The Core Curriculum: Cardiopulmonary Imaging, 1st Edition (2004)

Chapter 2. Imaging Modalities and Applications

Over the past three decades, the introduction of new imaging modalities has had a profound impact on the practice of cardiopulmonary radiology. Many of the new modalities have not only complemented but replaced the classic imaging arsenal (1). For example, conventional planar tomography and bronchography have been replaced by computed tomography (CT) with multiplanar reconstructions. No other imaging modality matches the level of detail of the lungs provided by CT. Dynamic imaging of the heart was markedly changed with the introduction of echocardiography and is currently again being transformed by cardiac magnetic resonance imaging (MRI) and CT. Fluoroscopy has become an almost completely lost art. In this chapter, we introduce the main imaging modalities in clinical use and elaborate on their principles and applications.

Chest Radiograph

Its origins going back to 1895, the plain radiograph still represents the cornerstone of modern cardiopulmonary radiology. Depending on the practice setting, around 30% to 60% of all radiologic examinations performed in the clinical practice of radiology are plain radiographs of the chest. These are simple, quick, cheap, and provide useful clinical information for diagnostic and follow-up purposes. The major indications for chest radiography are listed in Table 2.1 (2).

The single most commonly performed radiologic examination is the chest radiograph.

The basic principles of chest radiography are relatively simple (Fig. 2.1); a patient is placed between an x-ray source and a detector. A point source emits photons with energies in the x-ray range toward the patient. These photons penetrate the patient, are attenuated (mostly by scattering and absorption), and those that come out of the patient are captured by the detector. This generates a single projection image of attenuation properties of the patient onto the detector (Fig. 2.2). Lungs contain mostly air and thus show little attenuation of the photons traversing them. Other tissues, like the heart, great vessels, and chest wall, are denser and show greater attenuation of the x-ray beam.

Radiation in the diagnostic range is 60 to 150 keV.

The point source is an x-ray tube that emits radiation in the “diagnostic” x-ray range, which is between 60 and 150 keV. Photons with energies at the upper end of that range have better penetrance in the body and are therefore less attenuated. Photons with energies at the lower end of the range are more attenuated by the soft tissues and osseous structures of the chest. Tube voltage, measured in peak kilovoltage (kVp), determines the maximum energy of the emitted photons. Tube current, measured in milliampere (mA), combined with exposure time in seconds, determines the total number of photons emitted. Both kVp and mA are controlled by the operator. Typical values for a chest radiograph are 120 kVp and 5 mA.

Table 2.1: Indications for Chest Radiograph

Diagnostic
   Cardiopulmonary symptoms
      Cough, hemoptysis, shortness of breath, chest pain, etc.
   Preoperative for thoracic surgery
   Preoperative if known cardiopulmonary limitations
   Staging of thoracic tumors and extrathoracic malignancies
   Infection
      Pleural, parenchymal, mediastinal
Follow-up
   Previously diagnosed cardiopulmonary disease
      Pneumonia resolution to exclude endobronchial lesion
      Pulmonary edema
   Monitoring of intensive care unit patients
      Lung disease
      Pleural disease
      Lines and tubes positions
   Monitoring of postoperative patients

Figure 2.1 Principles of chest radiography. A source emits x-ray photons toward the patient. The radiation that comes out of the patient hits a detector. A grid is used between the patient and the detector to block scattered radiation.

Figure 2.2 Normal chest radiograph in the posteroanterior (A) and lateral (B) positions. The radiograph has wide latitude and can represent lung structures, cardiac and vascular structures, soft tissues of the chest wall, and ribs. Metallic nipple markers are used here.

Tube voltage (kVp) determines maximum energy of the x-ray photons.

The simplest type of detector is a plastic film covered with a light-sensitive emulsion, silver halide. Films alone have very limited sensitivity to x-rays; therefore, an amplifying screen must be placed against the film to produce multiple visible photons for each incident x-ray. Plain films can display a wide range of attenuations (wide latitude). More modern detectors are digital in nature (3). In computed radiography, the detector is a plate of phosphor or selenium phosphor, which captures x-rays. Once exposed, the plate is scanned point by point by a laser beam. Each point emits a radiation of intensity that is quantified and is proportional to the x-ray energy captured at that specific point on the plate. Computed radiography offers greater latitude than the screen–film combination. In direct radiography, the plate is usually made of a grid of tiny elements called charged coupled devices (like sensors in most modern digital cameras) or thin-film transistors (like flat screens for computers). These elements can directly record and quantify the number of photons striking each point of the plate without the need to scan the plate point by point with a laser beam. It is therefore much faster to process than computed radiography. With all types of detectors, grids and filters are usually used to block scattered photons and eliminate photons of too low energy, respectively.

Tube current (mA) × time (seconds) determines the number of photons emitted (mA).

The position of the patient with respect to the detector can vary. In the most common technique, posteroanterior and lateral views of the chest are obtained (Fig. 2.2). The patient is placed in the upright position, 180 cm (6 feet) from the source, facing the detector so the radiation crosses the patient from the back to the front (posteroanterior) and then with the left side facing the detector (lateral). These two projections allow a gross three-dimensional estimate of the structures in the chest. In bedridden patients, the x-ray plate or detector is typically inserted under the patient, and the radiation crosses the patient in a front to back direction (anteroposterior). The distance between the source and the patient is often decreased to 100 to 125 cm, which produces a larger magnification of anatomic structures in the anterior aspect of the chest, particularly the heart and superior mediastinum (Fig. 2.3). If the patient is placed on a table lying on their side, a lateral decubitus view is obtained. This is useful to assess for a pleural effusion (Fig. 2.4) (4) or air trapping on the dependent side and pneumothorax on the nondependent side. Oblique views with the patient at an angle to the film can be obtained to eliminate superposed structures. A common angled view is the apical lordotic view, with photons crossing the patient from a low anterior to a high posterior position. This provides a better view of the lung apices (5) by projecting the clavicles above them (Fig. 2.5).

Decubitus images are useful for evaluation of pleural fluid (free flowing or not), pneumothorax (when patient cannot sit upright), and air trapping (failure of dependent lung to collapse).

Figure 2.3 Posteroanterior vs. anteroposterior radiograph. On the anteroposterior radiograph (A) of this normal patient, the detector is against the back of the patient. A combination of decreased distance between the source and the patient and increased distance between the detector and the anterior mediastinal structures compared with the posteroanterior radiograph (B) leads to magnification of the heart.

