pressure of oxygen. Perfusion and ventilation imaging served as the starting point for the introduction of “functional” MRI of the lung.18,19 This field was boosted by adapting the technologies originally developed for cardiac MRI for the assessment of blood flow in the pulmonary arteries, shunt volumes between the pulmonary and systemic vasculature, as well as cardiorespiratory interaction, especially right heart compromise. Again, pulmonary hypertension emerged as the major clinical indication benefiting from these developments. An additional novel application is the ultrafast volumetric MRI during continuous breathing, either tidal or forced. Several goals can be pursued with this technique. First, it can be used to assess diaphragmatic motion, paradoxical movement, and relaxation, which were formerly investigated by fluoroscopy. Second, it may be used to evaluate chest wall motion and respiratory mechanics in general. Dynamic MRI is especially suited in the preoperative workup of the funnel chest (pectus excavatum).20 The assessment of respiratory mechanics has also been successfully applied to preoperatively assess hyperinflation, such as in patients with severe emphysema, and to later monitor the positive effects of lung volume reduction surgery or recent minimally-invasive techniques with similar goals. Third, it may be applied to the assessment of motion of lung cancer. Recent innovations in radiotherapy aim at the escalation of dose to cure lung cancer, but spare the adjacent normal lung parenchyma to prevent radiation-induced pneumonitis. Comprehensive knowledge about the motion of the tumor and the lung in an individual patient and close monitoring will be required.21,22 MRI might play an important role in this emerging new field. After the substantial progress in the last decade, MRI of the lung is now ready for prime time in the clinical arena. An easy-to-use standard protocol has been introduced.20,23 It is robust, fast (less than 30 min), and solves the common diagnostic queries in routine clinical practice (Table 1). MRI offers a number of advantages when compared with conventional nuclear medicine techniques and is challenging multidetector CT in many applications. Lack of ionizing radiation places MRI of the chest in a front-line position for all cross-sectional imaging studies in children. The unique combination of structural and functional information makes MRI attractive in all diseases in which the choice between innovative and expensive treatment options will actually require and benefit from an increased number of measurable parameters, including reproducible quantitative read-outs.24TABLE 1: Indications for MRI in Pulmonary Disease Beyond the Heart When Compared With CT, Nuclear Medicine, and Ultrasound TechniquesFurther improvements can be envisaged for MRI of the lung. For example, the visualization and characterization of inflammation represent a major challenge (Fig. 4). Three main goals should be addressed: (1) differentiation between infectious and noninfectious inflammation; (2) estimation of inflammatory activity in pneumonitis, such as pulmonary fibrosis, granulomatosis, and pulmonary manifestations of collagen vascular disease; and (3) airway inflammation in chronic bronchitis, chronic obstructive pulmonary disease, and asthma. Although these goals are indeed challenging, MRI seems much more likely to succeed in these efforts than multidetector CT.FIGURE 4.: A 2-year-old girl with complicated pneumonia and pleural effusion with multiple septa (T2-weighted images, A coronal and B transverse orientation). Especially in children, the possibility of imaging the lung without using ionizing radiation is very appealing. This patient underwent thoracoscopy to remove the septa and to place a draining tube. During the operation, bronchoscopy was performed, and a diagnosis of pulmonary abscess (arrow) was confirmed.