High Resolution Computed Tomography of the Lungs: A Practical Guide D Karthikeyan
Chapter Notes

Save Clear

Computed Tomography Techniques and AnatomyChapter 1

High-resolution computed tomography (HRCT) is now regarded as an indispensable tool for the investigation of patients with suspected or known parenchymal lung disease. The cross-sectional perspectives and high spatial resolution make HRCT superior to other imaging modalities and it has been integrated in to imaging algorithms for assessment of a number of diffuse lung process. Today HRCT has become a commonly requested imaging technique. Familiarity and basis of interpretation of HRCT images is critical for accurate diagnosis.
Parenchymal lung diseases include Airway disease, Air space disease and Interstitial diseases. The latter needs special attention since HRCT provides histospecific diagnosis in many cases.
Numerous studies have shown the basic superiority of HRCT over chest radiography in terms of improved detection of lung disease, provision of a specific diagnosis, and the identification of reversible disease.
HRCT today uses thin section CT images (0.625–2 mm slice thickness) often with a high-spatial-frequency reconstruction algorithm.
With advent of multi-detector CT scanners capable of acquiring data throughout the entire thorax in a single breath-hold, two variation of HRCT are available. The first and more common method entails obtaining axial HRCT images spaced at 10 to 20 mm intervals throughout the thorax. The second uses the ability of multi-detector CT (MDCT) scanners to provide volumetric data sets allowing spaced, contiguous, and/or overlapping HRCT images to be reconstructed (Figures 1-1 to 1-4).
zoom view
Figures 1-1A and B: Axial CT section. (A) In comparison with HRCT section; (B) At the same level shows the bronchiolar change and secondary architectural distortions very clearly in the HR sections
zoom view
Figures 1-2A and B: Axial and High-resolution CT section in a patient with DIP showing the advantage of using a HR protocol
zoom view
Figure 1-3: Scanogram with scan plan for MDCT volume HRCT
Indications for High-Resolution CT
  • To detect focal or diffuse lung disease in patients with normal or equivocal radiographic abnormalities.
  • To narrow the differential diagnosis or make a specific diagnosis, in patients with obvious but nonspecific radiographic abnormalities.
  • To guide the site of lung biopsy.
  • To investigate patients presenting with hemoptysis.
  • To assess the distribution of emphysema.
  • To evaluate disease reversibility, particularly in patients with fibrosing lung disease.
Recent progress in CT technology has resulted in profusion of slices with clinical scanners having up to 320 slices or dual source being used. Today thin sections up to .5/.6 mm can be obtained with excellent Z axis coverage and resolution giving isotropic coronal/sagittal images.
Two-dimensional multiplanar reformats help solve the inherent difficulties of assessing the craniocaudal extent of disease on axial images.
To obtain an optimum HRCT few technical modifica­tions to routine CT are to be considered.
  • Collimation: Thinnest available collimation should be undertaken. This includes .5/.6 to 1 mm collimation. Higher collimations (1 cm) reduces the ability of the scan to resolve small structures.
  • Reconstruction algorithm: High spatial frequency or sharp algorithm is used, that is bone algorithm which reduces image smoothing and increases spatial resolution, making structures appear sharper.
    zoom view
    Figures 1-4A to D: Axial with coronal, sagittal and volume reconstructions from isotropic volumetric datasets
  • kV, mA and scan time: In HRCT image noise is more apparent than with standard CT which appears as graininess or mottle, in addition using sharp reconstruction algorithm makes the noise more prominent. Most of this noise is quantum related, thus decreased by increased technique, that is kVp and mA or scan time. However as increasing scan time is not practicable and can lead to motion related artifacts the other two parameters are altered to reduce noise. The routine technique used for HRCT includes kVp of 120–140 and mA of 140–240. In recent years introduction of low dose HRCT using kVp of 120–140 and mA of 30 using 2 seconds scan time has proved to be an alternative approach. However low dose technique failed to demonstrate ground glass opacities in few of the cases. Low dose technique also proved to have more prominent linear streak artifacts. In conclusion, although conventional HRCT was more accurate than low dose HRCT this difference was not significant and both techniques provided similar anatomic information.
  • Matrix size: Largest available matrix size should be employed which is 512 × 512 for most commercial scanners including ours.