Figure 2.4 Lateral decubitus. The patient is positioned with their side against a table, and a radiograph is taken across the table. A free-flowing right pleural effusion is demonstrated (asterisk).

In most cases, the radiograph is taken with the patient in full inspiration. Several respiratory maneuvers can be performed by the patient immediately before radiographic exposure. A radiograph with the patient in expiration may demonstrate air trapping. Radiographs taken while the patient is performing a Valsalva or a Müller maneuver can also be taken to assess changes in vascular structures or masses.

Figure 2.5 Lordotic view. In this patient with a left apical neurofibroma, the abnormality is subtle on the posteroanterior radiograph(A), but the lordotic view (B) improves visualization of the lung apices, and the neurofibroma (asterisk) becomes more apparent.

Fluoroscopy, Conventional Tomography, and Bronchography

Fluoroscopic imaging is largely a dead or dying art. In the past it had been used to evaluate pulmonary nodules and mediastinal vasculature. It is still occasionally used to assess diaphragmatic motion (“sniff test”). This test involves dynamic evaluation of the simultaneous motion of the two hemidiaphragms during normal breathing and forced rapid inspiration (a “sniff”). If there is paralysis of one hemidiaphragm, it will not move as well as the contralateral part (Fig. 2.6) and may show paradoxical motion during rapid inspiration.

Figure 2.6 Bronchography, circa 1950. An endobronchial inhaled contrast agent was used to demonstrate the presence of bronchiectasis.

Bronchography is also an obsolete technique involving the administration of iodinated contrast via inhalation to coat the surface of the tracheobronchial tree and to evaluate the presence and extent of bronchiectasis. This has been completely replaced by high resolution CT.

Conventional tomography is an obsolete technique that used simultaneous movement of a source of x-ray and a detector in opposite directions with respect to a specific plane of imaging, to blur all the structures out of that plane of imaging. This has been replaced by CT (6).

Computed Tomography

Invented in 1972 by Hounsfield, CT is the imaging modality of choice for evaluating the thorax, after the chest radiograph, because the level of anatomic detail is unmatched by anything else in radiology. Indeed, the only other imaging modality that offers more extensive anatomic detail than CT is direct tissue examination under a microscope by a pathologist (which is much more invasive than CT). Note that whereas the in-plane spatial resolution of CT is 5 times lower than that of plain radiography (Table 2.2), the contrast resolution (the depiction of subtle differences in contrast) is 10 times greater than that of chest radiography. This improved contrast resolution, in combination with the tomographic nature of CT, enables CT to provide much better anatomic detail than plain radiography, despite the lower spatial resolution. The indications for chest CT are indicated in Table 2.3 (7).

After the chest radiograph, CT is the method of choice for most advanced imaging of the thorax.

The basic principles of CT are relatively simple (Fig. 2.7) and are an extension of plain film radiography. The point source of x-ray photons and the detectors are placed on opposite sides of the patient on a ring-like structure, called the gantry. The gantry rotates around the patient, located on a table at its center. Photons are emitted toward the patient, penetrate the patient, and are captured by one or more detectors. This generates a series of projection images of attenuation properties of the patient. Images representing photon attenuation at each point in the volume traversed by the photons can then be mathematically reconstructed from the different projections (Fig. 2.8).

Figure 2.7 Principles of computed tomography. The source of x-rays and the detectors are on opposite sides of the gantry with the patient at the center of the gantry. Radiation that crosses the patient is detected, producing a projection of attenuation information. By rotating the gantry around the patient, multiple projections are obtained, which are then used to mathematically reconstruct tomographic attenuation images.

Table 2.2: Spatial Resolution and Effective Radiation Dose of Thoracic Imaging Modalities

Modality

Resolution (mm)

Dose (mSv)

CXR

0.08

0.02 (PA)
0.04 (lateral)

DR

0.17

0.02 (PA)
0.04 (lateral)

CT

0.4

8
0.6–1.2 (screening)

MRI

1.0

0

Planar nuclear medicine

7.0

0.4 (Xe)–18 (Ga)

PET

3.0

7

Angiography

0.13

12

US

0.3

0

Background radiation

n/a

3 per year

CXR, chest radiography; DR, digital radiography; CT, computed tomography; MRI, magnetic resonance imaging; PET, positron emission tomography; US, ultrasound.
From Bushberg JT, et al. The essential physics of medical imaging, 2nd ed. Philadelphia: Lippincott, Williams & Wilkins, 2002, with permission.

Table 2.3: Indications for Thoracic Computed Tomography

Thoracic
   Further characterize CXR abnormality (e.g., nodule, mediastinal mass)
   Detection and follow-up of neoplastic disease (e.g., metastatic sarcoma, lymphoma)
   Characterization of lung nodules
      Benign vs. indeterminate
   Parenchymal lung disease (e.g., emphysema, interstitial lung disease, infection)
   Airway disease
      Central and peripheral airways
   Pleural disease
      Empyema, metastasis, mesothelioma
   Postsurgical complications
   Percutaneous biopsy guidance
   Localization for VATS
Cardiac
   Cardiac abnormalities on CXR
   Cardiac anatomy
   Coronary arteries
      Calcification, patency with CTA
      Aberrant coronary arteries
   Postcardiac bypass grafting complications
      Mediastinitis
Vascular
   Aorta: aneurysm, trauma, dissection, coarctation
   Pulmonary arteries: embolus, pulmonary hypertension
   Venous: SVC/brachiocephalic vein thrombus or obstruction

VATS, video-assisted thoracoscopic surgery; CXR, chest radiography; CTA, CT angiography; SVA, superior vena cava.

The source of radiation, the x-ray tube, rotates continuously around the gantry, using a slip-ring technology in more modern scanners. Collimation is used at the source. The x-ray tube voltage in CT is usually 120 kVp, similar to that of a plain chest radiograph. However, CT uses a longer exposure time, 1 second per gantry rotation, with combined tube current–exposure time value of 200 mA, which increases the radiation dose.

