Reflections and Acoustic Impedance

Sound waves and their reflections form the basis of ultrasound images. A transducer sends out pulses of high-frequency sound waves and receives their echoes. Reflection of sound waves can occur when the ultrasound beam encounters a different tissue. Two factors influence the amount of reflection: the acoustic impedance of the two tissues and the angle between the direction of the sound wave and the reflecting surface (Fig. 3.1).^8–9

Fig. 3.1 Factors influencing the amount of reflection.

The acoustic impedance is made up by the combination of density of the tissue and the sound velocity within the tissue. The latter is different for each tissue type (Table 3.1). The largest differences in acoustic impedance are found between bone and air, in which sound has a velocity of approximately 300 and 4000 m/s, respectively, whereas in muscle the sound velocity is approximately 1580 m/s.⁹ The larger the difference in acoustic impedance, the larger the amount of reflection of the ultrasound beam. A small difference in acoustic impedance will result in no reflection or reflection of only a small part of the sound wave, whereas most of the ultrasound beam can pass through to deeper layers. This is also the case in muscle: histologic structures in healthy muscle will mostly transmit sound (Fig. 3.2). Conversely, a transition with a very large difference in acoustic impedance will result in reflection of all sound waves. Beneath such a transition no imaging is possible. For example, the difference in acoustic impedance of air and skin is very large, which makes it impossible to produce ultrasound images without the help of a gel or another contact fluid between the transducer and the skin (Fig. 3.3). A transition from muscle to bone or calcifications will also cause a strong reflection. As the amplitude of the sound wave corresponds to the brightness of the image, the strong reflections of bone result in a bright spot on the ultrasound image. Additionally, because hardly any sound gets through, no structures beneath this transition can be displayed, resulting in a characteristic bone “shadow” (see Fig. 3.2).

Table 3.1 Sound Velocity in Various Tissues

Fig. 3.2 The interaction between the ultrasound beam and normal and pathologic tissues (left panel) with examples of an ultrasound image (right panel). When the ultrasound beam encounters tissue with a different acoustic impedance such as the transition from muscle to fascia, a part of the sound is reflected (upper left). Because healthy muscle contains only little fibrous tissue, only a few reflections occur, resulting in a relatively black picture (upper right). The acoustic impedance is very different between muscle and bone, causing a strong reflection, with hardly any sound passing through. This results in a bright bone echo with a characteristic bone shadow (middle pictures). In neuromuscular disorders such as Duchenne’s muscular dystrophy, muscle tissue is replaced by fat and fibrous tissue, resulting in many transitions with different acoustic impedance and much reflection of the ultrasound beam. This explains why the muscle ultrasound image appears white. Additionally, attenuation of the ultrasound beam occurs, causing the superficial part of the muscle to be whiter than the deeper part and invisibility of the humerus. B, Biceps brachii muscle; Br, brachialis muscle; F, fibula; H, humerus; P, peroneus longus muscle.

Fig. 3.3 Ultrasound image of the quadriceps muscle containing artifacts caused by air between the skin and the transducer. This results in a black line beneath the skin where the air is located (right arrow). When the air bubble is small, the changes are subtle, with only a slightly decreased echogenicity in the structures below, which can be seen beneath the left arrow F, Femur; R, rectus femoris muscle; V, vastus intermedius muscle.

The amount of sound reflection is also determined by the angle at which the ultrasound beam reaches a tissue transition.^8–9 A large angle will result in deflection (away from the transducer) instead of reflection of the sound. In this way, the sound wave will not return to the transducer and the structure will not be displayed (see Fig. 3.1). Because of this, fascia surrounding muscle, nerves, and bone will be especially echogenic when they are reached perpendicular by the ultrasound beam. Oblique scanning will cause deflection and diminishes the echogenicity of these structures (Fig. 3.4). Moreover, fascia between muscles that are in line with the ultrasound beam is difficult to visualize; the fascia remains either black or is displayed as an interrupted line (Fig. 3.5).

Fig. 3.4 Ultrasound image of the upper arm perpendicular (A) and after angulating the transducer 5 degrees (B). After angulation of the transducer, the sound reaching the fascia within and surrounding the muscle as well as the bone gets partially deflected. Because this sound does not return to the transducer, the image is relatively black compared with the image with a perpendicular position of the transducer. The double arrow indicates subcutaneous tissue.