  • Field of view and use of targeted reconstruction: Field of view should be large enough to encompass the patient, that is 35 cm. Retrospective targeting image reconstruction to a single lung instead of the entire thorax, using a smaller field of view significantly reduces the image pixel size and this increases the spatial resolution.
    zoom view
    Figure 1-5: Scanogram with acquisition plan for conventional HRCT, note the interslice gap
  • Image photography and window setting: Even though there is no specific or ideal window setting for demonstration of lung anatomy in HRCT, it is important that at least one lung window setting be used consistently in all patients as to make the comparison easy and effective. The so-called lung window of −700 window mean/1000 HU window width for routine lung examination. This can be modified for visualization of specific lesions and targeted area approach.
Few alterations include:
  • Window of −500 to −700/2000 HU could be employed for evaluation of pleuro-parenchymal abnormalities
  • Window of 30/350 HU can be employed for evaluation of mediastinum, and pleura
  • Low window setting of −800 to −900 HU with narrow window width of 500 HU can be deployed for evaluation of emphysema as the lung tissue will appear gray and the emphysematous tissue will stand out.
In conclusion, window mean of −600 to −700 HU and window width of 1000 to 1500 HU are appropriate.
Interpretation of large images is much easier than smaller ones, we provide 9 images in one 14 × 17 film for lung window images and twelve on one film for mediastinum.
The average density of each voxel is measured in Hounsfield Units; these units have been arbitrarily chosen so that zero is water density and −1000 is air density. The range of Hounsfield Units encountered in the thorax is wider than in any other part of the body, ranging from aerated lung (approximately −800 HU) to ribs (+700 HU).
The window width determines the number of Hounsfield Units to be displayed. Any densities greater than the upper limit of the window width are displayed as white, and any below the limit of the window are displayed as black. Between these two limits the densities are displayed in shades of gray.
In volume HRCT modifications (VHRCT), the natural contrast of lung helps to modify the techniques of acquisition in MDHRCT to deliver a lower radiation, 100–120 kV, 80–140 mAs is used for multi-detector row CT.
Images are acquired with a 16– or 64–detector row CT scanner during a single breath hold lasting about 4–10 seconds, minimizing respiratory motion artifacts. A high-frequency algorithm, 512 or 768 and a 25.6 to 35 cm field of view used, with a rotation time of 500–600 msec allows a marked decrease in cardiac pulsation artifacts.
With a detector size of 0.625 mm, images with a section thickness of approximately 1 mm are reconstructed at intervals of approximately 0.5 mm, thereby producing a voxel of almost cubic dimensions (isotropic) and allowing the creation of excellent 2D and 3D reformatted images.
With optimal scanning technique, such as limitation of the field of view to the parenchyma of both lungs, the spatial resolution is between 0.3 and 0.5 mm. Depending on orientation, position and contrast within the lung parenchymal structures as small as 0.3 mm are identified. Pulmonary artery branches down to the 16th and bronchi down to the 8th generation will be depicted.
Streak and motion artifacts are two common types of artifacts seen on HRCT.
zoom view
Figure 1-6: Axial HRCT section showing star artifacts due to breathing
Streak artifacts arise from edges of sharply marginated, high contrast structures such as bronchial walls, ribs, vertebral bodies and seen mainly overlying the posterior lung, paralleling the pleural surface and posterior chest wall.
Motion artifacts or star artifacts are commonly visible, particularly at lung base, adjacent to the heart due to cardiac pulsation (Figure 1-6).
Patient Related
  • Motion artifacts (respiratory, cardiovascular, or both) create pseudobronchiectasis, pseudo—ground-glass opacity, and star artifacts.
  • Dependent atelectasis mimics or masks early subpleural lung disease.
  • Image noise in large patients may obscure abnormalities.
  • Left heart failure with interstitial edema mimics infiltrative lung disease.
  • Incorrect window width and level.
  • Failure to detect bronchiectasis due to mucous plugging.
  • Mosaic perfusion (airway, vascular, or infiltrative lung disease).
  • Failing to suspect infection when septal lines and ground-glass opacity are present.