CT scanner technology has gone through several generations of refinements (8). Early scanners used incremental table displacement and produced one set of axial projections at each table position. Modern scanners use continuous table displacement while the gantry rotates and produce helical projections, with a stationary 360-degree array of detectors around the entire gantry. The latest scanners have multiple rows of detectors in the direction of slice thickness. An entire CT of the chest using a multislice CT scanner takes only a few seconds. The recorded projections are helical in nature, and planes of imaging are interpolated from the helical projection data. In single-row detectors, collimation determines slice thickness, whereas it is determined by detector configuration and reconstruction algorithm in multiple row scanners. The table travel per complete rotation of the gantry with respect to x-ray beam collimation is called the “beam pitch.” In multislice scanners, the table travel per complete rotation of the gantry with respect to the data slice thickness is called the “data pitch.”

Fast helical CT scanners can image the entire thorax in seconds, reducing respiratory motion artifact.

Beam pitch is table travel per gantry rotation divided by beam collimation.

Data pitch is table travel per gantry rotation divided by normal slice width. This is the preferred definition for multislice CT.

Figure 2.8 Computed tomography imaging. On a mediastinal window (A), the lungs are mostly black and the mediastinum and chest wall are emphasized. On a lung window (B), these structures are white and the fine structures of the lungs are emphasized.

Several postprocessing steps can be performed when the slices are reconstructed from the projections. Tissue-specific algorithms, such as “soft tissue” or “lung,” are available to more accurately reconstruct the CT projection data based on assumptions of x-ray photon energy and tissue attenuation. The projection data represent attenuation of multiple points through a volume of the chest. In practice, the projection data are used to reconstruct entire axial slices of the patient or only a subset of each slice, called the “field of view.” But one is far from limited to reconstructing axial images. One of the most powerful features of CT is the ability to reformat the image data into multiplanar and three-dimensional images. Axial and helical data can be reformatted to produce simple images in the coronal or sagittal planes (Fig. 2.9). To improve visualization, volumetric data can also be combined into a single image representing for each pixel the maximum or minimum value of that pixel through the volume (Fig. 2.10). Powerful algorithms can also produce three-dimensional images with shaded surface display (Fig. 2.11) or even endobronchial images (virtual bronchoscopy) (Fig. 2.12) (9,10).

The reconstructed images represent a range of attenuations of different points in a slice of the patient. CT image values are typically scaled to 4,096 different levels of attenuation recorded, measured in “Hounsfield units” (HU). By convention, air has a value of -1,000 HU and water a value of 0 HU. Hounsfield units in medical CT typically range from -1,000 and +3,000. Modern monitors cannot display the full range of attenuation values, because they are typically limited to 256 levels of gray. Moreover, the radiologist’s eyes can only distinguish between 30 to 90 levels of gray. The operator can choose to map any range of attenuation values into the displayed 256 levels of gray. The center of this range of displayed attenuation values is called the “window level” and the width of this range is called the “window width.” A soft tissue window typically displays attenuation levels in the range of -400 to +200 HU and emphasizes structures in the mediastinum, whereas a lung window displays attenuation levels in the range of -1,000 to -300 HU and emphasizes structures in the lung (Fig. 2.8).

Different window and level combinations are used to evaluate structures of different attenuation, such as lung, soft tissue, and bone windows settings.

Figure 2.9 Coronal and sagittal reconstructions. Multiplanar reconstruction of the helical projection data in the coronal (A) and sagittal(B) planes can be performed. This improves visualization of some structures, such as the lung apices and the great vessels.

 

In general, for cardiopulmonary imaging, the patients lay still in a supine position on the table and hold their breath for the duration of the CT acquisition. Scanning can be performed at end-inspiration and end-expiration and also in the prone or decubitus positions. Furthermore, intravenous contrast can be administered to the patient before scanning to emphasize vascular structures and to modify the attenuation properties of different parts of the body. For routine chest CT, contrast is injected at 2 mL/s, whereas for CT angiogram (CTA) contrast is injected at 4 mL/c.

CT angiography requires a faster rate of intravenous contrast administration than routine CT.

Figure 2.10 Maximal intensity projection reconstructions. Information from a stack of images representing a volume can be combined into a single image representing for each pixel the maximum value of that pixel through the volume, shown here in the coronal (A) and sagittal (B) planes.

Figure 2.11 Three-dimensional reconstructions. Data can be further processed to produce three-dimensional images with shaded surface of any chest structure, such as the heart, mediastinum, lungs or ribs.

Figure 2.12 Virtual bronchoscopy. Endoluminal views of the tracheobronchial tree, reconstructed from thin slice images of the thorax, improve visualization of this endotracheal papilloma (asterisk).

Figure 2.13 Lung nodule on computed tomography. The faint nodule projecting at the right lung base near the diaphragm (A) was further investigated by Computed Tomography, which revealed a calcified granuloma (B).

Different scanning protocols are used to answer different clinical questions (Figs. 2.13 and 2.14). Table 2.4 illustrates common CT protocols. In general, thicker slices (5 mm) provide enough anatomic detail for diagnosis of most cardiopulmonary disease. Thinner (1 to 1.5 mm) slices are used in three different contexts. For interstitial lung disease (11), thin slices are used with 10-mm spacing between slices. This covers only 10% of the chest but provides improved pulmonary parenchymal detail while minimizing radiation dose. For CT angiography, thin continuous slices are used to provide fine detail of the entire vascular tree of concern (12). For screening, thin continuous slices at very low doses are used to detect small lung nodules (13).

With multislice CT scanners and protocols involving thin continuous slices, radiation dose per examination has become a concern (14). Total effective dose for a chest CT equals up to 200 to 400 plain chest radiographs (15), which is quite high. The dose for a screening CT is, however, equal to only 30 to 60 plain chest radiographs (Table 2.2). Manufacturers are including several optimizations to their scanners to lower the dose, and scanning protocols are designed to use the lowest possible radiation doses that will provide an answer to clinical problems.

CT acquisition parameters should be optimized to obtain diagnostic images without unnecessary radiation exposure to patients.

Figure 2.14 High resolution computed tomography allows exquisite visualization of the fine detail of the lung parenchyma in this patient with Langerhan’s cell histiocytosis.

Magnetic Resonance Imaging

MRI for cardiopulmonary disease has been in clinical use since the 1980s (Fig. 2.15). It has proven extremely valuable to evaluate abnormalities of the chest wall, diaphragm, and mediastinum (Table 2.5) (16,17,17). MRI is also important for both anatomic and functional assessment of the heart, aorta, and pulmonary arteries (19,20). Its clinical value to assess lung disease is currently limited. Several experimental techniques are under active development, including pulmonary imaging using hyperpolarized gases (21).