Bi, Biceps brachii; Br, brachialis muscle; H, humerus.

Fig. 3.5 Ultrasound image of the peroneus longus (PL) muscle. Fascia surrounding or within the muscle that are perpendicular to the ultrasound beam appear as bright white lines, as does the bone (black lines). Conversely, fascia oblique to the ultrasound beam, such as the fascia between the peroneus longus muscle and the extensor digitorum longus (EDL) is less bright (gray line). The fascia between the peroneus longus and lateral gastrocnemius muscle (GL) is in line with the ultrasound beam and therefore not visible at all (dotted line).

When the ultrasound beam encounters structures smaller than the wavelength of the sound emitted, such as individual muscle fibers, the sound waves are not reflected but scattered (see Fig. 3.1).⁹ Only a small part of this sound will return to the transducer, and will therefore result in only a small increase of the gray level in the picture.

In summary, the gray value of a certain tissue is determined by the amount and amplitude of the reflection. This gray value is called echogenicity in this chapter.

Choice of Transducer Frequency

When the sound beam travels to deeper layers, attenuation of the ultrasound beam occurs. Less sound reaches deeper levels, caused by reflection and scattering, but also by absorption and transformation of the beam into thermal energy. The amount of absorption depends on the type of tissue and is, for example, very high in bone. It increases with higher transducer frequencies (Fig. 3.6). For average soft tissues, the loss amounts to approximately 1 dB per 10-mm tissue depth for each MHz. As a rule, the depth penetration is limited to about 200 mm for a 3-MHz probe, 100 mm for a 6-MHz probe, and 12 mm for 50-MHz probe.⁶

Fig. 3.6 Ultrasound images of the calf muscles made with two different transducers. A and B were made with a 17-5 MHz broadband linear transducer resulting in an image with relatively high resolution but also with strong attenuation of the ultrasound beam. The fibula is hardly visible because much of the sound is absorbed by the overlying soleus (S), gastrocnemius (G), and the subcutaneous tissue (double arrow). Increasing the time gain compensation in B amplifies the signal coming from deeper layers, making it possible to visualize deeper structures. But these deeper structures are still difficult to discriminate from surrounding tissue because the background noise is also amplified.

C and D were made with an 8-4 MHz broadband linear transducer. Because of the lower frequencies, less attenuation of the ultrasound beam occurred, resulting in a better display of deeper structures but with a lower resolution. Increasing time gain compensation has also amplified signals from deeper structures in D, resulting in a bright line of the fibula.

Axial resolution is determined by the transducer frequency. As higher frequencies are accompanied by smaller wavelengths, a transducer with the highest frequency will achieve the best axial resolution. The lateral resolution is best in the focal zone of the ultrasound beam and is several times larger than the axial resolution.⁶ This means that when the focus of the ultrasound beam is positioned at a depth that contains the structures of interest (i.e., the muscles under investigation) these structures will be visualized with the best lateral resolution.

In clinical practice, the transducer that can reach the depth of the tissue of examination with the highest frequency is preferable. In muscle ultrasound studies a 7.5-5 MHz probe is generally used. Now it is also possible to use broadband transducers (for example, a 17-5 MHz probe), combining the advantage of high frequencies to image superficial structures with high resolution (up to 0.1 mm) and lower frequencies to reach deeper structures (resolution decreases to 0.3 mm).

Histology

To understand what is visualized with muscle ultrasound, it is also necessary to know the architecture of muscles up to the maximum resolution of muscle ultrasound. Muscles are composed of individual muscle fibers that normally have a diameter of approximately 40 to 80 μm.¹⁰ The resolution of muscle ultrasound is around 100 μm, which means that either one large fiber or a group of small muscle fibers are the smallest structures within muscle that can be visualized. Each muscle fiber is surrounded by thin fibrous tissue called endomysium. A group of muscle fibers forms a muscle fascicle, surrounded by perimysium. This connective tissue also contains small blood vessels and nerves. The outside covering of the muscle is made of fibrous tissue called epimysium, which is thicker and stronger than perimysium (Fig. 3.7). The epimysium continues at the origin and insertion to form tendons and sheetlike tendons called aponeuroses.¹¹

Fig. 3.7 Normal muscle architecture.

