Imaging Techniques for Tuberculosis

Abstract

Imaging has an important complementary role in the diagnosis of pulmonary and extrapulmonary tuberculosis, particularly in patients with non-specific symptoms. It helps to detect lesions and confirm the diagnosis, evaluate the extent of disease and its complications, monitor its progress and response to treatment, as well as demonstrate residual or recurring disease after completion of therapy. Radiographs are an excellent screening tool and typically the initial radiological investigation, particularly for pulmonary and musculoskeletal tuberculosis. Computed tomography is utilized to further evaluate pulmonary, abdominal, urinary tract, and head and neck tuberculosis. Magnetic resonance imaging is the modality of choice for assessing tuberculosis in the brain, spine, and musculoskeletal system. Imaging modalities such as contrast fluoroscopy studies, ultrasound imaging, and nuclear medicine imaging, particularly positron-emission tomography, may also be of benefit in the management of selected patients suspected to have tuberculosis. Imaging techniques, such as ultrasound and computed tomography, are also used to guide diagnostic and therapeutic aspirations and drainages, as well as biopsies for histopathological confirmation of tuberculosis. As each of these modalities have their own advantages, disadvantages, and limitations, the modality of choice therefore varies, depending on the clinical indications, and should be tailored to each individual patient.

1 Introduction

Tuberculosis (TB) may infect virtually all organ systems, with the lung being the most common organ affected. Extrapulmonary TB comprises up to 15% of all tuberculous infections (Peto et al. 2009). The definitive diagnosis of TB is made by isolation of Mycobacterium tuberculosis from a bodily sample (Pai et al. 2016). However, as a laboratory diagnosis cannot be achieved in approximately 15–20% of individuals (Taylor et al. 2000), treatment can be instituted based on a presumptive diagnosis (“clinically diagnosed TB”) (World Health Organization 2020), particularly in the presence of risk factors (see the first Chapter on Epidemiology of Tuberculosis and the last Chapter on Diagnostic Algorithm of Tuberculosis).

Patients with TB may present with a range of diverse clinical manifestations, which can be broadly classified into three groups, namely: (1) signs and symptoms of pulmonary TB, (2) signs and symptoms of extrapulmonary TB, and (3) asymptomatic patients suspected to have latent TB. Patients with extrapulmonary TB typically have either non-specific constitutional symptoms or clinical features relating to a specific organ, which mimic some other pathology and are hence particularly challenging to diagnose clinically. The roles of imaging are to detect lesions, suggest the diagnosis, evaluate the extent of the disease, look for complications, follow its response to treatment, as well as demonstrate residual or recurring disease upon completion of therapy (Bomanji et al. 2015).

2 Imaging Techniques

Conventional radiographs are the initial radiological investigation for pulmonary and musculoskeletal TB in most instances. Further cross-sectional evaluation using ultrasound (US) imaging, computed tomography (CT), and/or magnetic resonance imaging (MRI) is commonly the subsequent step in the line of investigations. The use of other imaging modalities, such as intravenous urography (IVU), barium studies, as well as a whole host of different radiological techniques, may also be beneficial in the management of patients suspected to have TB. In patients in whom TB is not suspected at presentation, imaging findings may raise the first concern for a tuberculous etiology. Imaging modalities such as US imaging and CT have an additional role in aiding aspiration and biopsy for confirmation of the diagnosis, as well as guiding therapeutic interventional procedures, such as drainage of abscesses and fluid collections. This chapter provides an overview of the major imaging techniques employed in investigating patients with TB. Some specific techniques will be covered in more detail in the following chapters dealing with imaging of TB affecting various organ systems.

2.1 Radiography

Radiographs are often the first-line imaging examination requested for detection of disease and sometimes provide a diagnosis. Radiographs complement the referring physician’s clinical suspicion based on the patient’s history and physical examination, together with laboratory findings. This assessment should ideally be made in conjunction with the knowledge of established risk factors, i.e., host and environmental factors (see the first Chapter on Epidemiology of Tuberculosis and the last Chapter on Diagnostic Algorithm of Tuberculosis).

2.1.1 Advantages of Radiography

Radiography has several advantages, namely:

• 1. It is readily available, often being the primary and/or sole imaging modality in small hospitals and clinics with limited resources, particularly in developing countries.

• 2. It is relatively simple to perform. All radiographers who have finished their basic training should be competent.

• 3. It is cost efficient, being far less expensive compared to all other imaging modalities.

• 4. It provides a quick overall imaging assessment of a relatively large area of interest, with less radiation exposure compared to many other imaging modalities, e.g., CT.

• 5. Its identification and localization of lesion(s) serve as a guide for further imaging such as CT or MRI.

• 6. It is a simple imaging tool for follow-up of treated lesions and for detection of complications or disease recurrence.

• 7. Being potentially portable, it is quickly available, even to the sickest patients at the emergency department, ward, or intensive care unit (ICU).

2.1.2 Computed Radiography, Digital Radiography, and PACS

Although still utilized in many developing countries and in facilities with limited resources, the time-honored traditional film-screen system has been superseded by computed radiography (CR) and digital radiography (DR). Being digital imaging systems, both CR and DR allow digital display of the acquired images. These digital images can be uploaded into the picture archiving and communication system (PACS).

CR is a cassette-based digital imaging method, as the image is obtained using a cassette, before the data is converted into a digital image in a CR reader. In CR, photostimulable phosphor plates are used in cassettes that are similar in appearance and dimensions to those used in traditional film-screen radiography. This is in contradistinction to the film-screen system, where a film has to be inserted in a dark room prior to its use. As both systems employ cassettes, the same X-ray machines used in the traditional film-screen cassette system may be retained, making the transition from traditional analogue to digital imaging seamless.

In DR, a built-in detector unit allows the X-ray photons to be converted directly into electrical signals that can be read directly as digital images. There are two types of detectors: one which uses indirect conversion, and the other which uses direct conversion with amorphous selenium. In the former, phosphor is used to convert the X-ray photons into light, which is then detected by photodiodes in an amorphous silicon thin-film transistor (TFT) array. The photodiode, each representing a single pixel, then produces an electrical charge that is sent to the image processor. In the latter, amorphous selenium is used as a photoconductor that will pass on electrical charges to charge collectors in the TFT array upon irradiation (Allisy-Roberts and Williams 2008).

The PACS enables the rapid transmission of digital images from the site of image acquisition to multiple different locations within a hospital, and even across the globe. PACS allows remote viewing and storage of these images while retaining their original quality. In PACS, a more economical, filmless, and efficient workflow is set up, whereby the digital images acquired can be viewed almost in real time, by both radiologists and clinical teams simultaneously. It also allows convenient access across multiple modalities, allowing prior studies from the same patient to be referenced and compared immediately.

Both CR and DR offer several advantages over conventional film-screen radiography:

• 1. The wide dynamic range of the photostimulable phosphor plate and DR detector reduces exposure errors and the need for repeat examinations.

• 2. They allow immediate manual optimization of the display features desired for the particular anatomical part to be imaged.

• 3. When integrated into the PACS, there is an efficient and almost live transmission of the acquired images. The system also stores and archives the digital images for convenient and fast retrieval in the future, without compromising image quality.

• 4. Viewing of digital films has various added benefits of windowing, enhancing contrast, magnification, and a host of other methods to improve lesion detection, which traditional film-screen radiography would not be able to offer.

2.1.3 Disadvantages and Limitations of Radiography

Despite the improvements brought about by the usage of CR and DR, radiography has some inherent disadvantages and limitations. These include the following:

• 1. Intrinsically poor soft tissue contrast: For example, it is good for detecting a tuberculous lesion within the lung because of wide contrast between soft tissue of the lesion and air within alveoli, but poor for detecting a tuberculous lesion within the renal parenchyma because of lack of contrast with adjacent normal tissue.

• 2. It is relatively insensitive for the detection of small or early bone lesions, particularly in structures with a complex anatomy, for example, early tuberculous osteomyelitis in the scapula.

• 3. Ionizing radiation hazards, albeit small, exist. Standard clinical indications and the ALARA (as low as reasonably achievable) principle should always be followed.

2.1.4 Chest Radiographs

Radiographs remain the mainstay for mass screening and the initial imaging modality for investigating pulmonary TB. The combination of chest radiographs and clinical correlation has been reported to have a 100% sensitivity and negative predictive value for excluding TB (World Health Organization 2018). In addition to being the first-line investigation in suspected cases of TB, chest radiographs are also useful in the follow-up of patients during treatment and are obtained in all patients at the end of anti-TB drug therapy. Documenting the new baseline chest radiographs for these patients, which may show findings such as residual fibrosis and nodular opacities, allows easy detection of new findings upon future comparison (Nachiappan et al. 2017).

The frontal chest radiograph may be obtained as a posterior-anterior (PA) or anterior-posterior (AP) radiograph, with the former being the preferred standard technique. The PA radiograph is usually taken in a relatively well patient who is able to stand, facing the cassette in full inspiration, with the X-ray beam passing through the patient in a posterior-to-anterior direction (Fig. 1a). The lateral chest radiograph is useful for assessing the retrosternal and retrocardiac airspaces, as well as posterior costophrenic recesses for small effusions. The apical view radiograph is particularly useful in the context of pulmonary TB, where the lung apices are often involved by disease. In this view, there is good demonstration of the lung apices, which would otherwise be obscured by the overlying soft tissues, ribs, calcified costal cartilage, and clavicles (Fig. 1b).

Fig. 1

Pulmonary tuberculosis. (a) PA chest radiograph shows patchy opacities mainly in the right upper zone. Fibronodular scarring is also present in the periphery of the left mid zone. (b) Apical view radiograph in the same patient. In this projection, the opacities in the right upper lobe, as well as the left-sided scarring, are much better demonstrated, without the overlying clavicle and ribs

2.1.5 Other Radiographs

Radiographs may not yield much information in many patients with abdominal and/or urogenital TB. However, abdominal radiographs may still be useful in demonstrating focal globular calcifications associated with a granulomatous mass, hepatic or splenic calcifications, and calcified lymph nodes in abdominal TB. In urinary tract TB, a radiograph of the kidneys, ureters, and bladder (KUB) may show triangular ring-like calcifications within the collecting system (Gibson et al. 2004), characteristic renal parenchymal calcifications in a lobar distribution, a calcified “putty kidney,” and/or calcified bladder, particularly in end-stage disease (Engin et al. 2000). When there is involvement of bowel in gastrointestinal TB, ileocecal involvement is common, and may result in small bowel obstruction. The role of the abdominal radiograph is to detect dilated bowel loops and pneumoperitoneum. The erect abdominal radiograph is useful for demonstrating air–fluid levels in obstruction and free air from perforation.

Similar to pulmonary TB and contrary to urogenital and gastrointestinal TB, radiographs still have an important role in assessing musculoskeletal TB. In musculoskeletal TB, radiographs of the spine and extremities are ideally performed in two orthogonal views, i.e., AP and lateral projections. If orthogonal views are not possible, e.g., to evaluate the metatarsals of the foot, then alternatives such as AP and oblique projections are obtained. Special radiographic views may be required for specific bones with complex anatomy, e.g., scaphoid views or open-mouth view for the cervical spine odontoid process. Patients with spinal TB usually have a radiograph of the spine obtained at their initial presentation. Similarly, dedicated radiographs of the affected joint or part are also normally first obtained in patients subsequently known to have tuberculous involvement of large joints, long bones, and small bones of the hands and feet.

However, in early TB, these radiographs may be normal or show non-specific findings. For example, in the early stages of spondylodiscitis, while the initial spine radiograph may be normal, subsequent radiographs may reveal findings such as vertebral body destruction, progressive vertebral collapse, anterior wedging, and gibbus formation, as the disease progresses. In the affected extremity, be it joint, long, or short bones, the initial radiographs may depict non-specific findings of osteopenia, osteolytic foci with poorly defined edges, varying amounts of sclerosis, abnormal periosteal reaction, and/or soft tissue swelling (Engin et al. 2000) (Fig. 2a, b). These patients usually go on to have an MRI, which aims at evaluating the full lesion extent or revealing lesions not well seen on radiographs (Fig. 2c, d).

Fig. 2

Tuberculous arthropathy of the elbow joint. (a) AP and (b) lateral radiographs of the elbow show the small erosions in the proximal ulna (white arrows) and an associated joint effusion. (c) Sagittal and (d) axial fat-suppressed contrast-enhanced T1-W MR images of the elbow show diffuse synovial proliferation and enhancement at both the ventral and dorsal aspects of the elbow joint (white arrows). There is also marrow enhancement in the proximal radius and lateral humeral condyle (asterisks). The extent of these soft tissue changes is not appreciated on the radiographs

2.2 Intravenous Urography

Intravenous urography (IVU) provides a general overview of the whole urinary tract. Prior to imaging, patient bowel preparation, usually starting from one day before, is required and is an important part of the IVU technique. IVU consists of a series of radiographs, starting with a control radiograph obtained to look for any calcifications in the urinary tract (Fig. 3a). A nonionic iodinated contrast agent is then administered intravenously, and radiographs are obtained in the nephrogenic and excretory phases at different timings to comprehensively show the components of the urinary tract. The dose of contrast agent administered varies according to the weight of the patient, usually up to 1.5 ml/kg body weight. A typical series will comprise 1-min coned renal radiograph, 5-min coned renal radiograph with compression, 15-min full-length release radiograph, 30-min coned bladder radiograph, and postmicturition full-length radiograph (Fig. 3b, c). Modifications include addition of oblique radiographs and tomograms.

Fig. 3

Left renal and ureteric tuberculosis. (a) IVU control radiograph shows a cluster of calcifications projected over the lower pole of the left kidney (arrows). (b) Coned renal tomogram, obtained after intravenous contrast administration, shows contrast opacification of the right pelvicalyceal system with none seen in the left pelvicalyceal system. The left lower renal parenchyma is thinned (dotted arrow), and the calculi (solid arrow) seen on the control radiograph are located in the lower pole calyces. The left renal nephrogram is slightly dense with subsequent delayed contrast excretion into the left pelvicalyceal system. (c) Postmicturition full-length radiograph shows a stricture (arrow) at the left proximal ureter, with upstream hydroureteronephrosis. The left renal pelvis is contracted, relative to the dilated proximal ureter and calyces. Most of the calyces, particularly in the upper and mid-poles, are distorted and irregular, consistent with papillary destruction

IVU is relatively sensitive for the detection of renal TB, being normal in only 10–15% of patients known to be affected (Kenney 1990). In TB of the urinary tract, contrast material within the pelvicalyceal system, ureters, and bladder allows filling defects and areas of mucosal irregularity to be demonstrated. It is also able to depict segments of strictures and focal caliectasis, as well as ureteric angulation from scarring. However, while IVU provides a good assessment of the pelvicalyceal system, it yields relatively little direct information about the renal parenchyma, compared to CT urography. IVU also does not have the added benefit of a broader cross-sectional assessment, which enables a simultaneous search for extrarenal features of TB in the rest of the abdomen and spine, which is a great benefit of CT urography. For these reasons, in modern practice, IVU has largely been replaced by cross-sectional imaging techniques, particularly CT urography.

2.3 Contrast Fluoroscopy Studies

These are imaging studies used to evaluate various structures, particularly the gastrointestinal tract, typically in double contrast utilizing air and barium, under real-time fluoroscopic imaging. In a barium swallow and meal, the patient drinks a barium suspension and swallows fluid or tablets that produces effervescent gas. The aim is to coat the esophagus, stomach, and duodenum, as well as distend these structures, in the evaluation of the upper gastrointestinal tract. In a small bowel series, the small bowel may also be subsequently imaged with both real-time fluoroscopy and serial radiographs after ingestion of diluted barium suspension (Fig. 4). This is sometimes also known as a barium meal and follow-through study (BMFT). Barium enteroclysis, in which a nasogastric tube is used to intubate the proximal jejunum, enables better small bowel distention and detection of mucosal abnormalities, compared to BMFT. When evaluating the large bowel, there is a need for prior bowel preparation, and the patient should be able to tolerate air insufflation and barium introduced into the large bowel via a rectal tube. Furthermore, these studies require the patient to have a certain level of mobility, as they have to be able to turn and adopt various positions on the fluoroscopic table, in order to thoroughly coat the colonic mucosa with barium and to obtain images of different segments of the colon in various projections.

Fig. 4

Ileocecal tuberculosis. Small bowel series. Abdominal radiograph taken 3 h after the start of the study shows a short segment of luminal narrowing in the terminal ileum (black arrows), as well as mild deformity of a contracted cecum (white arrow). This patient had subsequent histological confirmation of terminal ileal tuberculosis

These double-contrast barium studies are able to identify ulcers, mucosal irregularity, strictures, as well as fistulas and have been extensively used in the past for the evaluation of gastrointestinal TB, particularly in the region of the ileocecal junction, which is the commonest site for tuberculous involvement of the bowel (Nakano et al. 1992; Leder and Low 1995). These studies are being increasingly replaced by CT, which does not require as much patient cooperation compared to barium studies and has the added benefit of evaluating bowel wall abnormalities and extraintestinal pathology. Regardless of the type of study the patient undergoes, these patients would usually be further evaluated with gastro- or colonoscopy, to complement the radiological findings and also obtain tissue for histopathological confirmation.

A water-soluble contrast agent (usually a nonionic iodinated contrast agent) may also be used in the assessment of the gastrointestinal tract; however, it is usually only used when barium is contraindicated, e.g., in suspected cases of perforation, and in the post-operative patient to look for anastomotic leaks. In these cases, the leakage of barium into the peritoneum may be complicated by barium peritonitis, acutely resulting in large-volume exudative ascites and hypovolemic shock. In the later stages, granulomas, fibrosis, and adhesions may develop, leading to small bowel obstruction. Water-soluble contrast follow-through series of the small bowel is also often performed in cases of small bowel obstruction to determine the transit time of contrast material into the large bowel, which would help surgeons to predict which patients may be managed conservatively. Thus, water-soluble contrast studies of the gastrointestinal tract have a limited but useful role in the diagnosis of gastrointestinal TB. Water-soluble contrast agents may also be used in sinography/fistulography, where the contrast agent is injected via a cannula into a cutaneous opening, prior to imaging with fluoroscopy or CT, to demonstrate possible communication with bowel and other deep structures, e.g., abscess, which will opacify with contrast material if there is a fistulous communication.

2.4 Ultrasound Imaging

Ultrasound (US) imaging is an excellent modality for targeted assessment of a superficial body part or organ in patients with extrapulmonary TB. In US imaging of enlarged cervical lymph nodes, features such as hypoechoic cortical thickening, decreased hilar fat, decreased internal vascularity, and necrosis may be detected. US imaging may also be used to assess the kidney in renal TB. However, it is less sensitive in identifying isoechoic masses, small calcifications, and cavities that communicate with the pelvicalyceal system (Gambhir et al. 2017). US imaging also allows assessment of the prostate gland, testes, and epididymis in urogenital TB, although findings may be non-specific. In musculoskeletal TB, it is particularly helpful in the evaluation of superficial soft tissue masses, tendons, synovium, and bursal spaces, as well as to look for abscesses and joint effusions (Fig. 5). It also allows quick comparison with the contralateral side or extremity, which is not possible with other imaging modalities, e.g., MRI, where only the affected side is imaged. There are several other specialized US imaging techniques, e.g., endoluminal US imaging to enable clearer visualization of bowel wall abnormalities, transrectal ultrasound imaging to evaluate prostate TB, and contrast-enhanced US imaging.

Fig. 5

Tuberculous abscess of the elbow. Longitudinal US image shows a predominantly anechoic fluid collection (between small yellow cursors), with few thick internal septations, in the subcutaneous plane on the posterior aspect of the elbow, with overlying skin thickening (asterisk). This fluid collection was subsequently aspirated and the diagnosis of a tuberculous abscess made

2.4.1 Advantages of US Imaging

• 1. There is no ionizing radiation hazard, as it utilizes sound waves, making US an eminently suitable imaging technique for young and pregnant patients.

• 2. It is portable and is readily available at the bedside for ill patients in the emergency department and ICU. US imaging is also useful in outpatient clinics.

• 3. It provides images in real time and can therefore be used in various interventional procedures. These include guiding biopsy needles in obtaining tissue for the histopathological confirmation of TB, fine-needle aspiration cytology (FNAC) of superficial tuberculous masses and lesions, as well as drainage of pleural effusions and abscesses in patients with TB.

2.4.2 Disadvantages and Limitations of US Imaging

• 1. Its range of anatomical assessment is limited as US does not cross tissue–bone and tissue–gas boundaries, thus preventing the evaluation of deeper structures beyond these barriers. Therefore, US imaging is not generally used as a modality in the investigation of pulmonary TB, except when looking for pleural effusion. In the abdomen, the presence of air in the stomach may obscure deep structures such as the pancreas. Abnormalities in areas surrounded by or deep to bony structures, e.g., hip joint, will be difficult to image.

• 2. It may be technically challenging in patients with an abnormally large body habitus, which limits the extent of examination.

• 3. US imaging is operator-dependent, i.e., much depends on the experience and expertise of the ultrasonologist, particularly for the musculoskeletal system. The quality of the equipment, e.g., use of high-resolution probes, is also important.

2.5 Computed Tomography

CT is the modality of choice for comprehensive assessment of the thorax, abdomen, pelvis, and head and neck and is also useful for the brain and musculoskeletal system. CT is rapid to perform and well tolerated by most patients, including those who are ill or less cooperative. There are many techniques to optimize image quality, such as choice of imaging parameters, multiplanar reconstructions, administration of contrast agents, and postprocessing techniques.

2.5.1 Image Optimization

Patient positioning is important in order to avoid beam hardening artifacts from the upper limbs, which may limit assessment in the region of concern. When imaging the head, neck, and cervical spine, the patient’s arms should be kept down at the side of the body. For the same reason, the arms are raised for imaging of the thorax, abdomen, pelvis, and thoracolumbar spine. A scanogram is initially obtained in order to plan the anatomical length and field of view (FOV) to be examined. This is important as decreasing the FOV to fit the region of interest helps to improve spatial resolution. Optimal kilovoltage peak is dependent on the indication for the scan and size of the patient, while tube current is modulated automatically in current machines. High-resolution filters are used for imaging of the lung and bones, while standard filters usually suffice for all other parts. The scan direction is usually craniocaudal.

Administration of an intravenous nonionic iodinated contrast agent is an important means for image optimization. It allows the delineation of abnormal from normal structures by improving the differential contrast enhancement of the various tissues, e.g., structures in the thorax, abdomen, and pelvis. An 18–20G intravenous cannula is set, ideally in the antecubital vein, in order to withstand the consistent high flow rate of 3–5 ml/s produced by a mechanical contrast injector. A test dose of saline is introduced to check if the cannula is properly set within the vein, as well as to test its patency, prior to contrast administration. This is to prevent contrast extravasation into the arm. Following intravenous contrast administration, a saline chaser is used to flush the contrast material along the arm veins, thus fully utilizing the injected contrast material.

It is worth mentioning that overt hyperthyroidism is an absolute contraindication to the intravenous administration of iodinated contrast material. Additionally, it should also not be administered in patients with known severe contrast allergy, severe asthma, or end-stage renal disease who are not on hemodialysis. Patients with known minor allergic reactions to contrast agents and asthma should be given corticosteroids prior to intravenous contrast administration. Patients with renal dysfunction should be hydrated prior to and/or after the study and have their renal function monitored. Patients with renal dysfunction and on metformin are at risk of developing lactic acidosis and should have the medication withheld for a total of 48 h after the study. As the exact protocol for patient preparation differs among institutions, it is prudent to check the local departmental guidelines.

2.5.2 Roles and Advantages of CT

• 1. CT is the modality of choice for the cross-sectional evaluation of patients with TB, particularly involving the thorax, abdomen, pelvis, and head and neck.

• 2. It has the benefits of multiplanar reconstruction in the coronal and sagittal planes and volume-rendering/three-dimensional (3D) reconstruction and allows the images to be viewed in the lung, soft tissue, and bone windows.

• 3. CT has a much faster scan time compared to MRI, and is better tolerated, cheaper, and more accessible.

• 4. CT is the preferred imaging-guided technique in percutaneous biopsies of deeper structures when a histological diagnosis of TB is necessary; these sites include the mediastinum, abdomen, and spine.

2.5.3 Pulmonary TB

CT has an important role in the diagnosis of pulmonary TB, when the initial chest radiograph is normal or inconclusive. Viewing the lung window allows assessment of the lungs and airways for changes indicative of TB (Fig. 6a, b). Added assessment of the soft tissue structures is possible using the soft tissue window, which reveals characteristic tuberculous lymphadenopathy (Fig. 7a), if present. With CT, the diagnosis of pulmonary TB is correctly made in 91% of cases and correctly excluded in 76% of patients (Lee et al. 1996). It also aids in determining disease activity, being able to characterize 80% of patients with active disease and 89% of those with inactive disease (Lee et al. 1996).

Fig. 6

Pulmonary tuberculosis. (a) Axial and (b) coronal CT images taken in the lung window setting. In the right upper lobe, there are areas of consolidation with internal cavitation (white arrows), as well as centrilobular tree-in-bud opacities (black arrows). Sputum cultures confirmed tuberculosis. (c) AP chest radiograph shows consolidation in the right upper and mid zones. However, neither the internal cavitation nor the tree-in-bud opacities could be appreciated, being only depicted with CT

Fig. 7

Pulmonary tuberculosis. (a) Axial CT image taken in the soft tissue window setting shows necrotic right hilar and mediastinal lymph nodes with a hypoattenuating center and rim enhancement (arrows), due to active tuberculosis. Small pleural effusions are also present bilaterally (asterisks). (b) AP chest radiograph of the patient shows a small area of consolidation in the left mid zone, with internal lucencies suggestive of cavitation (arrows). The hilar and mediastinal lymphadenopathy are not visible

In our institution, besides covering the entire thoracic cage, the coverage of CT of the thorax is usually extended further caudally to include the liver and adrenal glands. Images are obtained at 40–60 s after the intravenous administration of 70–80 ml of nonionic iodinated contrast agent at a rate of 2 ml/s, with typical parameters of 110 kVp, pitch value of 1.2, and slice thickness of 3 mm. High-resolution CT (HRCT) is a CT technique which uses a high spatial frequency reconstruction algorithm to postprocess thin-slice images of the thorax obtained, providing exquisite lung detail. HRCT of the thorax is valuable in the detailed assessment of various manifestations of pulmonary TB, such as interstitial lung involvement, bronchiectasis, and cystic lung lesions. CT aortography is useful for the detection of Rasmussen aneurysm and for preembolization mapping of bronchial and nonbronchial arteries.

2.5.4 Urogenital TB

CT urography is a dedicated technique that is preferred over CT of the abdomen and pelvis for imaging of urinary tract involvement in TB, as it best demonstrates all manifestations of renal TB (Gambhir et al. 2017). It has the additional benefit of having an initial unenhanced scan, which allows better detection of calcifications, as well as a delayed excretory phase. With contrast opacification of the pelvicalyceal system and ureters in the excretory phase, it is akin to an IVU, but with the added advantage of cross-sectional images (Fig. 8). CT urography comprises an unenhanced study, as well as a portovenous phase at 80 s and an excretory phase at 10 min, after intravenous injection of 80 ml of nonionic iodinated contrast agent at a rate of 4 ml/s. Its scan coverage extends from the diaphragm through to the level of the pubic symphysis, with typical parameters of 120 kVp, pitch value of 0.6, and slice thickness of 3 mm.

Fig. 8

Right renal and ureteric tuberculosis. Oblique coronal CT urography image reconstructed to display the whole length of the ureters shows a long segment of dilated right ureter with irregular wall thickening (long arrows). There is also right hydronephrosis and a large cavity (short arrows) in the right upper pole renal parenchyma. There is no normal contrast excretion into the right pelvicalyceal system. The left kidney, left ureter, and bladder are normal, with normal pelvicalyceal system opacification

2.5.5 Abdominal TB

CT is the mainstay of investigation of patients with abdominal TB, the findings of which are generally not evident on radiographs. It demonstrates tuberculous lymphadenopathy, which is the commonest manifestation of abdominal TB (Engin et al. 2000) (Fig. 9). It also comprehensively depicts the features of tuberculous infection of various organs such as the liver, spleen, adrenal glands, gastrointestinal tract, and peritoneum (Fig. 10). A routine CT of the abdomen and pelvis is obtained typically at 120 kVp, with a pitch value of 0.6 and slice thickness of 3 mm, scanned in the portovenous phase at 60 s, utilizing 60–70 ml of intravenous nonionic iodinated contrast agent in the average-sized patient, injected at a rate of 2 ml/s. Care should be taken to ensure that the cranial extent of the scan covers the hepatic dome, with the caudal landmark being the pubic symphysis.

Fig. 9

Tuberculous lymphadenopathy. Axial contrast-enhanced CT image shows a cluster of enlarged, low-density aorto-caval (black arrow), left para-aortic (paired black arrows), and mesenteric lymph nodes (white solid arrows). In this background, there are also a few calcified nodes (white dotted arrows) resulting from previously treated tuberculosis

Fig. 10

Ileocecal tuberculosis. (a) Axial and (b) coronal contrast-enhanced CT images show mural thickening at the ileocecal junction (long arrows) with adjacent inflammatory stranding (short arrows) and free fluid. This patient subsequently underwent colonoscopy, with histopathological confirmation of tuberculosis

CT enterography is an improved technique for the evaluation of the small bowel. It involves ingestion of 1.5–2 L of neutral or low-density contrast material (e.g., water–methylcellulose solution, polyethylene glycol, and low-density barium) over 45–60 min, enabling good distension of small bowel and better visualization of wall and intraluminal abnormalities. Virtual CT enteroscopy is a recent technique in which the small bowel is cannulated with a nasogastric tube and distended using carbon dioxide, following which CT images are obtained and reconstructed to produce endoluminal images of the small bowel. This technique produces superior images of mucosal wall and intraluminal abnormalities.

2.5.6 Cranial TB

CT may be helpful in detecting tuberculous pachymeningitis, cranial tuberculomas, and cerebral tuberculous abscesses. However, it is not the preferred modality in the assessment of central nervous system (CNS) TB, if MRI is available. The initial scanogram obtained should extend from the vertex of the skull to the level of C1 vertebra caudally. Typical parameters are 120 kVp, pitch value of 0.55, and slice thickness of 5 mm. For the average patient, 60 ml of intravenous nonionic iodinated contrast agent is administered at a rate of 2 ml/s, and the scan is obtained 1 min later. A CT venogram may also be performed to assess for complications of venous thrombosis in intracranial TB; this technique has some differences in the scanning protocol for a routine cranial CT. Firstly, the coverage for a CT venogram extends further caudally to the level of C3 vertebra to include a longer segment of the internal jugular veins. Scanning is also done earlier, at 40 s and at a faster rate of 3–5 ml/s, utilizing 70 ml of intravenous nonionic iodinated contrast agent.

2.5.7 Head and Neck TB

In the neck, CT may help in the assessment of pharyngeal and laryngeal TB and is ideal for demonstrating cervical lymphadenitis. As it is a better modality compared to MRI for showing calcifications, CT allows better appreciation of fibrocalcified nodes that may be found in patients treated for TB. The scan coverage for a neck CT includes the frontal sinus and extends caudally to the level of the tracheal carina. The images are obtained 40 s after the administration of 50–60 ml of nonionic intravenous contrast agent, at a rate of 2 ml/s, with typical parameters of 120 kVp, pitch value of 0.8, and slice thickness of 3 mm.

CT is also used for the evaluation of sinonasal disease, although findings are non-specific and are often similar to those of the more common infections. CT coverage for the sinuses extends from the vertex of the skull to the mandible and includes the ears and tip of the nose. Unenhanced images using a bone algorithm, as well as contrast-enhanced scans in soft tissue window settings, are obtained, following intravenous administration of 50 ml of nonionic iodinated contrast agent, delivered at a rate of 2 ml/s. Typical CT imaging parameters are 100 kVp, pitch value of 0.55, and a slice thickness of 3 mm.

CT of the temporal bone is essential in the diagnosis of tuberculous otomastoiditis. It allows assessment of the ossicles, walls of the tympanic membrane, and inner ear structures and reveals the presence of retroauricular abscesses. Images are acquired using a bone algorithm, with a slice thickness of 0.4 mm, at 120 kVp, and a pitch value of 0.55, which includes the frontal sinus to the level of the upper palate, with care taken to include both ears in the lateral extents.

2.5.8 Musculoskeletal TB

When viewed in bone window settings, CT is able to depict the degree and extent of cortical erosions, osseous destruction, and small calcifications in far greater detail compared to radiographs and MRI. However, assessment of soft tissue structures, intervertebral discs, and bone marrow is limited and inferior to MRI. CT has a useful and selected complementary role to radiographs and MRI in the diagnosis of tuberculous spondylitis, tuberculous arthritis, and tuberculous osteomyelitis. When imaging the upper or lower extremities, the entire region of interest should be included and scanned with typical parameters of 120 kVp, pitch value of 0.8, and slice thickness of 3 mm.

2.5.9 Dual-Energy CT

In dual-energy CT (DECT), two sets of X-ray sources and detectors are used to simultaneously acquire CT attenuation data at two different energy levels, namely 80 or 100 kVp and 140 kVp. This technique allows interrogation of different tissues in the body and how they behave at these different radiation energy levels. Initially, DECT was used in the detection of uric acid stones in the urinary tract. Its use, however, has been gradually extended to musculoskeletal imaging, starting with identifying uric acid deposition within various soft tissues in gout, and subsequently evolving into the detection of bone marrow edema. This technique allows calcium to be subtracted from cancellous bone, creating a virtual noncalcium image from an unenhanced image, thus depicting marrow edema in trauma (Pache et al. 2010).

Additionally, visual detection of attenuation changes in the bone marrow could be improved with the use of color-coded maps of the virtual noncalcium subtracted images (Pache et al. 2010). Extrapolating this, the use of DECT in the detection of marrow edema in trauma may be extended to the detection of marrow edema in tuberculous spondylitis, particularly in patients who are unable to undergo MRI. However, this does not replace MRI, which remains the gold standard for identifying marrow edema. Furthermore, there remains the inability of DECT to depict marrow alterations directly subjacent to cortical bone as a result of masking of the cortex and spatial averaging (Pache et al. 2010).

2.5.10 Disadvantages and Pitfalls of CT

• 1. Artifacts can occur, in particular, movement, beam hardening, and streak artifacts (Fig. 11) due to very high attenuation materials, and may cause problems for image interpretation. Movement artifacts can occur with swallowing in neck CT, breathing in CT of the thorax and abdomen, and movement of the patient in general. Thus, clear instructions to the patient are important. They need to be informed that they need to stay still during the duration of the scan and to carefully follow the instructions given. Shorter CT scan times also aid in reducing/removing movement artifacts, and this is generally achieved with new-generation CT scanners. Beam hardening artifacts occur when the X-ray beam “hardens” and mean energy increases, producing a dark area which obscures an area of interest. This occurs when the lower energy photons are absorbed more rapidly and the higher energy photons pass through a very dense area; for example, the posterior fossa in cerebral CT due to the close presence of dense petrous bones in the base of skull is a common site (Fig. 12). Manufacturers utilize filtration, calibration correction, and beam hardening correction software to overcome this artifact (Barrett and Keat 2014).

• 2. High radiation doses and radiation protection should be considered. It is advisable to tailor the imaging parameters in order to adhere to the ALARA principle. For example, using a high pitch, low kVp, and mAs, and limiting the region of interest, should be considered, especially in children. Increasing pitch decreases radiation dose proportionally. Similarly, there is a proportional increase in radiation dose with increase in tube current; doubling the tube current time product doubles the exposure to the patient. Tube current modulation is an essential imaging tool, which allows higher values of tube current in higher attenuation regions; the converse holds true for lower attenuation regions, potentially reducing the radiation dose. Modulation of tube current is also done along the length of the patient. Hence, the CT scanogram helps to appropriately lower the estimate of tube current modulation in relation to differing patient sizes without compromising imaging quality. Modulation of tube current has been known to reduce radiation dose by up to 40% per examination, providing appropriate settings catered to individual patient size and examination type, while maintaining consistency in image quality (Mayo-Smith et al. 2014). Likewise, a decrease in tube voltage will reduce radiation dose; for example, reducing tube voltage from 140 kV to 120 kV will reduce patient exposure and radiation dose by up to 35% (Huda and Mettler 2011). In recent times, automated tube voltage-assisted technology and selection software have been aiding automation of tube voltage based on different CT examination types, as well as each individual patient’s attenuation profile gathered from the initial CT scanogram, while providing diagnostically acceptable image quality (Mayo-Smith et al. 2014).

• 3. CT does not allow full assessment of the pathological changes in the brain, spinal cord, intervertebral disc, and bone marrow. It is also less sensitive than MRI in the depiction of meningeal abnormalities (Chaudhary et al. 2017). MRI has a much greater range of soft tissue contrast and better shows anatomical details in the brain and cord, thus providing greater appreciation of abnormalities within and distinct from normal structures. Similarly, MRI remains the gold standard in detecting discal and marrow edema, as well as assessment of the spine and paraspinal structures and is the modality of choice in spinal infections (Leone et al. 2012).

Fig. 11

Streak artifacts in two different patients. (a) Axial CT image shows prominent streak artifacts (black arrows) from dental amalgam, which limits assessment of the adjacent structures in the mandible. (b) Sagittal CT image in another patient shows streak artifacts (white arrows) from dental amalgam, which extends posteriorly to the cerebellum

Fig. 12

Beam hardening artifacts. Coronal CT image shows beam hardening artifacts across the cerebellum (arrows), caused by dense petrous bone at the skull base

2.6 Magnetic Resonance Imaging

MRI is a non-invasive imaging technology, which produces high-quality anatomical images in different orthogonal planes. A powerful magnet magnetizes hydrogen protons found in water that make up living tissues in the patient’s body, such that they align with the main magnetic field. Radiofrequency (RF) pulses are employed to stimulate and detect the responses of these protons, enabling MRI sensors to create images that show the differences among various types of tissues, normal or abnormal, based on these magnetic properties. It is often used for detection, diagnosis, and treatment monitoring of various diseases, including TB.

2.6.1 Advantages of MRI

• 1. There is no ionizing radiation hazard. For pregnant and young patients where comprehensive cross-sectional imaging of a body region or organ is required and in which ionizing radiation should be avoided, MRI is preferable to CT.

• 2. MRI allows direct acquisition of images in different planes without having to reposition the patient.

• 3. The superior soft tissue contrast of MRI makes it the imaging modality of choice for providing exquisite anatomical and pathological detail and is excellent for imaging of the brain, meninges, spinal cord, and spine, as well as joints and extremities, in CNS and musculoskeletal TB, respectively. It is the gold standard for imaging of tuberculous spondylodiscitis, demonstrating well the local extent of disease and any complications (Gambhir et al. 2017) (Fig. 13). MRI is also considered superior to CT and is the modality of choice in the detection and assessment of CNS TB (Trivedi et al. 2009; Skoura et al. 2015) (Fig. 14). In particular, it is more sensitive than CT in depicting abnormalities in meningeal TB (Chaudhary et al. 2017) (Fig. 15).

Fig. 13

Tuberculous spondylodiscitis. (a) AP and (b) lateral radiographs of the thoracic spine show the right lateral remnant of a severely compressed and largely destroyed T8 vertebral body (black arrow), loss of the normal intervening T7/8 and T8/9 disc spaces, as well as a large associated paravertebral mass (white arrows). (c) Sagittal contrast-enhanced CT image better shows the severely compressed T8 vertebral body with small retropulsed bony fragments (black dotted arrow). Large anterior prevertebral abscess (white solid arrows) and epidural extension (black solid arrows) are seen. (d) Corresponding sagittal STIR MR image shows hyperintense T7 and T9 vertebral bodies, with the severely compressed T8 vertebral body being barely visible. The T8 vertebral posterior elements and adjacent soft tissues show hyperintense signal (dotted arrows), indicating the full extent of infection. The anterior subligamentous spread of the paravertebral abscess (arrows) and epidural phlegmon causing cord compression are well shown

Fig. 14

Cerebral tuberculosis. (a) Axial T2-W MR image of the brain shows numerous tiny subcentimeter hyperintense foci in the centrum semiovale bilaterally (arrows). (b) Coronal contrast-enhanced FLAIR MR image shows multifocal enhancing foci scattered in the cerebrum and cerebellum (arrows). These lesions are better demonstrated following intravenous Gd-based contrast agent administration. (c) Coronal unenhanced CT of the same patient. These tiny subcentimeter foci seen on MRI are hard to identify in the corresponding CT image, particularly without intravenous contrast administration

Fig. 15

Leptomeningeal tuberculosis. (a) Axial FLAIR MR image shows abnormal hyperintense signal outlining the pons (arrows). Suppression of CSF signal in FLAIR sequences accentuates this area of leptomeningeal abnormality, which would otherwise be difficult to detect on corresponding axial (b) T1-W and (c) T2-W images. (d) Corresponding axial contrast-enhanced FLAIR MR image shows leptomeningeal enhancement. The use of intravenous Gd-based contrast agents better depicts leptomeningeal involvement

2.6.2 Pulse Sequences

Spin-echo (SE) T1-weighted images are acquired using a short time-to-repetition (TR) of <800 milliseconds (ms) and short time-to-echo (TE) of <30 ms. It is the favored sequence for the detailed depiction of anatomical structures. It is also the sequence used for imaging after intravenous administration of gadolinium (Gd)-based contrast agents, improving detection of enhancing tuberculous lesions, and allows assessment of its enhancement characteristics and extent of infection. Although Gd-based contrast agents may be administered by other means, e.g., intra-articularly in MR arthrography, these techniques have very limited applications in the context of patients with TB.

SE T2-weighted images are obtained using a long TR of >2000 ms and long TE of >60 ms. This sequence is sensitive to the presence of fluid, making it particularly useful in the detection of edema and pathological lesions, which typically have hyperintense T2 signal. Hence, this sequence is ideal for the detection of diseases such as tuberculous infection. Being time-saving, fast or turbo SE (TSE) sequences are currently often used in routine MRI in place of SE T2-weighted sequences. As fat appears hyperintense on TSE T2-weighted images, fat suppression is usually applied to differentiate the abnormal signal from the background fat, particularly in areas where the suspected lesion is in close proximity to fat, e.g., TB of the spine and musculoskeletal system.

The proton density (PD) sequence reflects the density of protons rather than the magnetic characteristics of hydrogen nuclei and is an intermediate sequence sharing features of both T1- and T2-weighted images. It has a long TR of >2000 ms and short TE of 20 ms, minimizing T1 and T2 differences, respectively. This sequence is used in musculoskeletal imaging as it is ideal for the assessment of the joints by providing excellent signal distinction between fluid, hyaline cartilage, and fibrocartilage.

The short tau inversion recovery (STIR) sequence has a TR >2000 ms, TE >30 ms, and inversion time (TI) of 120–150 ms. As fat has a relatively shorter T1 compared to other tissues in the body, its signal may be selectively nulled with this fat suppression sequence, which aims to increase the signal intensity difference between abnormal fluid and adjacent fat. While the imaging quality is usually not as good as that of conventional fat-suppressed T2-weighted images, it is less susceptible to magnetic field inhomogeneity and is favored in certain indications, e.g., where the patient has metallic implants close to the location of the tuberculous lesion. It is also important to note that the signal suppression in STIR is not specific to fat and may also nullify signal from any material with a short T1, e.g., melanin, methemoglobin, proteinaceous fluid, and more importantly gadolinium. Therefore, the STIR sequence cannot be used to demonstrate post-Gd-based contrast enhancement, and it is its most noteworthy limitation.

The fluid-attenuated inversion recovery (FLAIR) sequence is a special inversion recovery sequence with a long TI. Its typical acquisition parameters are TR >3000 ms, TE >80 ms, and TI of 1700–2200 ms. It is essential in imaging the CNS in TB, being able to suppress fluid signal from cerebrospinal fluid (CSF), thus accentuating parenchymal edema, and is helpful in distinguishing periventricular subependymal edema from obstructing hydrocephalus. For the same reason, it also better demonstrates exudates and meningeal enhancement, following Gd-based contrast administration (Fig. 15d).

Diffusion-weighted imaging (DWI) is an MRI technique that detects changes in the Brownian motion of water molecules within tissues (Brant 2012). In routine clinical practice, DWI b-values between 0 and 1000 are typically used; for example, a b-value of 1500 is used in MRI of the prostate. The b-value is a factor that reflects the timing and strength of gradients used to generate these images; the higher the b-value, the stronger the diffusion effects depicted. DWI is used to identify hypercellular tissue, which demonstrates restricted diffusion.

Apparent diffusion coefficient (ADC) measures the magnitude of diffusion of water molecules within tissues and is clinically calculated using MRI with DWI (Sener 2001). Higher ADC values indicate more mobile water molecules, while low ADC values indicate restricted movement of water molecules in tissues (Mukherjee et al. 2008). In neuroimaging of TB, the DWI sequence allows detection of cerebral infarction (Rodriguez-Takeuchi et al. 2019). This sequence may also help differentiate pyonephrosis from hydronephrosis in renal TB and spondylodiscitis from disc degeneration in spinal TB. Generally, in tissues infected by TB, restricted diffusion may be demonstrated; this may make it a useful sequence in identifying other sites of involvement, rather than just differentiating TB from other disease processes (Dunn et al. 2015).

Gradient-recalled echo (GRE) sequences are used to acquire fast images and are therefore useful in minimizing motion artifacts from breathing, heartbeat, and vessel pulsation, as well as peristalsis. Magnetization decay time in GRE is termed T2* and is much shorter than the T2 decay times in SE imaging. On GRE images, signal intensity arising from T2 relaxation characteristics of tissue is strongly affected by imperfections in the magnetic field. By utilizing GRE T2*-weighted MRI sequences, in addition to relatively long TE values, local magnetic field homogeneity effects are accentuated which is useful to detect blood products, iron, or calcifications (Brant 2012).

Susceptibility-weighted imaging (SWI) is a high-spatial-resolution 3D GRE MRI technique, which has increased sensitivity in the identification of blood products and/or calcium, which may not be as apparent on T2*-GRE imaging; but there are potentially more false positives in SWI. Additionally, SWI has the ability to distinguish paramagnetic (e.g., hemorrhage or iron) from diamagnetic (e.g., calcification) substances using filtered phase postprocessing images, where they demonstrate opposed signal intensity (Tong et al. 2008).

Chemical shift imaging, also known as opposed-phase or in- and out-of-phase imaging, is used to detect the presence of intracellular or microscopic fat, by taking advantage of the differences in precessional frequencies between fat and water. It is acquired simultaneously using two different TEs; the echoes are timed to coincide with out-of-phase and in-phase timings of the relevant spins, and the data is subsequently divided into two different image sets covering the same anatomical region (Roth and Deshmukh 2017). The presence of intracellular or microscopic fat is then identified when there is significant reduction in signal intensity on out-of-phase images.

The Dixon method is a fat suppression sequence, which is based on chemical shift imaging. This technique uses a nonspectrally selective pulse with TEs, which capture phase differences between fat and water protons, which, over time, alternate between being in-phase and opposed-phase. Acquiring both in-phase and opposed-phase images simultaneously allows the images to be either added or subtracted to produce water-only (fat-suppressed) or fat-only (water-suppressed) images, respectively. The Dixon method has the advantage of uniform fat suppression and ability to quantify fat.

All the different MRI manufacturers have developed their own special sequences; most of them are named with easy-to-remember and/or catchy acronyms. For example, Siemens’ fast low angle shot (FLASH) and volumetric interpolated breath-hold examination (VIBE) are spoiled GRE sequences, which utilize RF spoiling to eliminate transverse magnetization prior to each RF pulse. With FLASH, both T2*- and T1-weighted imaging may be obtained, and with its short repetition time enabling images to be acquired within a single breath hold, it is commonly used in abdominal imaging. VIBE is a modified form of FLASH, which allows high-resolution dynamic images to be acquired. Using fast 3D GRE sequences to produce T1 images, it is a form of volumetric imaging used in abdominal imaging to acquire dynamic contrast-enhanced images using short breath-hold-length acquisition times.

The double-echo steady state (DESS) is a 3D GRE sequence by Siemens, which produces higher T2*-weighted imaging. It is mainly used in musculoskeletal imaging, providing high-resolution cartilage imaging, as well as synovial fluid imaging. Half-Fourier acquisition with single-shot turbo spin echo (HASTE) is a product by Siemens, which uses a single-shot technique to acquire only slightly more than half of the required data within a single TR for an entire T2-weighted image. This is made possible by half-Fourier transformation, which enables computer reconstruction of the remaining data that was not acquired, thus reducing scan times, allowing for fast breath-hold imaging (Regan et al. 1998). Sampling perfection with application-optimized contrasts using different flip angle evolution (SPACE) is an isotropic 3D FSE acquisition sequence by Siemens. It enables acquisition of T1-, T2-, FLAIR-, and PD-weighted imaging in high-resolution 3D datasets.

MR spectroscopy (MRS), on the basis of the chemical shift phenomena, provides information about the presence and concentration of metabolites in tissue (Brant 2012). The more common metabolite peaks measured are lactate, N-acetylaspartate (NAA), citrate, creatine, and choline, and these change with different pathologies. For example, MRS helps to differentiate tuberculoma from pyogenic abscesses and neoplasms (Gupta et al. 1995).

Magnetization transfer (MT) imaging manipulates differences in relaxation of freely mobile unbound water protons, immobile protons with restricted motion (macromolecular pool), and protons at the boundary, where exchange of magnetization transfer occurs (Yousem and Grossman 2010). When MT pulses are applied to the macromolecular pool, some of this energy is transferred to the free water pool, which would partially saturate, resulting in reduced signal due to the MT effect. The MT effect can be quantified by obtaining two sets of images, one before the MT pulse (x) and the other with the MT pulse (y), and subsequently subtracting them digitally (x – y). The magnetization transfer ratio (MTR) for a given voxel is then taken to be (x – y)/x. MT imaging is used most frequently in time-of-flight (TOF) magnetic resonance angiography (MRA) to suppress the signal intensity of the background brain, thus improving visualization of the vessels. In combination with contrast-enhanced MRI, it also improves conspicuity of white matter-enhancing lesions with its greater suppression of white matter signal compared to gray matter (Yousem and Grossman 2010).

TOF MRA or MRV is an MRI technique used to visualize flow within arteries and veins, respectively, using a GRE sequence, without the need for intravenous contrast administration. In TOF imaging, unsaturated spins moving into the image slice have high signal, while the saturated stationary ones have low signal; this effect is thus termed flow-related enhancement. The source images are then postprocessed using different algorithms to depict only flowing blood, seen as the brightest pixels, and projected in any plane to better demonstrate the vascular anatomy in multiple views (Yousem and Grossman 2010).

When dynamic contrast-enhanced GRE T1-weighted sequences are performed to acquire contrast-enhanced MRA/MRV images, sequential images in different phases are obtained; the “first-pass” images are acquired in the arterial phase in the vessel of interest, and subsequent images acquired have varying degrees of mixed arterial and venous enhancement. From these, arterial as well as selective venous phase studies in the vein of interest may be obtained by subtracting the arterial phase study from the mixed arterial and venous phase study. The use of Gd-based contrast agent results in blood appearing bright.

2.6.3 Patient Positioning and Coils

It is imperative that the patient is in a comfortable position for the relatively long duration of the MRI scan. Typically, patients lie supine, “head in” first into the scanner gantry. The “feet in” first position is used when imaging the lower extremity and may be attempted in patients who are claustrophobic when imaging the spine and abdomen. Other positions may be adopted according to clinical circumstances, e.g., decubitus position for a patient with kyphosis. Pads, sponges, straps, and other mobilizing devices may be used to help the patient to keep still and maintain their position on the MRI couch.

Once within the bore of the magnet, the patient is surrounded by a series of coils, each with its own function. Working from outwards to inwards, the outermost coils are shim coils, which are fine-tuned to improve magnet field homogeneity and maintain as uniform a main magnetic field as possible throughout the imaging process. The role of gradient coils is primarily to allow spatial encoding of the MRI signal. Lastly, the RF coil can be used as a transmitter, generating RF magnetic field pulses perpendicular to the static main magnetic field, and/or used as a receiver for detecting the MRI signal emitted by the excited hydrogen protons within the tissues in the body part under examination. RF coils may also be grouped into volume and surface coils.

There are several types of RF coils including the following:

• 1. Volume coils: These are designed to provide homogeneous RF excitation across a large volume, thus providing a better magnetic field homogeneity compared to surface coils. The largest volume coil is the standard body coil, which is both a transmit and receive coil, incorporated as part of the scanner and used in imaging large parts of the body, such as the abdomen, pelvis, and chest. The smaller volume coils, for example, the head coil, also both a transmit and receive coil, are used in imaging of the brain. Small-volume coils are also used in imaging of the neck, cervical spine, and extremities, e.g., in the knee and wrist.

• 2. Surface coils are receiver coils placed as close as possible to the region of concern to be imaged, e.g., the orbits, in order to maximize the signal and obtain a high signal-to-noise ratio (SNR) and resolution. The drawback to this coil is its small FOV, which has been overcome by the advent of phased array coils.

• 3. Phased array coils are receive-only RF coils made up of a collection of four or more receiver coils, which together form a larger array. The individually received signals are combined to increase the SNR with a larger FOV, with all data acquired in a single sequence. An example of such usage will be MRI of multiple segments of the spine (Asher et al. 2010).

2.6.4 MRI Protocols

As a general principle, MRI protocols are designed with the aim of providing answers to the clinical problem by comprehensively evaluating the area of interest, yet in a timely and cost-effective manner. Besides coil selection to cover the specified region of interest and using the correct sequences, several other parameters have to be applied to achieve the best possible images, including the ideal imaging planes, FOV size, slice thickness, and interslice gap. It is worthwhile to emphasize the importance of having the radiologist available to review the images on completion of the scans, with the view of protocol modification, e.g., adding extra or special sequences, extending the scan region of interest, and assessing the need for contrast enhancement.

Routine MRI protocols will differ among institutions and the requirements of individual clinical practices. These differences will not be marked if MRI principles are followed. Even in a single institution, the MRI protocols will not be exactly the same. For example, in our department, we have MRI scanners from different manufacturers and magnets of 1.5 T and 3.0 T field strengths. In the following paragraphs, we have provided sample MRI protocols for the Siemens Avanto 1.5 T scanner in our institution that are used for imaging of suspected infection in various parts of the body. Note that the turbo inversion recovery magnitude (TIRM) sequence is the Siemens equivalent to the STIR sequence.

In imaging the spine in a patient with suspected tuberculous infection, we use a phased array coil system for spine imaging that is built into the table of the MRI machine, allowing imaging of multiple segments of the spine. For the cervical spine, there is an option to add an additional neck matrix coil, with multisegment attachment, that integrates with the spine coil and will produce better quality images (Leow et al. 2021). The scan coverage should cover the entire spinal area of interest, i.e., from above the craniocervical junction to T4 vertebral level in the cervical spine; C6–L1 vertebral levels in the thoracic spine; and T10–S4 levels in the lumbar spine. A sagittal localizer image of the entire spine is first obtained in order to correctly identify the vertebral levels and the presence of a transitional lumbosacral vertebral segment (Peh et al. 1999). MRI protocols for the cervical, thoracic, and lumbar spine are listed in Tables 1–3.

Table 1 MRI of the cervical spine protocol

Table 2 MRI of the thoracic spine protocol

Table 3 MRI of the lumbosacral spine protocol

Patients undergoing MRI of the brain and/or neck for suspected intracranial and head and neck TB are imaged using a combined head/neck coil, or a head coil for the brain and a neck coil for the neck, depending on the MRI machine used. These coils are volume coils that both transmit and receive RF signals. For the brain, the scan coverage should extend at least 1 cm above the vertex of the skull and through the skull base, in its craniocaudal extent, and cover both ears from side to side. For the head and neck, the scan coverage is from the frontal sinus to the distal sternoclavicular joint. MRI protocols for patients in whom brain and head and neck TB is suspected are listed in Tables 4 and 5.

Table 4 MRI of the brain protocol

Table 5 MRI of the neck protocol

Generally, when assessing for abdominal, gastrointestinal, and urogenital TB, as well as tuberculous lymphadenopathy, CT is the preferred imaging modality. MRI of the abdomen is reserved for patients in whom CT is contraindicated. In our institution, MRI of the liver, pancreas, or kidneys is usually reserved for troubleshooting and further characterization of a known lesion that has already been detected with US imaging or CT. A body coil is used when acquiring an MRI of the abdomen or MR urography. The coverage for MRI of the abdomen should include the liver and extend through the lower abdomen, clearing the inferior poles of the kidneys. MR urography covers the entire urinary tract, extending from the superior poles of the kidneys to the urinary bladder. Dynamic contrast-enhanced images of the kidneys are acquired, in addition to delayed imaging in the excretory phase with contrast opacification of the ureters and urinary bladder. Protocols of an MRI of the abdomen and MR urography are listed in Tables 6 and 7.

Table 6 MRI of the abdomen protocol

Table 7 MR urography protocol

MRI is the imaging modality of choice for imaging musculoskeletal TB. A large selection of coils may be used, tailored to the area of interest. This may range from using a whole-body coil to image extensive necrotizing fasciitis of the thigh to a dedicated wrist coil for tuberculous arthritis of the radiocarpal joint. As a principle, images are acquired in two orthogonal planes (usually axial plus either sagittal or coronal) as a minimum for lesions in or around long bones and three orthogonal planes for complex joints (e.g., ankle and knee). In our practice, intravenous Gd-based contrast agent is administered routinely as we have found that the additional information it provides enhances diagnostic confidence. Sample protocols for MRI of the upper limb long bones are listed in Tables 8 and 9.

Table 8 MRI of the humerus protocol

Table 9 MRI of the forearm (radius/ulna) protocol

2.6.5 Disadvantages and Pitfalls of MRI

MRI is generally less accessible, usually has longer waiting times, and is costlier compared to CT. The long acquisition times of MRI require the patient to lie still in the magnet for a relatively long time, and it is hence not ideal for patients with claustrophobia. It may also be difficult to image young children or patients with altered mental status, without first giving sedation. Motion blurring occurs and degrades the MR images, should the patient find it difficult to stay still throughout the duration of the study or hold their breath at the appropriate times. This is particularly important in imaging of the abdomen when breathing motion artifacts limit assessment of the solid organs, peritoneum, and lymph nodes.

It may also not be possible for patients with certain MRI-incompatible implants and foreign bodies to obtain an MRI. Some of these include the following:

• 1. Temporary transvenous pacing wires and abandoned intracardiac pacing wires are absolute contraindications, as current induced by RF pulses could result in thermal injuries.

• 2. Insulin pumps, as well as other drug infusion pumps for chemotherapy and analgesic agents, must also be removed prior to MRI as they may malfunction.

• 3. Certain catheters come with known metallic components, e.g., Swan-Ganz catheter.

• 4. Implantable neurostimulation systems, e.g., deep brain stimulators: There is the risk of possible thermal injury along wires and malfunction of the device.

• 5. Known metallic foreign bodies in the eye could heat up, move, or be displaced during MRI, resulting in injury to the eye and adjacent structures. When in doubt, a radiograph of the orbits should be obtained to search for these metallic objects. This search should also be applied to known metallic fragments elsewhere in the body, e.g., bullets, pellets, and metal shrapnel.

• 6. Smart contact lenses, which are used to record continuous intraocular pressures to guide glaucoma treatment, may result in thermal injury to the eye and must be removed.

This list is not exhaustive, and certain cardiac pacemakers, implantable cardioverter defibrillators, and cardiac monitors, as well as aneurysm clips, intraocular lenses, and cochlear implants, are deemed unsafe. The exact model of all known implants, prostheses, and devices should be cross-checked for MR compatibility prior to MRI.

As the more recent implants and prostheses are usually MR compatible, MRI is now possible for many patients with implants. Despite this, interpretation of MR images may still be challenging, due to image distortion or artifacts relating to or surrounding the implant or prostheses. This is particularly important when assessing a joint, extremity, or spine in the context of musculoskeletal TB. Many of these artifacts (e.g., magnetic susceptibility (Fig. 16), motion (Fig. 17), CSF flow, truncation) can be identified, reduced, or corrected by various measures and techniques (Peh and Chan 2001; Shikhare et al. 2014).

Fig. 16

Susceptibility artifacts. (a) Axial and (b) coronal MR images of the pelvis show susceptibility artifacts due to a left hip prosthesis. The large area of signal void produced by the susceptibility artifact prevents assessment of the affected region

Fig. 17

Motion artifacts. Axial MR image of the liver shows severe motion blurring, making detection of any abnormality difficult. The patient was unable to breath hold

It is important to note that some Gd-based contrast agents have been linked to a very small incidence of nephrogenic systemic fibrosis, a rare debilitating systemic fibrotic condition seen in patients with renal dysfunction (Kaewlai and Abujuneh 2012). Recognizing this disease is very important as it is progressive, potentially fatal and may affect multiple organs. In patients with renal impairment, the prolonged excretory half-life of these agents increases the likelihood of deposition of Gd ions in the skin, bone, liver, and brain. This results in hardening of the skin, contractures, and involvement of the skeletal muscles, including the diaphragm, causing respiratory failure.