Optimal studies of the lung parenchyma for diffuse pathology are best performed at full inspiration to promote uncrowding of vascular structures as well as to avoid gravity-dependent fluid accumulation and atelectasis. Such physiologic gravity-dependent densities can often be seen in the posterior aspect of the lower or upper lobes in the supine patient.
Scans should be done with breath held in full inspiration. However the use of expiratory CT has been reported in patients with emphysema, asthma, McLeods syndrome, cystic lung disease and in patients with variety of large and small airway obstruction. In normal subjects lung parenchyma increase uniformly in attenuation following expiration, but in presence of obstruction lung volume remains same and the lung parenchyma remains lucent.
The regional inhomogeneity of the lung density is accentuated and small or subtle areas of air-trapping may be revealed on CT performed at end-expiration. Expiratory CT may be helpful in differentiating between the three main causes of a mosaic pattern (infiltrative lung disease, small airways disease and occlusive pulmonary vascular disease) which may be problematic on inspiratory CT (Figures 1-7 and 1-8).
Expiratory Views
Three postexpiratory views are routinely performed at the level of the:
  1. Aortic arch.
  2. At the tracheal carina.
  3. Above the diaphragm.
These images are performed with 1–2 mm collimation at end expiration.
Dynamic studies where sections are obtained in rapid succession at a given level during forced expiration may improve the conspicuity and apparent extent of air-trapping. Dynamic sequence (acquired at the level of the arch, carina and 2 cm above the right hemidiaphragm) is obtained as a cine-acquisition without table incrementation, and reconstructed at 1 s intervals. Window level/width settings of −700/1000HU are recommended.
Interspaced high-resolution CT sections in the supine position are usually sufficient for the diagnosis of diffuse lung disease. However in many normal subjects a “dependent density” or “subpleural lines” is seen in the dependent lung areas. These normal findings can mimic early lung fibrosis.
zoom view
Figures 1-7A and B: Coronal color coded HRCT image showing changes of inspiration and expiration
zoom view
Figures 1-8A and B: Axial HRCT in inspiration Figure A, and expiration Figure B, note the posterior tracheal wall status in inspiration and expiration (arrows)
Additional scans obtained with the patient prone are occasionally necessary to detect or exclude subtle disease in the posterior parts of the lung. In case of lung pathology dependent density will persist whereas in normal subjects they disappear (Figures 1-9A and B). Most of the time prone scans are obtained when normal physiologic dependent opacification needs to be differentiated from a true pathologic abnormality (Figures 1-10A and B).
Breathing during CT causes motion artifacts in the lung parenchyma that are detrimental to the diagnosis of small airways disease.
zoom view
Figures 1-9A and B: Axial spine and prone HRCT images showing the dependent subpleural line (A) and disappearing in prone (B) Imaging
zoom view
Figures 1-10A and B: Axial supine and prone scans showing dependent atelectasis and bronchovascular changes which resolve in prone scans (arrows)
Lateral decubitus positioning causes the dependent hemithorax to be relatively splinted, thereby restricting movement of the thoracic cage on that side. The dependent lung is more opaque than the upper lung because of gravitational differences in perfusion and inflation. This technique is useful in children (Figure 1-11)
ECG Gating
ECG gating improves image quality and reduced cardiac artefacts as can be seen easily in data sets obtained during coronary CT angiography (Figure 1-12). However, it has not yet been determined whether ECG gating actually improves the diagnostic accuracy of HRCT and whether it should be used in routine imaging, but it should be easier to detect subtle parenchymal and airway abnormalities on images with less motion in subset of patients were routine images show movement degradation.
To counteract the noise inherent in thin sections, high exposure factors (120–140 kVp and 240–300 mA) were originally recommended. Today lower dose techniques are being used, very low dose techniques failed to demonstrate subtle ground glass opacity and emphysema. An adequate standoff would be to try and use ranges from 70–140 Mas (Figures 1-13A to C).
zoom view
Figure 1-11: Decubitus HRCT demonstrating the physiological changes, the side which is down behaves physiologically as in expiration due to splinting (arrow)
zoom view
Figure 1-12: Axial ECG gated step and shoot HRCT
zoom view
Figures 1-13A to C: Axial HRCT sections at 60, 80, 100 Mas – Note there is no significant change in image quality
Maximum intensity projections: MIP consists of pro­jecting the voxel with the highest attenuation value on every view throughout the volume onto a 2D image. The major application of MIP is to improve the detection of pulmonary nodules and assess their profusion. MIP also helps characterize the distribution of small nodules which help in characterizing their location (Figures 1-14A to D).