(Adapted from The structure of muscle and associated connective tissues. lvyRose Holistic: Holistic Health, Alternative Medicine, Human Biology, and Anatomy & Physiology. Available at www.ivy-rose.co/uk/humanbody/muscles/muscle_structure.php. Accessed June 2008.)

The organization of muscle fascicles results in the macroarchitecture of a muscle. In some muscles, the fibers run parallel to the long axis of the muscle and a few of them run through its whole length (Fig. 3.8A). In other muscles, the fibers are obliquely oriented to the long axis of the muscle. Because of their resemblance to feathers these are called pennate muscles (see Fig. 3.8B). They can be divided in unipennate muscles when the fibers have one linear origin resembling one half of a feather; bipennate muscles when the fibers arise from a broad surface resembling a whole feather; multipennate muscles when septa extend into the attachment of the muscles, dividing them into several feather-like portions; and circumpennate muscles when the fibers converge on a tendon extending into their substance (like the feathers of a dart). In other muscles, the fibers converge from a wide attachment to a fibrous apex (see Fig. 3.8C). This type of muscle is described as a fan-shaped or triangular muscle.¹²

Fig. 3.8 The macroarchitecture of a muscle is formed by the way muscle fascicles are organized. This can be parallel such as the biceps brachii muscle (A), pennate such as the tibialis anterior muscle (B), or triangular such as the latissimus dorsi muscle (C). This structure becomes visible on ultrasound when measuring longitudinally.

(Part of the anatomical drawings were modified and used with permission from PreventDisease.com: Muscle atlas. Available at www.preventdisease.com/home/muscleatlas.shtml. Accessed June 2008.)

Apart from a structural or histologic organization, muscles also have a functional or physiologic organization. This functional unit is called a motor unit, consisting of one motor neuron and the muscle fibers it controls (Fig. 3.9). The number of muscle fibers in a motor unit varies from less than 10 to several hundred, according to the size and function of the muscle. Large motor units, where one neuron supplies several hundred muscle fibers, are found in the trunk and thigh muscles, whereas motor units of small eye and hand muscles, where precision movements are required, contain only a few muscle fibers. The muscle fibers of a motor unit are not grouped together but spread among the muscle fascicles.¹¹ Therefore, the diameter of a motor unit can range from less than one millimeter to more than one centimeter wide. Because muscle fibers have no anatomic features that attribute them to one particular motor unit, it is impossible to tell with an imaging technique which muscle fibers belong to which motor unit. Hypothetically, if pathology was confined to one motor unit or a single motor unit was activated and would contract, imaging techniques might be able to visualize this particular motor unit provided that the resolution is sufficient.

Fig. 3.9 The functional unit of a muscle is a motor unit, consisting of the motor neuron and the muscle fibers it controls.

Ultrasound of Normal Muscle Tissue

The ultrasonographic appearance of muscles is fairly distinct and can easily be discriminated from surrounding structures such as subcutaneous fat, bone, nerves, and blood vessels (Fig. 3.10).^13–14 Normal muscle tissue is relatively black, that is to say, it has a low echogenicity. In the transverse plane, which is perpendicular to the long axis of the muscle, muscle tissue has a speckled appearance because of reflections of perimysial connective tissue that is moderately echogenic (see Fig. 3.10). In the longitudinal plane (parallel to the long axis of the muscle) the fascicular architecture of muscle tissue becomes visible. Reflections of the perimysial connective tissue result in a linear, pennate, or triangular structure on the ultrasound image (see Fig. 3.9). The boundaries of the muscle are clearly visible because the epimysium surrounding the muscle is a highly reflective structure. In normal subjects, the bone echo beneath the muscle is strong and distinct with an anechoic bone shadow underneath (see Fig. 3.2B). Subcutaneous fat has a low echogenicity, but several echogenic septa of connective tissue may be visible within this tissue layer (Fig. 3.11). Nerves and tendons are relatively hyperechoic compared with healthy muscles, whereas blood vessels are hypoechoic or anechoic circles or lines depending on the direction of the ultrasound beam (Fig. 3.12A). When one is unsure about the nature of a round or linear structure, Doppler imaging can confirm the presence of arteries or veins, by showing blood flow (Fig. 3.12B).

Fig. 3.10 A, Normal ultrasound measurement of the biceps brachii muscle and surrounding tissues, measured at two thirds of the distance from the acromion to the antecubital crease of the left arm. B, Depicts the different structures schematically.