In patients with a normal renal function, Gd is excreted by the kidneys with a half-life of 90 min. Current practice requires determining the necessity of intravenous Gd-based contrast administration in answering the clinical question. Assessment of the renal function is recommended in patients not undergoing dialysis and not having acute kidney injury, and according to which type of Gd-based contrast agent is used (Mathur et al. 2020). According to the consensus statement by the American College of Radiology and National Kidney Foundation, since the risk of nephrogenic systemic fibrosis is so low with Gd-based contrast media, the potential harms of delaying or withholding contrast administration for MRI in a patient with acute kidney injury or estimated glomerular filtration rate less than 30 ml/min per 1.73m² are likely to outweigh the risk in most clinical situations (Weinreb et al. 2021).

2.7 Nuclear Medicine Imaging

Nuclear medicine imaging is based on the use of radiopharmaceuticals, which are chemical compounds labeled with single-photon-emitting radioisotopes such as technetium-99m (^99mTc) or positron-emitting isotopes such as [fluorine-18]-fluoro-2-deoxy-d-glucose (¹⁸F-FDG). In general, nuclear imaging is rarely used to diagnose active TB, and its impact on the clinical care of patients with TB is limited. However, ¹⁸F-FDG positron-emission tomography (PET)/CT is an important non-invasive method for assessing disease activity, detecting extrapulmonary TB, and determining the treatment response (Sathekge et al. 2012).

2.7.1 Conventional Nuclear Medicine Imaging Techniques

The most commonly used and studied single-photon-emitting tracers for the diagnosis and management of TB are thallium-201 chloride (²⁰¹Tl-chloride), gallium-67 citrate (⁶⁷Ga-citrate), ^99mTc-sestamibi, and ^99mTc-tetrofosmin. ^99mTc-methylene diphosphonate (^99mTc-MDP) bone scintigraphy is used to evaluate spinal TB, even though it is non-specific.

2.7.1.1 Thallium-201 Scintigraphy

²⁰¹Tl-chloride scintigraphy is usually done in dual phases, with the early phase around 15 min and delayed phase around 3 h after injection. It has the potential to differentiate benign from malignant pulmonary lesions. Ratios of uptake are calculated from the early and delayed tracer uptake, compared to the normal contralateral lung in both phases. The retention index of the lesion, derived from the formula (delayed ratio – early ratio)/(early ratio) × 100%, is significantly different for benign lesions and malignant lesions. The retention index of benign lesions, including tuberculomas, has been found to be significantly low, compared to that of malignant lesions (Suga et al. 1993; Yu et al. 2004). The combination of HRCT and thallium-201 scintigraphy is more useful and relevant clinically, with improvement of sensitivity and specificity to near 100% (Kashimada 1998).

2.7.1.2 ⁶⁷Ga-Citrate and ^99mTc-MDP Scintigraphy

⁶⁷Ga-citrate scintigraphy has a high sensitivity to diagnose suspected active pulmonary TB but has limited specificity. The degree of tracer uptake is proportionally related to the bacterial load in the sputum. This technique is better than chest radiographs and fairly similar to HRCT and ²⁰¹Tl-chloride scintigraphy in detecting parenchymal lesions. They are especially useful in the subgroups of patients with TB who are sputum negative and have disseminated, military, or diffuse pulmonary involvement. ⁶⁷Ga-citrate scintigraphy can also correctly predict the involvement of extrapulmonary sites of TB, such as the spine, peritoneum, and lymph nodes, efficiently (Sathekge et al. 2012). ^99mTc-MDP and/or ⁶⁷Ga-citrate are the most commonly used tracers in spinal TB.

These methods are sensitive and permit whole-body evaluation, but are less specific and have poor spatial resolution. Their limitations can be overcome and improved by single-photon emission computed tomography (SPECT)/CT which provide better resolution and anatomical localization. The three-phase bone scan can evaluate the inflammatory process associated with intense infections of early stage and identify bone remodeling associated with late-stage spondylitis. Combination of ⁶⁷Ga-citrate scintigraphy with bone scan efficiently diagnoses both osseous and soft tissue infections, monitors therapeutic response, and detects reactivation (Rivas-Garcia et al. 2013).

2.7.1.3 Other Single-Photon-Emitting Radiopharmaceuticals

^99mTc-sestamibi and ^99mTc-tetrofosmin have been studied extensively for the evaluation of pulmonary TB and have been found to show increased uptake in active lesions, proportional to the disease activity. Hence, they can be used for differentiation between active and inactive disease and to monitor the therapeutic response with good negative predictive value (Ahmadihosseini et al. 2008). Multiple radioisotopes, such as N-isopropyl-p-[1²³I] iodoamphetamine (¹²³I-IMP), ^99mTc-dimercapto succinic acid (^99mTc-DMSA), ^99mTc-glucoheptonate, indium-111 octreotide (¹¹¹In-octreotide), ^99mTc-hexamethylpropyleneamine oxime-white blood cell (^99mTc-HMPAO-WBC), ^99mTc-ciprofloxacin, and ^99mTc-ethyl cysteinate dimer (^99mTc-ECD), have been studied for the evaluation of pulmonary and extrapulmonary TB, with variable results produced with nominal clinical impact (Sathekge et al. 2012).

Recently, ^99mTc-ethambutol has been evaluated for its diagnostic value in TB, as it specifically binds to mycolic acid in the cell wall of the M. tuberculosis bacteria. Early results have been promising with good sensitivity, specificity, and diagnostic accuracy. ^99mTc-ethambutol scintigraphy can detect and localize both pulmonary and extrapulmonary TB. This procedure has no side effects and can be performed safely, even in pediatric patients (Kartamihardja et al. 2018).

2.7.2 Positron-Emission Tomography/CT

2.7.2.1 ¹⁸F-FDG PET/CT

¹⁸F-FDG is used in the diagnosis of TB, based on its ability to detect increased glucose metabolism occurring in the disease process, due to increased macrophage and neutrophil activity. ¹⁸F-FDG PET/CT is sensitive for detecting diseases such as TB, which can cause both acute and chronic infections and relative uptake quantification. Standard uptake value (SUV) measurement can be used to distinguish between residual active and inactive lesions. Although morphological imaging is the current cornerstone for diagnosing and staging TB, ¹⁸F-FDG PET/CT has been evaluated and proven to be a valuable adjunct for differentiation of active and nonactive lesions, monitoring the treatment response and follow-up (Sathekge et al. 2012; Vorster et al. 2014).

2.7.2.2 Evaluation of Active and Inactive Disease

There are two different patterns of ¹⁸F-FDG PET activity in TB, namely the pulmonary pattern and the lymphatic pattern. In the pulmonary pattern, ¹⁸F-FDG uptake is primarily seen in cavitary and noncavitary consolidation and adjacent micronodules, with mild-to-moderate uptake in mediastinal and hilar lymph nodes. In the lymphatic pattern, increased ¹⁸F-FDG uptake is seen in the enlarged mediastinal and hilar lymph nodes and at extrapulmonary sites of involvement (Fig. 18) (Soussan et al. 2012).

Fig. 18

Disseminated tuberculosis. (a) Whole-body PET MIP image shows increased ¹⁸F-FDG uptake in the right cervicothoracic region and superior mediastinum. Faintly increased ¹⁸F-FDG uptake is seen in left hypochondrium. Corresponding (b–j) axial CT, fusion PET/CT, and axial PET images show increased tracer uptake in the conglomerate right supraclavicular lymphadenopathy (b–d), right paratracheal lymph node and consolidation in the medial right upper lobe medially (e–g), and omental nodule in left hypochondrium (h–j). These lesions in pulmonary and extrapulmonary locations of disseminated tuberculosis are arrowed (courtesy of Dr. Godwin Jeeva, Gemini Scans, Chennai, India)

Active pulmonary tuberculomas usually have significant ¹⁸F-FDG uptake with SUVmax greater than 4. Double-phase acquisition of ¹⁸F-FDG PET/CT acquired at 60 and 120 min postinjection can better differentiate active from inactive lesions, with analysis of multiple-point SUV (SUVmax_E – early phase, SUVmax_D – delayed phase, %DeltaSUVmax – difference). With SUVmax_E of 1.05, SUVmax_D of 0.97, and %DeltaSUVmax of 6.59 as cutoff values, active and inactive lesions can be differentiated with a specificity of 100% (Goo et al. 2000; Kim et al. 2008). Similarly, dual-phase ¹⁸F-FDG PET/CT is also helpful in identifying extrapulmonary sites of TB with more diagnostic accuracy. In a study conducted on 16 patients with TB, ¹⁸F-FDG PET/CT identified sites of lymph node, bone, and joint involvement, which were initially missed on CT (Sathekge et al. 2010a).

2.7.2.3 Differentiation of Malignant Lesions and Tuberculoma

Increased ¹⁸F-FDG uptake is non-specific for tumors, as well as infectious and inflammatory conditions. ¹⁸F-FDG-avid lesions in the TB-endemic regions are always a diagnostic challenge for differentiating between neoplasm and tuberculoma. Unfortunately, there is no significant difference between the SUVmax values of tuberculomas and malignant nodules, and ¹⁸F-FDG PET/CT cannot confidentially differentiate them. The SUV values of involved lymph nodes in TB and malignancy are also not statistically different. Sathegke et al. (2010b) found the SUVmean values of involved lymph node basins to be 6.5 (3.4–9.2 range) in TB and 8.0 (2.5–20.1 range) for malignancy, with significant overlap. These findings are in agreement with the data derived from other studies conducted for the same purpose. Hence, we can safely conclude that ¹⁸F-FDG PET/CT is not indicated for differentiating malignancy from TB (Fig. 19) (Chen et al. 2008; Sathekge et al. 2010b).

Fig. 19

Patient with known right buccal carcinoma and concomitant pulmonary tuberculosis. Axial (a) CT and (b) fusion PET/CT images show increased ¹⁸F-FDG uptake at the site of the primary malignancy (arrow). Staging axial (c) CT, (d) PET, and (e and f) fusion PET/CT images taken in (e) lung and (f) soft tissue window settings show increased ¹⁸F-FDG uptake within an irregular spiculated nodule with small cavitation in the left upper lobe. This was indeterminate for metastases and active tuberculosis. Subsequent histological examination revealed the left lung nodule to be due to tuberculous infection (courtesy of Dr. Godwin Jeeva, Gemini Scans, Chennai, India)

2.7.2.4 Monitoring Treatment Response and Follow-Up

This is the most accepted and important clinical application of ¹⁸F-FDG PET/CT in patients with TB. After completion of treatment, morphological changes of response such as reduction in size and decrease in enhancement often take a longer time to manifest. In endemic areas where multidrug-resistant TB is common, assessment of the treatment response at an early stage is needed, so as to facilitate change or modification of treatment in nonresponders. As there is correlation between decrease in ¹⁸F-FDG uptake and successful treatment, quantitative PET assessment is helpful to evaluate the treatment response. Some lesions may even increase in size, instead of decreasing in size, with good treatment response. In those patients, decreased ¹⁸F-FDG activity in those lesions likely suggests that the tuberculoma is responsive to anti-TB treatment and current treatment should be continued. Similarly, the SUVmax of the involved lymph node basins is significantly higher in pre-treatment scans and interim scans in nonresponders. The metabolic response may indicate clinical response and guide the duration of the antimicrobial therapy (Park et al. 2008; Sathekge et al. 2013; Vorster et al. 2014).

2.7.2.5 Other Positron-Emitting Tracers

Various other positron-emitting tracers, including ¹¹C-choline, ¹⁸F-fluorothymidine (¹⁸F-FLT), and gallium-68 (⁶⁸Ga), have been investigated for diagnosis of TB. Among them, early results of ¹¹C-choline have been encouraging. Hara et al. (2003) concluded that combined use of ¹⁸F-FDG PET/CT and ¹¹C-choline PET/CT can differentiate lung cancers, TB, and atypical mycobacterial infections.

2.8 Interventional Radiology

Interventional radiology has a role, not only in establishing the diagnosis, but also in the management of TB (Nachiappan et al. 2017).

2.8.1 Percutaneous Biopsy

Tissue diagnosis is needed when the imaging features are not classical of TB or mimic other infections or neoplasms. Specimens should be sent for histology, acid-fast bacilli (AFB) staining, and culture. If there is a lack of AFB, differentiating TB from other granulomatous lymphadenitis could be challenging. Polymerase chain reaction (PCR) testing of lymph node tissue for M. tuberculosis may help in the final diagnosis. Biopsy or aspiration can be performed under imaging guidance, depending upon the region. For superficially located lesions such as lymph nodes in the head and neck region, US imaging is the modality of choice (Fig. 20). However, for deep-seated lesions such as mediastinal lymphadenopathy or retroperitoneal lymphadenopathy, CT guidance is necessary. CT is usually also needed when a biopsy is required in the musculoskeletal system.

Fig. 20

High-resolution US image shows an enlarged cervical lymph node with biopsy needle in situ

2.8.1.1 General Principles

Informed consent should be obtained from the patient. The procedure can be performed under local anesthesia, with or without moderate sedation. Care should be taken with the dosage of these medications in elderly patients. The laboratory test values, especially full blood counts and clotting profile, should be carefully noted, and they should be within the acceptable range. Relative contraindications include significant coagulopathy, low platelet count (less than 50,000/mm³), severely compromised cardiopulmonary function, inaccessible sites or lack of safe pathway, and patient’s inability to cooperate during the procedure.

2.8.1.2 Imaging Guidance

The biopsy can be performed by US imaging guidance in superficial lesions such as enlarged lymph nodes in the neck (Fig. 20). US imaging is preferred for superficial lesions because of multiple advantages such as real-time imaging, low cost, absence of ionizing radiation hazard, and dynamic imaging capabilities. Color Doppler US imaging aids in assessing the vascularity of the lesion and location of vessels along the needle trajectory or that lie close to the lesion.

CT guidance is preferred for deep-seated lesions such as enlarged lymph nodes in the mediastinum or abdomen or focal lesions in the lungs. Fluoroscopy or CT can be used for biopsy in the musculoskeletal system, such as spine lesions. However, CT is the imaging modality of choice as it enables precise localization and is safer compared to fluoroscopy, especially to avoid vital structures such as nerves and blood vessels. The technique of CT fluoroscopy uses less radiation dose and tracks the needle almost close to real time. Up to six frames per second can be reconstructed while performing CT fluoroscopy. A viewing monitor is usually set up within the CT suite. The CT couch can be moved using the console or operated manually, and image acquisition is usually done with a foot switch. The needle tip can be tracked by looking for low-attenuation beam hardening artifacts. Most often, the needle is inserted perpendicularly in the axial plane. Rarely, the angulated approach may be necessary to reach the target.

2.8.1.3 Tools and Techniques for Percutaneous Biopsy

Fine-needle aspiration samples are often too small, and usually core biopsy is preferred. For FNAC, needles of sizes 20–22G are preferred. Core needle biopsies are often performed with needles of sizes 16–20G, with larger caliber outer coaxial needles. The length of the “throw” or distance advanced by the needle after firing is indicated on the needle package and should be selected carefully. For example, if the lesion is 1.5 cm in diameter, a 1 cm throw needle can be safely used. The coaxial needle is usually introduced first, and the spring-loaded cutting needle is then advanced through the coaxial needle. This avoids multiple punctures. If there is excessive bleeding from the outer coaxial needle after biopsy, the track can be sealed with a Gelfoam plug. If pneumothorax occurs during a lung biopsy, a wire can be inserted through the coaxial needle and a drainage catheter inserted over the wire.

For bone lesions with shell of overlying bone, the cortex is penetrated with a trephine needle (14.5 or 15G), and then a cutting needle can be inserted using a coaxial technique to obtain the tissue sample. If the lesion is completely sclerotic or is an osteolytic lesion with a predominant bone component, then the lesion is sampled with a trephine needle. The cortex is penetrated by a corkscrew rotational motion of the trephine needle with a diamond-tip stylet. After penetration of the cortex, when the margin of the bone lesion is reached, the stylet is removed and a syringe is attached. The plunger is withdrawn to create a vacuum. A jiggling motion is applied to dislodge the tissue when the vacuum is being applied. The plunger is then released, and the needle is removed. The tissue sample, which is usually present within the needle, is removed with the help of a blunt stylet. Intermittent CT fluoroscopic screening should be performed when the needle is advanced. Multiple passes should be obtained with needle directed in different directions to obtain samples in different parts of the lesion.

2.8.2 Drainage and Aspiration of Collections

Drainage of collections can be performed under US imaging or CT guidance (Yin et al. 2015). When US imaging guidance is used, fluoroscopy can be utilized to position the catheter in the desired location. For large uncomplicated pleural collections, direct trocar puncture technique can be used under US imaging guidance. However, for deep-seated or complex collections, drainage catheters can be inserted by the Seldinger technique. Initially, the collection is accessed under US imaging or CT guidance by a 18G or 21G needle, and the track is dilated to accommodate an 8Fr or 10Fr pigtail drain. For deep-seated abdominal, paraspinal, or mediastinal collections, CT guidance is preferred (Fig. 21).

Fig. 21

Tuberculous spondylodiscitis of the thoracic spine. (a) Axial contrast-enhanced CT image shows a large paraspinal abscess (arrowheads). (b) Axial unenhanced CT image obtained postdrainage of the abscess shows a drainage catheter in the abscess cavity. Frontal radiographs obtained (c) before and (d) after drainage of the paraspinal collection (black arrow)

2.8.3 Embolization for Hemoptysis

Embolization is the initial treatment of choice in patients with pulmonary TB presenting with massive hemoptysis (approximately 200–1000 ml over a 24-h period). Massive hemoptysis results from a hypertrophied bronchial artery or nonbronchial systemic collaterals supplying the diseased lung (Fig. 22). Rarely, aneurysms from the pulmonary artery (Rasmussen aneurysm) can be the cause of bleeding (Seedat and Seedat 2018) (Fig. 23). Contrast-enhanced CT obtained in the arterial phase is helpful to identify the source of hemoptysis. CT helps assess the degree of lung damage and shows the anatomy and course of hypertrophied bronchial arteries. If CT is not performed, an aortic angiogram using a pigtail catheter may be performed to identify the bronchial arteries. Endovascular stenting with stent-grafts together with anti-TB drug therapy has been reported to be an alternative to traditional open surgery in patients with mycotic aneurysms of major arteries caused by TB (Zhao et al. 2019). Deployment of such stent-grafts is usually done in the angiography suite under fluoroscopic guidance.

Fig. 22

Post-primary pulmonary tuberculosis involving the upper lobe of the right lung in a 68-year-old woman who presented with massive hemoptysis. (a) Selective bronchial angiogram shows a hypertrophied and tortuous right bronchial artery supplying the right lung (long arrow). (b) Postembolization angiogram of the right bronchial artery (short arrow) shows absent flow distally. Patient did not have further episodes of hemoptysis after embolization

Fig. 23

Rasmussen aneurysm in a 34-year-old man with post-primary pulmonary tuberculosis who presented with massive hemoptysis. (a) Coronal contrast-enhanced CT and (b) selective right pulmonary angiogram images show a small Rasmussen aneurysm arising from the distal right lower lobe pulmonary artery, along the margin of a cavitary lesion (arrow). Corresponding (c) coronal contrast-enhanced CT and (d) selective right pulmonary angiogram images show successful embolization (arrow). Patient had resolution of hemoptysis after coil embolization

3 Conclusion

The role of imaging in the diagnosis of TB complements the clinical and laboratory findings. Its vast role also extends to assessment of the extent of disease and its complications, monitoring the disease response to treatment and detecting residual disease at the end of therapy, as well as guiding biopsies and therapeutic drainage. Conventional radiographs are important as an initial tool for screening, while CT and/or MRI are necessary for further cross-sectional assessment. Each of these modalities has its own advantages, disadvantages, and limitations; thus, the modality of choice varies based on clinical indications and should be catered to each individual patient. In the absence of known contraindications, CT remains the modality of choice in detailed evaluation of pulmonary, abdominal, urinary tract, and head and neck TB, while MRI is the preferred choice in the assessment of intracranial, spinal, and musculoskeletal TB. Other imaging modalities, such as US imaging and ¹⁸F-FDG PET/CT, may be selectively employed in the management of patients with TB.