Minimum intensity projection: MIP consists of projecting the voxel with the lowest attenuation value on every view throughout the volume onto a 2D image. Difference in density between the endobronchial (pure) air and the lung parenchyma, corresponding to a difference in attenuation of 50–150 HU, permits visualization of the bronchi below the subsegmental level. MIP is the optimal tool for the detection of attenuation differences (Figure 1-15).
Volume Rendering
Three-dimensional volume rendering (VR) assigns a color and opacity to each attenuation threshold chosen based on the tissue charecteristics. All data in the volume are overlapped which may sometimes produce an near pathologic pattern (Figure 1-16).
zoom view
Figures 1-14A to D: Axial and coronal HRCT image with corresponding MIP image showing profusion of nodules which are seen very easily in the MIP images
zoom view
Figure 1-15: Coronal MinIP highlighting air in trachea and segmental bronchi
zoom view
Figure 1-16: Volume rendering image in a patient with fibrosis showing the architectural distortion well
An accurate understanding of normal lung anatomy along with pathological anatomical alterations in disease is essential in the interpretation of HRCT images (Figures 1-17 and 1-18).
Right Lung
Right Upper Lobe
Right Middle Lobe
Right Lower Lobe
Medial basal
Anterior basal
Lateral basal
Posterior basal
Left Lung
Left Upper Lobe
Apical posterior
Superior lingular
Inferior lingular
Left Lower Lobe
Anterior medial basal
Lateral basal
Posterior basal
zoom view
Figure 1-17: HRCT image of the lung
zoom view
Figure 1-18: Images showing axial CT anatomy
The accurate interpretation of HRCT requires clear understanding of normal lung anatomy and pathological alterations in it.
The lung is maintained in a stable position within the chest by its hilar connection, to which the central or core structures are ultimately attached. The bronchovascular structures form the stem upon which the functional parenchymas are distributed.
The parenchyma of the lung includes the pulmonary alveolar epithelium and capillary endothelium and the spaces between these structures, together with the tissues within the septa including the perivascular and perilymphatic tissues. More centrally it includes the peribronchiolar and peribronchial tissues.
The major core structures include pulmonary artery and its branches and the bronchi, which characteristically run in a parallel fashion. Both are enclosed within a connective tissue sheath, within which amorphous interstitial collagen, lymphatics and small lymph nodes can be found. The lung lymphatic system drains the visceral pleura and courses within the interlobular septa in parallel with septal veins.
There are approximately 23 generation of dichotomous branching in airways, from the trachea to the alveolar sac.
The division of the trachea gives rise to the left and right mainstem bronchi, which further divide into lobar and segmental bronchi. Segmental bronchi divide, and after 6 to 20 divisions they no longer contain cartilage in their walls and are referred to as bronchioles.
The trachea, bronchi and bronchioles down to the level of the terminal bronchiole constitute the purely conductive portion of the airway.
The terminal bronchiole subdivide into respiratory bronchioles from which alveoli arise. Respiratory bronchioles give rise to alveolar ducts which give rise to alveolar sacs. These conduct air and also participate in gas exchange.
The central bronchi down to the segmental level can be identified routinely with 10 mm collimation CT sections, and 8th order branching airways can be seen with thin sections.
Airway diseases may be divided into those conditions affecting primarily the large airways, such as the trachea, main, lobar, and segmental bronchi, and those conditions primarily affecting smaller airways, particularly those airways less than 3 mm in size. Large airway diseases may be further subdivided into focal and diffuse airway abnormalities, primarily related to either narrowing or dilation.
Communication between adjacent alveoli occur through pores of Kohn, communication between distal bronchiole and alveoli occur through canals of Lambert.
Acinus: It is a unit of lung supplied by the terminal bronchiole, this typically include 3 respiratory bronchioles and their corresponding alveolar ducts and alveoli. An individual acinus measures 7-8 mm and is recognizable when filled with fluid or cells (Figures 1-19 to 1-22).
zoom view
Figure 1-19: 3D volume rendering of lung
zoom view
Figures 1-20A to C: 3D volume rendering of trachea, major bronchi and lung
zoom view
Figures 1-21A and B: Schematic of acini and alveolus
zoom view
Figure 1-22: Schematic of tracheobronchial tree
The most useful subsegmental unit of the lung in terms of HRCT is the secondary pulmonary lobule (SPL). Defined as the smallest unit of lung function marginated by connective tissue septae. It is a polyhedral structure that measures approximately 1-2.5 cm on each side, and is supplied by 3-5 terminal bronchioles, and therefore include 5-17 acini within the borders formed by interlobular septa. The secondary pulmonary lobule is divided into core and septal structures, the core structures include the pulmonary arteriole, 15terminal bronchiole–termed as centrilobular artery and bronchus, and accompanying lymphatics. The septal structures include pulmonary veins, lymphatics and the septum itself (Figure 1-23).
Normally they are visible over the upper lobes, the anterior and lateral aspects of right middle lobe, lingula, and along the diaphragmatic surface of lower lobes, the centrilobular arteries and bronchus branch dichotomously within the secondary pulmonary lobule (SPL) producing interlobular arteries, respiratory bronchioles and acinar arteries. Terminating in pulmonary gas exchange unit.
Interlobular Septae
Secondary lobules are marginated by septae which extend inward from the pleural surface. These septae are well defined in the anterior, lateral and diaphragmatic surface, they measure about 100 mmu (0.1 mm) in the subpleural location. Few normal septae are often visible on routine HRCT. They appear as thin straight lines of uniform thickness and are 1-2 cm long.
In central lung these septae are thinner and less well defined (Figure 1-24).
Lobular Core
The HRCT appearances and the visibility of structures in the core are determined by the size. Secondary lobule is supplied by arteries and bronchioles that measure approximately 1 mm in diameter, intralobular bronchioles and arteries measure 0.7 mm. Acinar bronchioles and arteries measure 0.5 mm. The visible lobular core structures do not extend to the pleural surface.
Important rule to remember is that on routine HRCT intralobular bronchioles are not normally visible and bronchioles are normally not seen within 2-3 cm of pleural surface.
It is necessary to identify the peripheral pulmonary arteries to define the center of the pulmonary lobule.
The substance of the secondary lobule which surrounds the lobular core and is contained within the interlobular septae consists of functioning lung parenchyma namely the alveoli, alveolar ducts, and vessels.
zoom view
Figure 1-23: Schematics of secondary pulmonary lobule and alveoli
zoom view
Figure 1-24: Axial HRCT section of lungs showing interlobular septal prominence outlining the secondary pulmonary lobule
This parenchyma is supported by a network of central and peripheral fibers to form a fiber skeleton of the lung.
Lung is supported by a network of connective tissue, called lung interstitium. This extends from the hila to the periphery of lung. This interstitium is not usually recognizable on HRCT and its clear visualization depicts pathology. The interstitium is considered to have various components.
Peribronchovascular Interstitium
It is a system of fibers that invests the bronchi and pulmonary arteries. It extends from the hila to the periphery of the lung to invest the bronchioles and terminal arteries in the secondary pulmonary lobules and is referred to as centrilobular interstitium.
Subpleural Interstitium
It is located beneath the visceral pleura and envelops the lung in a fibrous sac. From this fibrous connective tissue, septae penetrate the lung parenchyma to form interlobular septa. These interlobular septa form the basic structure imaged in the periphery of the lung called the secondary pulmonary lobules. The terms secondary pulmonary lobules, secondary lobule and pulmonary lobule are often used interchangeably. Secondary pulmonary lobules are irregular, polyhedral in shape and somewhat variable in size, measuring approximately 1-2.5 cm in diameter in most locations (Figure 1-25).
Secondary pulmonary lobule has got three main components:
  1. Interlobular septa.
  2. Centrilobular region.
  3. Lobular parenchyma and acinus.
zoom view
Figure 1-25: Schematic of divisions of lung interstitium
Interlobular Septa and Contiguous Subpleural Interstitium
As mentioned above subpleural interstitium invests the lung with its extension into lung parenchyma to form interlobular septa, which contains terminal pulmonary veins and lymphatics. The veins are sometimes seen as linear or arcuate structures 5-10 mm from the centrilobular structures. The septa are better defined and visualized in the apical and anterior segment of the upper lobe, anterior and lateral aspects of middle lobe and lingula, diaphragmatic surface and along the mediastinal pleura. Within the central lung the interlobular septa are thinner and less well defined than the periphery (Figures 1-26A and B).
Centrilobular Region
The central portion of the secondary pulmonary lobule is referred to as centrilobular region or lobular core and contains the pulmonary artery and terminal bronchioles. It is invested with centrilobular interstitium, an extension of peribronchovascular interstitium. These usually appear as central linear, branching or dot like pattern in center of secondary pulmonary lobule about one centimeter from the pleura (Figure 1-27).
Lobular Parenchyma and Acinus
The substance of lung parenchyma within the pulmonary lobules consists of acini and intricate network of capillary beds. This vascular network along with walls of the acini and small airways are supported by a connective tissue stroma, a fine network of very thin fibers within the alveolar septa called the intralobular interstitium, or alveolar interstitium. This is invisible on HRCT in normal individuals. Pulmonary acini are the functioning structure of lung. There are a dozen or less of acini in each secondary pulmonary lobule, and vary in size measuring 6–10 mm each.
Intralobular Interstitium
It was referred by Weibel as septal fibers has been already described as a component of secondary pulmonary lobule constituting the lobular core.
zoom view
Figures 1-26A and B: Coronal and sagittal CT showing the interlobular septal and fissural prominence
zoom view
Figure 1-27: Schematic of secondary pulmonary lobule
Earliest manifestation of fibrotic lung disease on HRCT is abnormal thickening of interlobular interstitium (Figure 1-28).
Cortical Lung
It consists of 3-4 rows of secondary pulmonary lobule forming a layer of 4 cm at the lung periphery and along the lung surface. Bronchi and pulmonary vessels are small at this level.
zoom view
Figure 1-28: Axial HRCT section showing the intralobular interstitial involvement
Medullary Lung
Pulmonary lobules in the central lung are smaller and more irregular in shape, parahilar vessels and bronchi are large and are easily seen on HRCT (Figures 1-29A and B).
Zonal Predilection
The upper and lower lung zones are physiologically quite different. These differences explain, in part, the regional predilection of some diffuse lung diseases. The lower lung receives the greatest amount of ventilation, perfusion, and lymphatic drainage and is more frequently affected by various fibrotic processes and those affecting the lymphatic system.
zoom view
Figures 1-29A and B: Axial and coronal images showing the distribution of disease
zoom view
Figure 1-30: Axial HRCT showing subpleural sparing
• Sarcoidosis
• Lymphoma
• Lymphangitic carcinomatosis
• Sarcoidosis
• Idiopathic pulmonary fibrosis
• Lymphangitic carcinomatosis
• Collagen vascular diseases
• Rheumatoid arthritis
• Sarcoidosis
• Chronic interstitial pneumonias
• Coal workers' pneumoconiosis
• Silicosis
• Langerhans cell histiocytosis
• Ankylosing spondylitis
• Sarcoidosis
• Chronic medications
• Lymphangitic carcinomatosis
• Neurofibromatosis
• Vasculitis
• Silicosis
• Idiopathic pulmonary fibrosis
• Lymphangitic carcinomatosis
• Collagen vascular diseases
• Asbestosis
• Lymphangioleiomyomatosis
However, the upper lung has a relatively higher oxygen tension and pH, but less efficient lymphatic drainage. The upper lung is often affected by inhalational (e.g. silicosis) and granulomatous diseases and Langerhans cell histiocytosis. The lung may be further divided into an axial, parenchymal (middle), and peripheral compartment.
The axial compartment includes the peribronchovascular bundles and lymphatics and is contiguous with the mediastinum. The middle or parenchymal compartment is formed by the alveolar walls. The peripheral compartment includes the pleura, subpleural connective tissue interlobular septa, pulmonary veins and lymphatics, and the walls of the cortical alveoli (Figure 1-30). Although these compartments communicate with one another, relatively selective involvement may be seen in a number of disease processes.
The radiologist can take advantage of this zonal predilection when formulating a differential diagnosis and directing potential transbronchial or open lung biopsies.