Fig. 3.11 A, Ultrasound image of the calf showing several layers of fascia in the subcutaneous tissue. B, Depicts the different structures schematically.

Fig. 3.12 Transverse (A), and longitudinal (B and C) ultrasound images of the proximal forearm. The median nerve is visible as a hyperechoic circle in the transverse plane with a honeycomb appearance. In the longitudinal plane the fascicle structure of the nerve becomes visible as hyperechoic lines between the relatively hypoechoic muscles (flexor carpi radialis [FC] above and flexor digitorum [FD] profundus below). The ulnar artery is visible as an anechoic line in the longitudinal plane. In the transverse image Doppler imaging is applied, which confirms that this structure indeed is an artery. The double arrow indicates subcutaneous tissue. IM, Interosseous membrane; R, radius; U,ulna.

All superficial muscles can easily be imaged with ultrasound, although it can be difficult to identify individual small muscles when multiple muscle groups overlap. Recent ultrasound technology using higher transducer frequencies with corresponding higher axial resolutions has made it possible to image individual superficial small muscles in, for example, the hand (Fig. 3.13). Deeper muscles, especially in the pelvic region or around the trunk, can be difficult to visualize with sufficient resolution because of the reflection or absorption of sound by superficial tissue layers, such as skin, subcutaneous tissue, or other muscles.

Fig. 3.13 Ultrasound image of small hand muscles. The resolution of the 4-8 MHz transducer (A) is not sufficient to visualize individual hand muscles, whereas with the 5-17 MHz transducer (B) the abductor pollicis brevis (Ab), flexor pollicis brevis (F), and adductor pollicis (Ad) can be distinguished. The tendon of the flexor pollicis longus is visible as a bright spot (arrow). MC1, First metacarpal bone; MC2, second metacarpal bone.

Muscle Thickness

Muscle ultrasound is a reliable method to measure muscle thickness and cross-sectional areas,^15–17 with a test-retest correlation of 0.98 to 0.99^15,18 and a 0.99 correlation MRI.¹⁵ Several studies have used ultrasound to establish muscle thickness in healthy subjects. During childhood, muscle thickness increases rapidly. The main variable to predict muscle thickness in this age group is weight.¹⁹ Gender differences do not influence muscle thickness until puberty, when men start to develop bigger muscles than women.^20–22 After puberty muscle thickness increases further in both genders, until a peak is reached between 25 and 50 years of age. Thereafter, muscle thickness declines.^21–24 The influence of age and gender is different for each muscle group and should be taken into consideration when evaluating muscle ultrasound scans of individual patients.

Muscle Echogenicity

Muscle structure can also be evaluated with ultrasound by evaluating muscle echogenicity. Normal muscles are relatively black, but different muscles have specific appearances on ultrasound because of the variability in proportion of fibrous tissue and the orientation of muscle fibers. For example, the anterior tibial muscle has generally a whiter appearance than the rectus femoris.¹⁹

Muscle echogenicity increases with age, which is caused by age-related replacement of muscle fibers by fat and fibrous tissue.^25,26 Figure 3.14 shows an example of this age-related increase in muscle echogenicity of the biceps brachii muscle. Fat and fibrous tissue have a different acoustic impedance, increasing the number of reflecting interfaces in the muscle, which gives the muscle a whiter appearance (see Fig. 3.2B). Essentially, the same mechanism also causes increased muscle echogenicity in neuromuscular disorders.^1,27 The visibility of structures in and surrounding the muscle can provide additional clues about the presence of structural muscle changes. In several neuromuscular disorders, such as muscular dystrophies and spinal muscular atrophy, this finding can be as prominent as the increase in muscle echogenicity.^28,29

Fig. 3.14 Age-related changes in muscle echogenicity of the biceps brachii muscle, based on measurements in 194 healthy volunteers. The line represents the best nonlinear fit. In childhood, boys and girls have approximately the same muscle echogenicity. After the age of 16, men have a lower echogenicity than do women. Muscle echogenicity increases after the age of 40 in men and women.

(Adapted from Pillen S, Arts IMP, Zwarts MJ: Muscle ultrasound in neuromuscular disorders, Muscle Nerve 37:679-693, 2008.)

Visual detection of slightly increased muscle echogenicity can be difficult and its accurate interpretation depends on the experience of the observer, especially because different muscles have different appearances, and muscle echogenicity increases with age. