Unlocking the Potential of Chest MRI: Strategies for Establishing a Successful Practice

Abstract

Chest MRI is a valuable tool for assessing chest structures, particularly when CT produces inconclusive results. MRI provides exceptional soft-tissue resolution and enables the determination of lesion location, size, and invasion into neighboring structures. Its applications span various clinical scenarios, including the differentiation of non-tumorous and tumorous conditions in the mediastinum or pleura, planning of surgical interventions and treatments for such tumors, evaluation of post-treatment recurrence, staging of lung cancer, and diagnosis of progressive massive fibrosis. Despite the technical hurdles posed by cardiac and respiratory motion, advancements in sequence and scan techniques have enabled high-quality chest MRI examinations to be conducted across diverse clinical settings. This pictorial essay aims to offer comprehensive resources and strategies for radiologists to integrate chest MRI into clinical practice and to overcome its present challenges.

INTRODUCTION

MRI is a valuable imaging modality owing to its benefits, which include the absence of ionizing radiation, exceptional soft tissue contrast, and the potential to obtain functional data. However, despite these advantages, chest MRI is not used as often as CT for diagnosing thoracic diseases, primarily because several obstacles hinder its implementation in clinical practice. Key challenges include inherent physical limitations, such as proton deficiency, magnetic field inhomogeneity, and motion artifacts, and operator-associated challenges, such as limited thoracic radiologist expertise and MRI interpretative confidence, insufficient chest MRI training, the absence of standardized protocols, and the different sequences recommended by the manufacturers. However, the introduction of advanced pulse sequences for dedicated MRI in the first decade of this century has notably improved its image quality and lesion detection capabilities for thoracic diseases (1). Consequently, there is growing expectation regarding the diagnostic potential and value of chest MRI.

This pictorial essay aims to offer comprehensive resources and strategies for radiologists to integrate chest MRI into clinical practice and to overcome its present challenges.

POTENTIAL BARRIERS TO IMPLEMENTING CHEST MRI INTO PRACTICE AND THE BEST STRATEGIES TO OVERCOME THESE BARRIERS

There are several potential barriers to the use of chest MRI in clinical practice; however, MRI techniques are currently available to overcome and reduce these barriers.

LOW PROTON DENSITY OF THE LUNGS

Lungs have a lower proton density than other soft tissues in the chest wall and mediastinum because they comprise only around 800 g of tissue and accommodate 4–6 L of blood. Furthermore, the presence of air in the lungs leads to local field inhomogeneities at the tissue-air and liquid-air interfaces, thus causing T2* shortening (2,3). Consequently, the lung parenchyma typically appears dark or nearly black in most MRI sequences.

Emerging techniques, such as ultrafast echo time (UTE) or zero echo time (ZTE) MRI, offer potential solutions for signal loss in the lung parenchyma (4,5,6). During UTE imaging, the dephasing of spins between excitation and acquisition is significantly reduced, and susceptibility artifacts are absent. Moreover, 3D radial projection-based UTE techniques, such as Pointwise-Encoding Time reduction with Radial Acquisition (PETRA) or 3D spiral UTE Volumetric Interpolated Breath-hold Examination (VIBE), have now been utilized (7,8). During ZTE imaging, readout gradients are activated before RF excitation, enabling simultaneous signal excitation and encoding, thus effectively reducing the TE to zero with no susceptibility artifacts and a shorter image acquisition time. In addition, ZTE imaging can reduce acoustic noise by eliminating the need for rapidly switching gradients between repetition time intervals.