MRI is particularly useful for evaluation of cardiovascular and posterior mediastinal structures within the thorax.

Figure 2.15 Magnetic resonance imaging of the chest. Magnetic resonance provides multiplanar imaging emphasizing different tissue contrast information. A T1-weighted sagittal image (A) provides excellent anatomic detail, whereas a T2-weighted image (B) can reveal increased water content, and thus increased signal, found in many pathologies.

Table 2.4: Thoracic Computed Tomography Protocols

Name

Slice Thickness (mm)

Pitch

Start

End

mA

Intravenous Position

Intravenous Contrast

Lung nodule

5

1.5–2

Lung apices

Lung bases

Low

Supine

No

1.25

1.0

Top of nodule

Bottom of nodule

Regular

Supine

No

Routine chest

5

1.5–2

Lung apices

Lung bases

Regular

Supine

Yes

Lung cancer

5

1.5–2

Lung apices

Adrenal glands

Regular

Supine

Yes

High resolution

1–1.5

10-mm spacing

Lung apices

Lung bases

Regular

Inspiration
Expiration
Prone
Supine

No

Trachea

1–1.5

1.5–2

7 cm below carina

Above epiglottis

Regular

Inspiration
Expiration
Supine

Yes

Pulmonary veins

1–1.5

1.5–2

Lung bases

Lung apices

Regular

Supine

Yes

SVC

2.5

1.5–2

Lung apices

Lung bases

Regular

Supine

Yes

Pulmonary embolus

1–1.5

1.5–2

2 cm below diaphragm

Lung apices + legs

High

Supine

Yes

Thoracic aorta

2.5

1.5–2

2 cm above arch

Diaphragm

Regular

Supine

Yes

Cardiac

1–1.5

1.5–2

Aortic

Diaphragm

Regular

Supine

Yes

Lung cancer screening

1–1.5

1.5–2

Apex

Lung bases

Low

Supine

No

SVC, Superior vena cava.

Only enough physics to understand the basic principles of MRI is discussed here (22). In terms of MRI, the atomic nucleus is a system that has two important properties (Fig. 2.16). First, the nucleus has a “magnetic moment” represented by a small magnetization vector along its axis. Second, the nucleus has an “angular momentum” or a “spin” that can be thought of as a rotation about its axis. Only nuclei with an odd number of protons or an odd number of neutrons possess angular momentum. When an object possessing a magnetic moment is placed under a strong external magnetic field (B0, typically 10,000 times greater than the earth’s magnetic field), it experiences a force to align it with the external magnetic field, like a compass. If the object also has angular momentum, it experiences a force called a “torque” that generates a motion of “precession” of its magnetic moment at a specific angle around the direction of the external magnetic field (or around its opposite direction), similar to the motion of a spinning top (Fig. 2.17A). The frequency of the precession, called the “Larmor frequency,” is directly proportional to the intensity of the external magnetic field around the object.

Table 2.5: Indications for Thoracic Magnetic Resonance Imaging

Thoracic
   Chest wall neoplasm (especially superior sulcus tumors)
   Mediastinal tumors (e.g., bronchogenic cysts)
   Lung parenchyma: limited, experimental
   Thoracic outlet and brachial plexus
Cardiac
   Congenital heart disease: shunts, complicated anatomy
   Myocardium
      Cardiomyopathy
      Ischemic disease
      Hypertension
      Right ventricular dysplasia
   Pericardium: thickening, effusion, tamponade, pericardial cyst
   Masses: thrombus, tumors
   Valves (limited): stenosis, regurgitation
Vascular
   Aorta: aneurysm, trauma, dissection, coarctation
   Pulmonary arteries: embolus, pulmonary hypertension
   Venous: SVC thrombus or obstruction

SVC, superior vena cava.

Figure 2.16 Magnetic properties of nucleus. A hydrogen nucleus has two important magnetic properties: a magnetic moment, represented by an arrow along its axis, and an angular momentum or spin.

Figure 2.17 Precession. A precessing magnetic moment (A) under an external magnetic field will generate a signal in a coil (B) that is too small to measure. It is the signal from trillions of precessing magnetic moments that is measured in practice.

A precessing magnetization vector induces an electric current in an external loop of wire (a “coil”) and therefore produces a nuclear magnetic resonance signal (Fig. 2.17B). This signal is proportional to the component of the magnetization vector perpendicular to the external magnetic field. The current induced by the magnetic moment of a single precessing nucleus is too small to measure. Under a strong magnetic field, there is a small excess of nuclei precessing around B0 compared with those precessing in the opposite direction (few per million). If there is a sufficiently large number of precessing spins such that there is a dominant direction of alignment of magnetic moments, the “net magnetization M0” (Fig. 2.18) is measurable as long as it too is precessing. At equilibrium, there is no component of M0perpendicular to B0; thus, no signal is available. To detect M0, one needs to disturb the equilibrium and apply a second smaller external magnetic field, B1, which rotates around the B0 field at the Larmor frequency (Fig. 2.19). This exerts two effects on the precessing magnetic moments of a large number of nuclei: It will synchronize their phase of precession and exert another force, or “torque,” that rotates the net magnetization away from the direction of B0. An effective means to generate a rotating B1 field is by application of radiofrequency energy. Both effects of the B1 field increase the component of M0 perpendicular to B0, Mxy, and therefore increases the signal induced by M0.

In MRI, the signal is generated by the induction of an electric current in a coil (longitudinal relaxation; local tissue environment) and T2 (vertical relaxation; rejects local spin dephasing).

Figure 2.18 Net magnetization vector. When an external magnetic field is applied, there is a small excess of spins pointing toward the external magnetic field compared with those pointing in the opposite direction (A). The net difference (trillions of atoms) will produce a non-zero net magnetization, M0, pointing toward the external magnetic field (B).

Figure 2.19 Rotating B1 field. To produce a non-zero net transverse magnetization, a small second rotating field, B1, is applied orthogonal to the external magnetic field. This synchronizes spins and creates an additional torque on the net magnetization away from the external field.

Once the B1 field is stopped, the net magnetization returns slowly to its equilibrium state in the direction of B0, a phenomenon called “relaxation” (Fig. 2.20A). Relaxation of the magnetization vector of each tissue is different and is most influenced by two important quantities (Fig. 2.20B): T1 characterizes the relaxation of the component of the magnetization vector that is parallel to the B0 field (longitudinal relaxation), and T2 characterizes the relaxation of the component that is perpendicular to the B0 field (transverse relaxation). T1 reflects local tissue environment, whereas the much shorter T2 reflects local spin dephasing. The latter can be partly reversed (and thus signal recovered) using a refocusing technique by a 180-degree pulse perpendicular to B0 (spin refocused echo).

MR images are generated using gradients (varying magnetic fields) and pulse sequences (external pulses).

MRI is a technique that generates images of the human body, using these basic physical principles together with spatially varying magnetic fields, called “gradients,” and complex sequences of external pulses, called “pulse sequences.” These gradients encode spatial positional information. Pulse sequences are constructed to generate images that emphasize one or more aspects of these T1 and T2 components (Fig. 2.15). In these sequences, preparatory pulses are occasionally used. Examples of such preparatory pulses are an inversion 180-degree pulse, which is used to eliminate signal from a specific tissue (fat, blood, myocardium), and a chemical saturation pulse to eliminate the signal of fat. The signal of blood is affected by its motion in and out of the imaging plane, a characteristic that many sequences take into account to image blood vessels.

Figure 2.20 Relaxation. Once the B1 field is discontinued, the net magnetization vector slowly returns to its state of equilibrium (A). The temporal variations of the longitudinal, Mz, and transverse, Mxy, components (B) of that magnetization vector are, respectively, determined by the T1 and T2 values of the local tissue environment.

An MRI protocol is simply a set of pulse sequences, chosen to emphasize a specific structure or pathologic process under consideration. Sample cardiopulmonary MRI protocols are illustrated in Table 2.6. For cardiac imaging, spin-echo sequences produce dark signal in the blood and are commonly called “black-blood sequences.” Gradient-echo sequences produce high signal in the blood and are called “bright-blood sequences” (Fig. 2.21). MRI is often performed using the intravenous administration of a gadolinium-based contrast agent to better visualize vascular structures, such as the aorta (Fig. 2.22A) (19) and the pulmonary arteries (Fig. 2.22B); to better characterize masses; and to characterize myocardial perfusion and viability. Cardiac MRI is now considered the gold standard for the assessment of left ventricular ejection fraction (20).

Table 2.6: Sample Thoracic MRI Protocols

Name

Sequence 1

Sequence 2

Sequence 3

Sequence 4

Sequence 5

Sequence 6

Chest mass

Localizer

Axial T1

Axial T2

Multiplanar T1 post-Gd

 

Optional sequences

Thoracic outlet

Localizer

Sagittal T1

Sagittal gradient

Axial T2

Optional sequences

Cardiac mass

Localizer

Multiplanar bright blood

Multiplanar black blood

Axial T1

Axial T1 post-Gd

Optional sequences

RV dysplasia

Localizer

Multiplanar bright blood

Axial black blood

Axial black blood with fat-sat

Optional sequences

Pericarditis

Localizer

Multiplanar bright blood

Multiplanar black blood

Axial T1

Coronal T2

Optional sequences

Cardiac viability

Localizer

Multiplanar bright blood

Short-axis
perfusion
post-Gd

Multiplanar delayed hyperenhancement

Optional sequences

Aorta

Localizer

Black blood axial

3D gradient
   sagittal
   oblique
   post-Gd

3D gradient oblique coronal post-Gd

Optional sequences

MRI, magnetic resonance imaging; RV, 3D, three-dimensional; fat-sat, fat saturation.

The length of an MRI examination is 30 to 60 minutes, much longer than the few seconds of a CT examination. MRI offers much better tissue contrast than CT but typically has lower in-plane spatial resolution. Contraindications to MRI examinations include patients with pacemakers, aneurysm clips, metal fragments in the eyes, or some metallic implants. Pregnant patients are often scanned, usually without the administration of intravenous contrast.

Figure 2.21 A–D: Cardiac magnetic resonance imaging enables dynamic evaluation of cardiac structure and motion. This image composite illustrates bright-blood images in four different standard planes during diastole.

Figure 2.22 Magnetic resonance angiography. Magnetic resonance angiography of the aorta and its branches is useful to evaluate aortic dissection (A). Magnetic resonance angiography of the pulmonary arteries enables good visualization of the pulmonary arteries (B) and can be used to rule out pulmonary embolism.

Nuclear Medicine

In chest radiography and CT, the source of radiation is located outside the patient. In nuclear medicine studies, the source of radiation is administered to the patient in the form of a radiopharmaceutical, either via inhalation or intravenous injection. Once properly distributed into the body, the source emits radiation by radioactive decay, which is then captured by detectors outside the patient (Fig. 2.23).

In nuclear medicine imaging, a radioactive agent is administered to a patient and images are detected based on emission and radioactive decay.

A radiopharmaceutical is a complex made up of two parts: a molecule or aggregate of molecules that has a specific biological function and a radionuclide that is the source of radiation. In chest radiology, the radionuclide is either a gamma emitter or a positron emitter. Corresponding detectors for these are, respectively, a gamma camera and a positron emission tomography (PET) camera. A gamma camera records a projection of the radiopharmaceutical distribution in the patient. This is called planar imaging. Information from multiple projections can also be combined into volumetric data and produce tomographic images. This is called single photon emission CT. A PET camera is made of a ring of detectors surrounding the patient (Fig. 2.24), similar to CT. These detect collinear photons in opposite directions emitted by the annihilation of a positron–electron pair and can therefore pinpoint with relative precision the position of origin of the pair of photons in the body and produce tomographic imaging.

Figure 2.23 Principles of planar radionuclide imaging. The source of radiation is located inside the patient. Radiation that leaves the patient is captured by the detector, which produces projection images of the distribution of the radiotracer in the patient.

Figure 2.24 Principles of positron emission tomography. Emitted positrons quickly collide with tissue electrons, and their annihilation produces two collinear photons in opposite directions, which are captured by the detectors. The position of the photon pair and difference in timing between the two detections allow determination of position of the emitter.

The main clinical indications for radionuclide imaging in radiology are indicated in Table 2.7. The most common radionuclide examinations for cardiopulmonary disease are discussed.

Lung Perfusion Scan

Technetium-99m labeled human albumin microspheres or microaggregated albumin are complex aggregates that measure from 10 to 150 μm in size. They are administered via a peripheral vein and lodge themselves into precapillary arterioles in the lungs after going through the heart. The distribution of radioactivity throughout the lungs reflects the distribution of perfusion. In thromboembolic disease, part of the arterial vascular supply to the lungs is compromised and cannot be reached by the radiopharmaceutical. These regions will show absence of or decreased radioactivity (Fig. 2.25). This examination is very sensitive but not very specific for pulmonary emboli (23), because many etiologies of lung pathology can cause a decrease in lung perfusion, including malignancy and infection.

Perfusion scintigraphy reflects relative, not absolute, perfusion of the lungs.

Table 2.7: Indications for Thoracic Radionuclide Imaging

Pulmonary
   Ventilation/perfusion
      Pulmonary embolus (acute and chronic)
      Assessment of split function for surgical planning
      Emphysema surgery patient selection
   PET, SPECT
      Cancer staging and follow-up
      Indeterminate pulmonary nodule
   Gallium scan
      Infection (HIV)
      Neoplasm (lymphoma)
      Fever of unknown origin
Cardiac
   Cardiac perfusion
   Ischemia and infarct characterization
   Ejection fraction
   Wall motion

PET, positron emission tomography; SPECT, single position emission computed tomography; HIV, human immunodeficiency virus.

Figure 2.25 Ventilation-perfusion scan. There are multiple perfusion defects in both lobes, which are not matched by ventilation defects. This is consistent with high probability of a pulmonary embolus.

Lung Ventilation Scan

To improve specificity for pulmonary emboli, a lung perfusion scan is combined with a lung ventilation scan. In the latter, xenon (133Xe) or a99m Tc radiolabeled aerosol is administered to the patient via inhalation into a closed system. Xenon requires a special ventilation system, so technetium is generally preferred. Detectors measure radioactivity in the lungs during the wash-in, equilibrium, and wash-out phases of the examination. Classically, thromboembolic disease manifests as normal ventilation in areas of abnormal perfusion (Fig. 2.25), whereas other lung diseases typically show both abnormal ventilation and perfusion in the affected areas. However, in many cases of pulmonary emboli, matched defects can be seen (23).

Gallium Scan

Gallium (67Ga) citrate is administered to the patient intravenously. Gallium will distribute slowly throughout the body, accumulating mainly in the liver, spleen, bone marrow, nasopharynx, lachrymal glands, thymus, and breasts. Imaging is performed more than 24 hours after injection. Uptake of gallium in the lungs is increased in infection and inflammation and in neoplastic processes. It may be used to follow disease activity in lymphoma (Fig. 2.26) (24) and to differentiate between active tumor and residual fibrosis when soft tissue remains after treatment. It is also used in human immunodeficiency virus infection to detect lung disease in a setting of a normal chest radiograph.

The exact mechanism by which gallium uptake occurs is not well understood.

Positron Emission Tomography

A glucose analogue, 18FDG(fluorodeoxyglucose)-6-phosphate, is administered intravenously. FDG distributes in the body and localizes to different organs proportionally to the local glucose uptake and metabolism. Other less commonly used PET agents include 13N-ammonia, L-methyl-11C-methionin, and 18F-fluoromisonidazole. The two main organs of distribution of FDG are the brain and the heart. FDG also localizes to active tumor. It can detect tumors as small as 7 mm in size, which is better than single photon emission CT using thallium or99mTc sestamibi, which has poor sensitivity for tumors below 2 cm in size (25). For the heart, FDG allows assessment of local myocardial viability (26). PET lacks the spatial resolution of CT but enables the distinction between metabolically active and silent tissues, which CT cannot do. For the lungs, it enables distinction between benign and malignant tumor and between recurrent tumor and postoperative changes (Fig. 2.27) (27,28). In combination with CT, images of both modalities can be registered and provide a combination of anatomic detail and metabolic activity (29). Combined CT–PET scanners now permit simultaneous acquisition for better anatomic registration.

A negative PET in a patient with a lung nodule greater than 7 mm in diameter is a strong indicator of a benign lesion. However, some tumors, such as bronchoalveolar cell carcinoma, are known to yield false-negative PETs.

Figure 2.26 Gallium scan. There is increased gallium uptake in the multiple lymphoma lesions in the chest of this patient, the largest in the right supraclavicular area. (Courtesy of Dr. C. Bui, Ann Arbor, Michigan.)

Cardiac Perfusion Scan

Thallium chloride (201Tl) is a potassium analogue that distributes in tissues with an intact sodium-potassium pump. Technetium sestamibi distributes in tissues by passive diffusion, whereas teboroxime is avidly extracted by the myocardium. Either radiopharmaceutical can be used to assess local myocardial perfusion (30) by quantifying local differences in extraction of the radiopharmaceutical by the myocardium in the rest and stress states (Fig. 2.28). The stress state is induced either by exercise or by the administration of a drug like dipyridamole or adenosine, which increases heart rate and blood pressure. This study can also assess ventricular wall motion.

A positive PET in a patient with a lung nodule indicates a metabolically active lesion, such as a lung cancer, and also infection, such as tuberculosis.

Cardiac Function

Red blood cell labeling multiple gated acquisition studies (MUGA) is performed by different techniques involving technetium (99mTcO4-). Gating is used to assess differences in blood pool activities within the left ventricle, between systole and diastole. From these values, a left ventricular ejection fraction is determined. Wall motion can also be assessed by this study.

Figure 2.27 Positron emission tomography. Active tumor at the right lung apex shows increased metabolic rate and thus increased uptake of 18-fluorodeoxyglucose.

Figure 2.28 Cardiac nuclear medicine. Under stress, there is a defect of uptake of 99mTc-sestamibi in the middle to distal anterior wall and cardiac apex, which is not present at rest, consistent with ischemia. (Courtesy of Dr. J. Corbett.)

Infarct Scan

Pyrophosphate (99mTc-pyrophosphate) parallels calcium metabolism and binds to denaturated proteins in damaged myocardium. Increased update of the radiopharmaceutical is seen in myocardial infarction, as well as cardiomyopathies and myocardial injury.

Angiography

Fluoroscopic imaging is similar in principle to standard radiography, except that the detector is inherently digital, records dynamic information of temporal change in attenuation, and displays the information in real time. It is therefore equivalent to a rapid movie-like sequence of simple radiographs using the same projection, although the dose of radiation per image is lower. Angiography is the use of fluoroscopic imaging before, during, and after administration of intravenous iodinated contrast agents for the purpose of visualizing vascular structures. Contrast is administered either via a peripheral vein or by a catheter placed directly into the vessels of the thorax, including the aorta, coronary arteries, pulmonary arteries, bronchial arteries, or the heart. Although the risks related to angiography used to be significant, the introduction of pigtail catheters and low osmolarity contrast agents has significantly decreased these risks (31).

Angiography most often uses a digital subtraction approach for imaging (Fig. 2.29) (32,33). This involves two sets of images, respectively, without and with the intravascular contrast agent, that are subtracted from each other. If patient motion between the two sets of images is minimal, then the subtracted image will display mostly intravascular contrast and will minimize background structures.

Clinical indications for angiography are indicated in Table 2.8. In cardiopulmonary radiology, angiographic examinations for the lungs include pulmonary angiography (Fig. 2.30) (32) and bronchial artery angiography. For the heart and great vessels, aortography (33,34), coronary angiography, and ventriculography are performed. Aortography and pulmonary angiography are slowly being replaced by noninvasive techniques, like CT or MR angiography. Both CT and MR angiography are commonly performed for diagnosis before percutaneous interventional procedures like aortic stent grafts (35). Angiography is useful in emergency situations or when the results of noninvasive examinations are equivocal.

Many simple diagnostic angiography studies have been replaced by CT angiography and MR angiography. Thoracic angiography studies are now more often coupled to an intervention, such as embolization, stent-graft, and thrombolysis, than in the past.

Figure 2.29 Subtraction technique. The injury to the thoracic aorta is difficult to assess given the cardiac and rib shadows (A). If the background, the same image without contrast, is subtracted from the image with contrast, then the aortic injury becomes more conspicuous (B). (Courtesy of Dr. K. Cho, Ann Arbor, Michigan.)

Table 2.8: Indications for Thoracic Angiography

Pulmonary arteries
   Pulmonary embolism (acute or chronic)
   Pulmonary hypertension
      Chronic PE
      Extrinsic compression
      Mediastinal fibrosis
   Congenital abnormalities
      AVMs, hypoplastic, stenosis, anomalous venous drainage
Aorta
   Traumatic injury
   Ruptured aneurysm or dissection
   Stent graft placement
Cardiac
   Coronary artery disease
   Valvular disease

PE, pulmonary embolism; AVMs, arteriovenous malformations.

Figure 2.30 Pulmonary angiogram. There are filling defects in branches of the pulmonary arteries, consistent with pulmonary emboli. (Courtesy of Dr. K. Cho, Ann Arbor, Michigan.)

Figure 2.31 Principles of ultrasonography. Sound waves are emitted by a transducer, are reflected at tissue interfaces, and the reflected waves are captured by the same transducer. The intensity of the reflected signal and the time at which the signal was received are used to determine the exact position of the sonographic interfaces.

 

Ultrasonography

The basic principles of ultrasonography are as follows. A hand-held transducer containing a piezoelectric crystal is put into direct contact with the part of the body to be imaged and produces a narrow beam of sound waves, which travel through a physical medium by inducing compressions and rarefactions in that medium. These sound waves are in the ultrasound range from 1 to 20 MHz in medical applications and are beyond the audible range for humans (from 20 Hz to 20 kHz). These sound waves are either reflected back to the transducer, refracted, scattered, or attenuated at tissue interfaces. The waves that are reflected back to the transducer are recorded and produce a signal that, once processed, indicates the different interfaces in the penetrated tissues and their respective depths (Fig. 2.31). By rapidly changing the orientation of the beam of sound waves, two-dimensional imaging of a section of the body can be performed.

Diagnostic ultrasound waves are in the 1- to 20-MHz range (audible sound range is 20 Hz to 20 kHz).

Ultrasound is mostly used for cardiac imaging, not pulmonary imaging (36). Echocardiography is largely the province of cardiology in the United States and so is only covered briefly here. Because of its good spatial resolution and its exquisite temporal resolution, echocardiography is the first modality of choice in clinical practice for the anatomic and functional assessment of heart disease (Fig. 2.32). Its main clinical indications are listed in Table 2.9. It can be performed by either a transthoracic approach or a transesophageal approach. Doppler echocardiography can be used to quantify blood velocity through the different cardiac chambers and valves.

Ultrasound has a very limited diagnostic role in the chest outside the heart. It is commonly used to distinguish between pleural effusions and pleural masses (37,38) (although this is mostly done by CT) (Fig. 2.33) and to assess motion of the diaphragm (39). Ultrasound has a somewhat greater role in invasive procedures (40), including thoracentesis guidance and to guide a biopsy needle for chest wall masses and pulmonary masses adjacent to the chest wall. Most other lung biopsies are routinely performed under CT guidance.

Ultrasound is most commonly used by diagnostic radiologists for the evaluation and treatment of pleural disease.

Figure 2.32 Cardiac echocardiography. Echocardiographic imaging shows normal appearance of the heart in four standard planes. (Courtesy of Dr. W. Armstrong, Ann Arbor, Michigan.)

Table 2.9: Indications for Thoracic Ultrasound

Thoracic
   Pleural effusions
   Pleural masses
   Diaphragm
   Interventional procedures (guidance or localization)
Cardiac
   Congenital heart disease: shunts, complicated anatomy
   Myocardium
      Cardiomyopathy
      Ischemic disease
      Hypertension
   Pericardium: thickening, effusion, tamponade
   Cardiac masses: thrombus, tumors
   Valves: stenosis, regurgitation, prosthetic, vegetations
Vascular
   Aorta: dilatation, dissection
   Pulmonary arteries: pulmonary hypertension

Figure 2.33 Pleural effusion. A large anechoic area above the left hemidiaphragm is consistent with a left pleural effusion.

References

1. Fraser RS, Müller NL, Colman N, et al., eds. Fraser and Pare’s diagnosis of diseases of the chest. St. Louis: WB Saunders, 1999.

2. American College of Radiology. ACR standard for the performance of pediatric and adult chest radiography. Reston, VA: American College of Radiology, 1997.

3. Yaffe MJ, Rowlands JA. X-ray detectors for digital radiography. Phys Med Biol 1997;42:1.

4. Rigler LG. Roentgen diagnosis of small pleural effusion: a new roentgenographic position. JAMA 1931;96:104.

5. Rundle FF, Delambert RM, Epps RG. Cervicothoracic tumors: a technical aid to their roentgenologic localization. AJR Am J Roentgenol1959;81:316.

6. Grenier P, Maurice F, Musswet D, et al: Bronchiectasis: assessment by thin section CT. Radiology 1986;161:95–99.

7. American College of Radiology. ACR standard for the performance of pediatric and adult thoracic computed tomography (revised). Reston, VA: American College of Radiology, 1998.

8. Bushberg JT, et al. The essential physics of medical imaging, 2nd ed. Philadelphia: Lippincott, Williams & Wilkins, 2002.

9. Remy-Jardin M, Remy J, Artaud D, et al. Volume rendering of the tracheobronchial tree: clinical evaluation of bronchographic images.Radiology 1998;208:761–770.

10. Higgins WE, Ramaswamy K, Swift RD, et al. Virtual bronchoscopy for three dimensional pulmonary image assessment: state of the art and future needs. Radiographics 1998;18:761–768.

11. Webb WR, Müller NL, Naidich DP, eds. High resolution CT of the lung. New York: Lippincott-Raven, 1996.

12. Mayo JR, Remy-Jardin M, Müller NL, et al. Pulmonary embolism: prospective comparison of spiral CT with ventilation-perfusion scintigraphy. Radiology 1997;205:447–452.

13. Henschke CI, Yankelevitz DF. CT screening for lung cancer. Radiol Clin North Am 2000;38:487.

14. Slovis TL. CT and computed radiography: the pictures are great, but is the radiation dose greater than required? AJR Am J Roentgenol 2002;179:39–41.

15. Wall BF, Hart D. Commentary: revised radiation doses for typical x-ray examinations. Report of a recent review of doses to patients from medical x-ray examinations in the UK by NRPB. Br J Radiol 1997;70:437–439.

16. Naidich DP, Zerhouni EA, Siegelman SS, eds. Computed tomography and magnetic resonance of the thorax. New York: Raven Press, 1991.

17. Fortier MV, Mayo JR, Swensen SJ, et al. MR imaging of chest wall lesions. Radiographics 1994;14:597.

18. Heelan RT, Demas BE, Caravelli JF, et al. Superior sulcus tumor: CT and MR imaging. Radiology 1989;170:637.

19. Kersting-Sommerhoff BA, Higgins CB, White RD, et al. Aortic dissection: sensitivity and specificity of MR imaging. Radiology1988;166:651–655.

20. Manning WJ, Pennell DJ, eds. Cardiovascular magnetic resonance. New York: Churchill Livingstone, 2002.

21. Kauczor HU. Current issues in hyperpolarized gases in MRI: biomedical investigations and clinical applications. NMR Biomed2000;13:173.

22. Brown MA, Smelka RC. MR: basic principles and applications. New York: John Wiley & Sons, 1995.

23. The PIOPED Investigators. Value of the ventilation/perfusion scan in acute pulmonary embolism. JAMA 1990;263:2753–2759.

24. Front D, Israel O. The role of Ga-67 scintigraphy in evaluating the results of therapy of lymphoma patients. Semin Nuclear Med1995;25:60–71.

25. Goldsmith SJ, Kostakoglu L. Nuclear medicine imaging of lung cancer. Radiol Clin North Am 2000;38:511.

26. Mesotten L, Maes A, Vande Werf F, et al. PET radiopharmaceuticals used in viability studies in acute myocardial infarction: a literature survey. Eur J Nucl Med Mol Imag 2002;29:3–6.

27. Sazon DAD, Santiago SM, Hoo GWS, et al. Fluorodeoxyglucose-positron emission tomography in the detection and staging of lung cancer. Am J Respir Crit Care Med 1996;153:417.

28. Chinn R, Ward R, Keyes JW, et al. Mediastinal staging of non-small-cell cancer with positron emission tomography. Am J Respir Crit Care Med 1995;152:2090.

29. Wahl RL, Quint LE, Greenough RL, et al. Staging of mediastinal non-small cell lung cancer with FDG PET, CT and fusion images: preliminary prospective evaluation. Radiology 1994;191:371.

30. Hachamovitch R. Clinical application of rest thallium-201/stress technetium-99m sestamibi isotope myocardial perfusion single-photon emission computed tomography. Cardiol Rev 1999;7:83.

31. Hudson ER, Smith TP, McDermott VG, et al. Pulmonary angiography performed with iopamidol: complications in 1434 patients.Radiology 1996;198:61–65.

32. van Rooij WJ, den Heeten GJ, Sluzewski M. Pulmonary embolism: diagnosis in 211 patients with use of selective pulmonary digital substraction angiography with a flow-directed catheter. Radiology 1995;195:793.

33. Guthaner DF, Miller DC. Digital substraction angiography of aortic dissection. AJR Am J Roentgenol 1983;141:157.

34. Davies ER, Roylance J. Aortography in the investigation of traumatic mediastinal haematoma. Clin Radiol 1970;21:297.

35. Thurnher SA, Dorffner R, Thurnher MM, et al. Evaluation of abdominal aortic aneurysm for stent-graft placement: comparison of gadolinium-enhanced MR angiography versus helical CT angiography and digital subtraction angiography. Radiology 1997;205:341–352.

36. Otto CM. Textbook of clinical echocardiography. St. Louis: WB Saunders, 2000.

37. McLoud TC, Flower CDR, Hadfield JW. Ultrasound of the pleura: sonographic, CT and MR imaging. AJR Am J Roentgenol1991;156:1145–1153.

38. Yang PC, Luh KT, Chang DB, et al. Value of sonography in determining the nature of pleural effusion: analysis of 320 cases. AJR Am J Roentgenol 1992;159:29–33.

39. Gottesman E, McCool FD. Ultrasound evaluation of the paralyzed diaphragm. Am J Respir Crit Care Med 1997;155:1570–1574.

40. O’Moore PV, Mueller PR, Simeone JF, et al. Sonographic guidance in diagnostic and therapeutic interventions in the pleural space. AJR Am J Roentgenol 1987;149:1–5.