RESPIRATORY MOTION ARTIFACT

Minimizing the impact of respiratory motion is essential to ensure the interpretability of chest MRI images. Employing breath-holding techniques during image acquisition can effectively reduce breathing-induced respiratory motion artifacts (Fig. 1). Typically, healthy individuals can sustain breath-holding for approximately 20 s; thus, it is advisable to limit the acquisition time for chest MRI sequences to optimize image quality. However, several adjustments can be made for patients unable to sustain breath-holding for 20 s, or when sequences cannot be confined to this duration. One option involves dividing the tissue volume into smaller stacks to facilitate a shorter acquisition time. Alternatively, increasing the number of concatenations per sequence allows for the distribution of slices across multiple repetitions, thereby enabling multiple shorter breath-holding periods. In addition, radial k-space sampling can reduce respiratory motion artifacts by oversampling and continuously updating the center of the k-space by overlapping spokes (Fig. 2). Techniques like periodically rotated overlapping parallel lines with enhanced reconstruction (PROPELLER) (9) and StarVIBE (10) leverage radial sampling to produce motion-insensitive images. These methods can be combined with radial k-space sampling to enhance the image quality and reduce the impact of respiratory motion. Respiratory gating offers a solution when patients are unable to hold their breath (5,8,11). This technique involves acquiring images at specific phases of the respiratory cycle, commonly at end expiration, and can be achieved using an external pneumatic belt or a respiratory-navigated MRI sequence. When respiratory-navigated MRI sequences are used, imaging information is acquired only when the diaphragm dome is located within a predetermined acceptance window. The patients were instructed to remain still and to breathe comfortably and evenly before initiating the procedure.

Fig. 1. Chest MR image in a 53-year-old female.

A, B. Respiratory motion artifacts on the coronal T1-weighted turbo spin echo image (A) disappear on the coronal T2-weighted single shot fat spin echo image (B) with cardiac gating and breath holding.

Fig. 2. K-space sampling method. Cartesian sampling uses frequency- and phase-encode gradients to scan k-space line-by-line, thus collecting signal samples until reaching the opposite edge. Radial sampling, unlike Cartesian, oversamples the k-space center and uses overlapping spokes, wherein all spokes equally contribute to the image. The PROPELLER method, a type of radial sampling, collects data in multiple parallel lines (blades) that rotate incrementally until the entire k-space is covered.

CARDIAC MOTION ARTIFACT

Cardiac motion can also adversely affect chest MR images by blurring the structures adjacent to the heart owing to its distribution throughout the k-space during sequence acquisition. Furthermore, cardiac motion and related vascular pulsations can propagate throughout the acquired slice in the phase-encoding direction, thereby causing phase-ghosting artifacts. To address this issue, both prospective and retrospective gating techniques are employed (12,13). Prospective gating involves acquiring k-space data during a selected portion of the R-R interval, which enables the generation of motion-free image stacks. On the other hand, retrospective gating involves acquiring k-space data throughout the R-R interval, and subsequently selecting data from the desired segment of the cardiac cycle after data acquisition.

PULSE SEQUENCES UTILIZED IN CHEST MRI PROTOCOLS

In the past, 1.5T MRI was predominantly used to evaluate non-cardiac thoracic diseases because MRI with a magnetic field strength of ≥3T was prone to signal loss due to susceptibility artifacts, despite its higher signal-to-noise ratio (SNR). However, recent technological advancements have broadened the clinical applications of high-magnetic-field MRI, offering higher SNRs, greater tissue contrast and temporal resolution, shorter imaging times, and various novel fat suppression techniques (e.g., CHESS, SPIR, and SPAIR) (14,15).

Chest MRI protocols can be broadly categorized into the following: evaluations of anatomical and morphological abnormalities, assessments of functional abnormalities, and evaluations of structural motion. The details of the commonly used imaging techniques are listed in Table 1.

Table 1. Pulse Sequences for Chest MRI.

Sequences	Vendor Specific Techniques	Plane	Slice Thickness	ECG	Respiration	FS	CE
T2W single shot fast SE	SSFSE, HASTE, UFSE	Axial/coronal	≤5	+	Multiple BH	-	-
T2W STIR/T2W SPAIR	FAST STIR, TURBO STIR, STIR TSE/SPAIR, ASPIR	Axial/coronal	≤3	+	Multiple BH	+	-
T2W radial acquisition	PROPELLER, BLADE, MultiVane	Axial/coronal	≤5	+	RG	-	-
T1W SE	TSE, FSE	Axial/coronal/sagittal	≤5	-	FB	-	-
In- and opposed-phase T1W GRE	In- and opposed-phase	Axial	≤5	-	BH	-	-
T1W GRE with DIXON	LAVA FLEX, DIXON, mDixon	Axial	≤5	-	FB	+	-
Proton density 3D ultrashort GRE	ZTE 3D or 4D, 3D UTE spiral VIBE	Axial	≤3	-	RG	+	-
ECG-gated bSSFP	FIESTA, TrueFISP, BFFE	Four chamber view	≤5	+	BH	-	-
Real time bSSFP	FIESTA, TrueFISP, BFFE	Coronal	≤5	-	FB	-	-
T1W 3D fast GRE-pre-contrast	LAVA, StarVIBE, THRIVE	Axial/coronal/sagittal	≤3	-	FB with RG or BH	+	-
T1W 3D fast GRE-post-contrast (with or without DCE)	LAVA, StarVIBE, THRIVE	Axial/coronal/sagittal	≤3	-	FB with RG or BH	+	+
DWI	SS- or MS-EPI DWI, DWI-HASTE	Axial	≤3	+/-	FB or multiple BH	-	-

BH = breath-hold, CE = contrast enhancement, DWI = diffusion-weighted imaging, ECG = electrocardiogram, FB = free breathing, FS = fat-saturated, GRE = gradient echo, RG = respiratory gating, SE = spin echo, T1W = T1-weighted image, T2W = T2-weighted image, ZTE = zero echo time

SEQUENCES FOR ANATOMICAL/MORPHOLOGICAL ABNORMALITIES

The following pulse sequences were used to evaluate the anatomical structures and tissue characteristics of the chest in terms of the T1 and T2 properties.

T2-WEIGHTED IMAGING

T2-weighted imaging offers exceptional anatomical clarity and is valuable for characterizing lesions, particularly for assessing the presence of internal fluid or edema (Fig. 3). A single-shot fast spin-echo sequence (SSFSE) with half-Fourier acquisition is optimal for T2-weighted chest imaging because it enables rapid image acquisition and offers the potential for cardiac gating (16). Furthermore, incorporating the double inversion recovery technique into SSFSE can eliminate blood pool signals and facilitate dark blood imaging. For fat-suppressed T2-weighted imaging, short tau inversion recovery (STIR) can be added to the SSFSE sequence (17). A combined technique using radial sampling with an FSE image (PROPELLER or BLADE) can produce images that are less sensitive to motion artifacts (9).

Fig. 3. T2-weighted images obtained using the HASTE (left), BLADE (middle), and STIR (right) sequences in a 74-year-old male with a thymic cyst. A 15-mm sized, well encapsulated, anterior mediastinal nodule of high signal intensity is seen on the T2-weighted images.

T1-WEIGHTED IMAGING

T1-weighted imaging (T1-WI) can serve two purposes. First, it offers non-fat-suppressed T1-weighted images with a high SNR, which aids in assessing lesion signal characteristics, identifying hemorrhages, and delineating anatomical details such as fat planes between lesions and adjacent structures. Second, it facilitates microscopic fat-suppressed T1-WI for the detection of intravoxel lipids (microscopic fat) (Fig. 4) (18). Utilizing a dual-echo technique, in which one breath-hold captures precisely co-registered in- and opposed-phase data with identical imaging parameters, except for echo time, is preferred. This technique enables the computation of the lesion signal intensity index and chemical shift ratio (18). Dixon fat-suppressed gradient echo (GRE) or turbo spin echo (TSE) T1-WI provides additional means of obtaining chemical shift images of the thorax (19,20). By leveraging the chemical shift principles, this technique achieves uniform fat suppression by exploiting the different precession rates of water and fat molecules. By simultaneously acquiring both in- and opposed-phase images, mathematical combinations produced four sequences: in-phase, opposed-phase, fat-only, and water-only (Fig. 5). Water-only images can serve as fat-suppressed images and enable the evaluation of micro- and macroscopic lipids in a single breath-hold, particularly when using the DIXON technique.

Fig. 4. A 59-year-old male with a thymic hyperplasia.

A. Axial contrast-enhanced CT image shows soft-tissue lesions (arrows) with a contour bulging configuration in the anterior mediastinum.

B. The anterior mediastinal lesion shows a signal decrease on the opposed-phase images of both the single breath-hold dual echo (left upper, in-phase; right upper, opposed-phase) and DIXON techniques (left lower, in-phase; right lower, opposed-phase) of T1-weighted gradient echo sequence, thus demonstrating the presence of microscopic fat tissue.

Fig. 5. A 16-year-old female with a mature teratoma.

A. The anterior mediastinal mass has a heterogeneous high signal intensity with internal septa on a T2-weighted image obtained using the HASTE sequence.

B. T1-weighted gradient echo image obtained using the DIXON technique (left upper, in-phase; right upper, opposed-phase; left lower, water only; right lower, fat only) shows fat components (arrow) within the mass.

C. T1-weighted gradient echo image obtained using the VIBE sequence (left, pre-contrast T1-weighted gradient echo image with fat saturation; right, post-contrast T1-weighted gradient echo image with fat saturation) shows enhancement of the solid portion (arrows) within the mass. Inset: The enhancement of the solid portion is ambiguous on contrast-enhanced axial CT.

BALANCED STEADY-STATE FREE PRECESSION

Balanced steady-state free precession (bSSFP) is a sequence that provides excellent SNR and rapid acquisition times, and includes trueFISP, FIESTA, and bFFE (Fig. 6) (21,22). This imaging technique does not rely solely on pure T1 or T2 weighting, but rather on the tissue T2*/T1 ratio. In SSFP images, fluids or blood appear with a high signal intensity, thus making it useful for vascular evaluation, particularly when contrast agents cannot be used. It can also be used in cine-based protocols designed to capture respiratory and cardiac motions.

Fig. 6. A 42-year-old male with a mediastinal lipoma.

A. Non-enhanced axial chest CT image shows a well-defined nodule with fat density (arrow) in the retro-cardiac area of the mediastinum.

B. Cine-based balanced steady state precession image with breath-hold and retrospective electrocardiogram gating shows a 1.5-cm sized, well-encapsulated nodule (arrow) of high signal intensity in the mediastinal retrocardiac area. No cardiac motion artifact is seen.

C. T1-weighted gradient echo images (left, fat only image; right, water only image) obtained using the DIXON technique demonstrate that the mediastinal nodule is composed entirely of fat.

PRE-AND POST-CONTRAST IMAGING

Fat-saturated 3D fast GRE sequences are commonly employed to generate pre- and post-contrast images because of their ability to deliver high spatial resolution images that provide robust anatomical depictions and enable lesion detection (Fig. 7) (3,17,23). The timing methods used for contrast bolus administration of post-contrast 3D fast GRE sequences include the standard arterial phase for 20–30 s and the portal venous phase for 60 to 70 s, followed by 3- and 5-min acquisitions.

Fig. 7. A 70-year-old male with a non-invasive thymic epithelial tumor (WHO classification B2). Pre- (left) and post-contrast (right) T1-weighted gradient echo images obtained using the VIBE sequence show a well-encapsulated mass with homogeneous enhancement.

SEQUENCES FOR FUNCTIONAL ABNORMALITIES

Functional protocols such as diffusion-weighted imaging (DWI) and dynamic contrast enhancement are techniques used to measure tissue-specific parameters.

DIFFUSION-WEIGHTED IMAGING

DWI is used in clinical practice for thoracic diseases to enhance lesion conspicuity, characterize tissues, and evaluate treatment response. However, DWI in thoracic MRI is technically challenging due to respiratory and cardiac motions as well as susceptibility artifacts from air-tissue interfaces (24,25). However, advancements in single-shot spin-echo echo planar imaging have accelerated image acquisition despite inherent geometric distortions, and improved segmentation techniques, gradient systems, and spectral fat suppression have enhanced the thoracic DWI image quality (25,26). DWI provides measures of water molecule diffusion by analyzing the signal decay between opposing gradient pulses, which are quantified using b values. Although at least two b value sequences are required to calculate an apparent diffusion coefficient (ADC) image, the National Cancer Institute Consensus Conference on DW MRI in Cancer stated that imaging at least three b values should be performed, including low (b = 0 sec/mm²), intermediate (b ≥ 100 sec/mm²), and high (b ≥ 500 sec/mm² and ≤ 1000 sec/mm²) values for quantification (Fig. 8) (27). However, while additional b values can enhance the measurement accuracy, they prolong the scan time, thus necessitating the balancing of requirements for accuracy and efficiency during the DWI protocol design.

Fig. 8. A 61-year-old male with thymic carcinoma.

A. An axial T2-weighted image obtained using the HASTE sequence shows a relatively well-encapsulated mass with internal high signal intensity foci in the anterior mediastinum. Note the small left pleural effusion.

B. Diffuse weighted image (left, b value 800 sec/mm²; right, ADC image) demonstrates diffusion restriction within the mass. The ADC value was 0.696 × 10⁻³ mm²/s.

ADC = apparent diffusion coefficient

DYNAMIC CONTRAST ENHANCED MRI

Dynamic contrast-enhanced (DCE) MRI involves acquiring sequential MR data sets at various time points regarding the injection of intravenous contrast, thus enabling the generation of time-intensity curves to assess enhancement patterns (28). High-temporal-resolution DCE data provide quantitative parameters, such as time-to-peak, maximum relative signal enhancement, wash-in and wash-out degrees, Ktrans (a measure of the rate of blood transfer from the plasma to the extracellular space), Kep (a measure of the rate of blood transfer from the extracellular space to the plasma), Ve (the volume of extracellular space inside a voxel), and Vp (the volume of plasma space inside a voxel), which are crucial for evaluating blood flow and tissue characteristics. 3D T1-weighted spoiled fast GRE sequences are commonly used for DCE data acquisition during chest MRI (Fig. 9) (29,30). Advancements in imaging techniques, such as radial k-space acquisition and compressed sensing, have improved image quality by enhancing temporal resolution and mitigating motion artifacts (31). Furthermore, integrating DCE data with DWI provides a comprehensive multiparametric approach for lesion characterization, particularly before and after treatment (32).

Fig. 9. A 68-year-old female with a mediastinal schwannoma.

A. Contrast-enhanced axial CT image shows a well-defined mass (arrow) with central faint enhancement in the middle mediastinum.

B. An axial T2-weighted image obtained using the HASTE sequence shows a well-defined mass with heterogeneous signal intensity in the middle mediastinum. The mass exhibits the target sign, characterized by a central area of low signal intensity surrounded by a T2 hyperintense rim.

C. Pre- (left) and post-contrast enhanced (right) T1-weighted gradient echo images obtained using the StarVIBE sequence display intense enhancement at the center of the mass.

D. The time-intensity curve of dynamic contrast enhancement at the center of the mass.

SEQUENCES FOR STRUCTURAL MOTION

Cine-based protocols, including retrospective gating sequences and real-time cine, were designed to capture respiratory and cardiac motions.

RETROSPECTIVE GATING SEQUENCES

Retrospective gating sequences can aid in evaluating the lesions near the heart, assessing the relationships between intraluminal thrombi or tumors and vessel walls, and confirming the intravascular artifacts observed in other MRI sequences or CT scans (Fig. 6). Although typically based on bSSFP, spoiled GRE sequences can also be utilized (13,15).

REAL-TIME CINE SEQUENCES

Chest MRI real-time cine sequences require the acquisition of a 2D volume with a relatively high temporal resolution, typically without electrocardiogram gating. Specifically, 2D bSSFP sequences are well-suited for this application. These sequences enable motion assessment in the chest under free-breathing conditions and include the evaluation of diaphragm function and tumor adherence to relevant structures, such as the pleura, diaphragm, chest wall, spine, and aorta (Fig. 10) (17,33,34).

Fig. 10. A 79-year-old female with a localized, fibrous, pleural tumor invading the left hemi-diaphragm.

A. Contrast-enhanced axial T1-weighted image with VIBE sequence shows a lobulated enhancing mass attached to the left anterior hemi-diaphragm (arrows).

B. Real-time cine images with a bSSFP sequence show that the mass and the left hemi-diaphragm move together when the patient breathes.

CLINICAL INDICATIONS OF CHEST MRI

Chest MRI can be utilized for the diagnosis of various chest conditions, including abnormalities of the chest wall, diaphragm, pleura, and mediastinum, lung cancer screening and staging, and disorders of the aorta and pulmonary vasculature (1,6,17,35,36).

CHARACTERIZATION OF MEDIASTINAL NON-TUMOROUS LESIONS

Chest MRI is useful for characterizing mediastinal non-tumorous lesions, such as mediastinal cysts (thymic, bronchogenic, and pericardial cysts) (Figs. 11, 12), thymic hyperplasia (Fig. 4), and hemorrhage (17,36,37). Bronchogenic cysts exhibit various signal intensity patterns on T1-WI owing to their diverse contents, which include protein, hemorrhage, and mucoid material, and often display an air-fluid level (Fig. 11) (38). Subtraction techniques on post-contrast T1-weighted images may aid in the visualization of contrast enhancement and provide accurate registration of pre- and post-contrast images without significant motion artifacts (Fig. 11) (37).

Fig. 11. A 60-year-old female with a bronchogenic cyst.

A. Contrast-enhanced coronal CT image shows a 3-cm sized, paravertebral mass (arrow) without definite enhancement.

B. T2-weighted image obtained using the BLADE sequence shows a 3-cm sized, paravertebral mass of high signal intensity. Little motion artifact is seen, even in a free-breathing state.

C. Axial T1-weighted image obtained using the VIBE sequence shows an air-fluid level within a cystic mass.

D. Gadolinium-enhanced T1-weighted images obtained using the VIBE sequence (left) and the subtraction technique (right) show no mass enhancement.

Fig. 12. A 71-year-old female with an esophageal duplication cyst. Gadolinium-enhanced T1-weighted image obtained using the StarVIBE sequence and respiratory gating shows a 4-cm sized, cystic mass (arrow) with peripheral enhancement in the posterior mediastinum. No motion artifact is observed in the free-breathing state.

For thymic hyperplasia, employing T1-weighted gradient-echo in- and opposed-phase MRI sequences can assist in lesion detection and differentiation based on the detection of signal intensity reduction on opposed-phase images (Fig. 4) (39).

The appearance of a hemorrhage on MRI depends on the age of the blood product. During the subacute phase, hemorrhage typically appears as T1 hyperintensity, whereas more chronic foci may exhibit T1 and T2 hypointense rims due to hemosiderin deposition.

DIFFERENTIAL DIAGNOSIS OF MEDIASTINAL TUMORS

Chest MRI assessments of the enhancement patterns and signal intensities of internal components aid in differentiating mediastinal tumors (36,40). Notably, mature teratomas typically exhibit heterogeneously high signal intensities with internal septa on T2-weighted images (Fig. 5) (38). T1-weighted GRE sequences obtained using the DIXON and VIBE techniques can be used to evaluate the solid portions of teratomas. Detecting fat components involves observing signal reductions in opposed- versus in-phase images, while the solid portions demonstrate contrast enhancement. Other masses containing fatty components, such as lipomas, thymolipomas, and liposarcomas, may be suspected when a fat component is evident on T1-DIXON fat and water images of the mediastinal masses (Fig. 6).

Imaging findings of thymic epithelial tumors depend on the risk group (41). Low-risk thymomas typically exhibit smooth borders, complete capsules, and homogeneous enhancement (Fig. 7). Conversely, high-risk thymomas or thymic carcinomas may exhibit lobulated or irregular borders, necrotic or cystic changes, calcification, lymph node enlargement, or greater vessel invasion (Fig. 8). However, both the risk groups demonstrated enhancement on contrast-enhanced T1-weighted images. Recently, the malignancy of thymic epithelial tumors has been evaluated by assessing DWI ADC values or enhancement patterns of DCE images (42,43).

Mediastinal tumors can be differentially diagnosed based on their location, characteristic findings, internal components, and enhancement patterns (44). Neurogenic tumors are commonly located in the paravertebral space, which houses the central canal and intervertebral foramen, and contains neural tissue. The target sign, characterized by a central area of low signal intensity surrounded by a T2 hyperintense rim, was indicative of a peripheral nerve sheath tumor (Fig. 9). This pattern is usually associated with localized neurofibromas and histopathologically corresponds to a dense central area of collagenous stroma surrounded by myxomatous tissues (45). In addition, widening of the intervertebral foramina suggested a potential neurogenic tumor (Fig. 13).

Fig. 13. A 67-year-old male with a neurogenic tumor. Pre- (left) and post-contrast enhanced (right) axial T1-weighted gradient echo images obtained using the VIBE sequences depict a paravertebral mass with homogeneous enhancement and widening (arrows) of the intervertebral foramen.

SURGICAL AND TREATMENT PLANNING OF THORACIC TUMORS

Chest MRI plays a crucial role in the surgical and treatment planning of thoracic tumors (17). It aids in the identification of potential biopsy sites by detecting areas of contrast enhancement and diffusion restriction. Additionally, chest MRI enables the evaluation of the disease extent based on assessments of direct invasion and fat plane obliteration between the tumor and adjacent structures.

EVALUATION OF POST-OPERATIVE OR-TREATMENT RECURRENCE

Chest MRI can also play an important role in assessing post-operative or -treatment recurrence (Fig. 14). Usually, locally recurrent masses can be observed on T2 HASTE Half Fourier Single Shot Turbo Spin Echo), DWI, and T1-VIBE-FS-CE sequences.

Fig. 14. A 60-year-old male with a synovial sarcoma.

A. Contrast-enhanced CT image obtained 12 months after surgery shows left pleural thickening (arrow) and effusion.

B. Axial T2-weighted fast spin echo image obtained on the same day as (A) shows a mass (arrow) with heterogeneous high signal intensity in the diaphragmatic pleura.

C. Coronal contrast-enhanced T1-weighted image with VIBE sequence shows a lobulated mass (arrow) with heterogeneous enhancement.

STAGING OF LUNG CANCER, INCLUDING THAT OF APICAL TUMORS

Chest MRI is a valuable tool for staging lung cancer, particularly lung apical tumors, because of its exceptional ability to detect soft tissue involvement. This technique is optimal for evaluating lung apical tumors and tumor resectability based on lesion extent (46). Tumors that invade structures such as the brachial plexus, vertebral bodies, and vascular structures are classified as T4 tumors. Regarding the invasion of the brachial plexus, involvement of more than the lower trunk or C8 nerve root is typically deemed inoperable, and invasion of the C8 nerve root, trunk level of the brachial plexus, and 1st rib can be visualized on sagittal T1-weighted TSE images. Various MRI sequences are utilized to assess brachial plexus invasion, such as non-fat-suppressed sagittal T1-weighted TSE for nerve delineation, fat-suppressed sagittal T2-weighted TSE for tumor identification, fat-suppressed coronal-T2-weighted TSE for highlighting neurovascular structures, and contrast-enhanced fat-suppressed axial/sagittal/coronal T1-weighted-TSE for lesion enhancement (Fig. 15) (47). In addition, 3D STIR SPACE was used for multi-planar and curved planar reconstruction (48). During lung cancer T staging, MRI was more accurate than CT for evaluating mediastinal invasion (49), and dynamic cine MRI detected pleural invasion. For N staging, STIR TSE and DWI (often combined with PET-CT) can be used to detect thoracic metastatic lymph nodes (50,51).

Fig. 15. A 69-year-old male with a lung apical tumor.

A. Contrast-enhanced CT shows an ill-defined mass (arrows) suspicious for invasion of the right subclavian artery.

B. Axial T1-weighted image shows an ill-defined, lobulated mass (arrows) with iso-signal intensity.

C. Contrast enhanced sagittal T1-weighted image shows an ill-defined lung apical tumor invading the right subclavian artery (arrow) and right 1st rib (arrowhead). The middle and lower trunks of the right brachial plexus are not visible due to tumor invasion.

D. Multi-planar reconstructed coronal T2-weighted image obtained using the STIR SPACE sequence highlights the tumor and neurovascular invasion.

DIAGNOSIS OF PROGRESSIVE MASSIVE FIBROSIS

Chest MRI can help differentiate lung cancer from progressive massive fibrosis (PMF) (52). In contrast to lung cancer, PMF appears as a low signal intensity on T2-WI (Fig. 16). Furthermore, contrast-enhanced T1-WI may show persistent rim enhancement (53).

Fig. 16. A 61-year-old male with a progressive massive fibrosis.

A. Lung window image of non-enhanced chest CT scan shows a mass-like consolidation (arrow) in the right upper lobe posterior segment. Note the subpleural and centrilobular nodules with interlobular septal thickenings in both upper lobes.

B. Axial T2-weighted image obtained using the BLADE sequence shows a mass-like consolidation (arrow) of low to intermediate signal intensity in the posterior segment of the right upper lobe.

C. The mass (arrow) does not exhibit diffusion restriction on diffusion-weighted images (b value, 800 s/mm²).

D. Dynamic contrast-enhanced T1-weighted gradient echo image (left, pre-contrast image; right, post-contrast image at 5 mins after contrast injection) obtained using the VIBE sequence demonstrates persistent gradual enhancement of the mass for up to 5 mins after contrast injection.

OTHER DISORDERS AMENABLE TO THE CLINICAL APPLICATION OF MRI

In clinical practice, there may be occasions when there is insufficient evidence to support the use of MRI or when CT has better diagnostic accuracy than MRI. However, MRI can also be used effectively in several clinical scenarios.

LUNG NODULE DETECTION AND DIFFERENTIAL DIAGNOSIS

Similar to CT, chest MRI can detect lung nodules and enable evaluation based on their appearance. The overall detection rate of MRI was 82.5%, as compared to 97.0% for multidetector CT, wherein MRI and CT malignant nodule detection rates were similar (54). In a comparative study on the effectiveness of 3T MRI combined with UTE, low-dose CT, and standard-dose CT, no notable differences were observed in nodule detection or between nodule type assessments (ground-glass, part-solid, and solid nodules) (Fig. 17) (55). 3T MRI with a UTE sequence is also useful for assessing nodule morphology. In a study evaluating the size, attenuation, margins, and internal lucencies of pulmonary nodules as determined by CT and UTE MRI, Wielputz et al. (56) concluded that UTE-determined nodule sizes were 1–2 mm smaller than those determined by CT. On the other hand, UTE MRI had greater sensitivity for assessing factors predictive of malignancy for nodules detected by CT.

Fig. 17. A 65-year old male with a fibrocalcified pulmonary tuberculosis. Contrast-enhanced axial T1-weighted images with the ultra echo time sequence (right) show small nodules (yellow arrows) <5 mm in size, appearing similar to those seen on the CT image (left).

DIFFERENTIAL DIAGNOSIS OF MEDIASTINAL LYMPH NODES

STIR turbo spin echo images are effective for assessing mediastinal lymph nodes. These images improve tissue contrast, particularly for malignant lymph nodes, as both T1 and T2 relaxation times increase, and have better diagnostic accuracy (85%–92%) than CT and PET-CT for N staging (50). In addition, DWI is useful for N staging and has a high diagnostic accuracy of 80%–95%, outperforming PET-CT (51).

DIFFERENTIAL DIAGNOSIS OF PLEURAL LESION

Chest MRI is comparable to CT for evaluating pleural disease (Fig. 18) and can differentiate and characterize pleural effusions, including exudates, transudates, and hemothoraces. In addition, 3T MRI combined with DWI and DCE sequences can provide a more accurate diagnosis in patients suspected with malignant pleural effusion (57,58). Coolen et al. (57) reported that DWI can differentiate benign and malignant pleural effusions (sensitivity, 71.4%; specificity, 100%), and that adding DCE-MRI to inconclusive DW findings improved the accuracy from 87.1% to 93.5%.

Fig. 18. A 67-year-old female with a pleural tuberculosis.

A. Contrast-enhanced axial CT image shows a pleura-based nodule (arrow) with mild peripheral enhancement.

B. Pre- (left) and post-contrast enhanced (right) axial T1-weighted images show the nodule with central necrosis and peripheral homogeneous enhancement.

OPTIMIZING CHEST MRI PROTOCOLS FOR CLINICAL PRACTICE

Chest MRI examinations are often tailored to address specific clinical questions arising from prior CT examinations to facilitate efficient targeted imaging (15). Anatomic coverage can be adjusted based on lesion location, thereby reserving the whole thoracic coverage for staging or for cases without known lesions. Essential sequences are typically T1-weighted, T2-weighted, and bSSFP, with provisions for identifying macroscopic fat and considering intravoxel fat using in- and opposed-phase imaging, if necessary. Contrast administration, if indicated, enables the evaluation of lesion and vascular enhancements, although contraindications such as pregnancy must be considered. Functional imaging, such as DWI, aids in lesion characterization and treatment response evaluation, whereas DCE can provide quantitative data on perfusion kinetics. In contrast, cine imaging, particularly electrocardiogram-gated and real-time cine imaging, provides insights into cardiac and diaphragmatic functions as well as the proximities of lesions and vascular structures.

Taking into account these technical considerations, chest MRI protocols should aim to minimize examination time by selecting appropriate pulse sequences based on specific clinical needs (15).

CONCLUSION

MRI is a valuable tool for assessing chest structures, particularly when CT produces inconclusive results. MRI provides exceptional soft-tissue resolution and enables the determination of lesion location, size, and invasion into neighboring structures. Its applications span various clinical scenarios, including the differentiation of non-tumorous and tumorous conditions in the mediastinum or pleura, the planning of surgical interventions and treatments for such tumors, and the evaluation of post-treatment recurrence. Despite the technical hurdles posed by cardiac and respiratory motions, advancements in sequencing techniques have enabled high-quality chest MRI examinations to be conducted across diverse clinical settings.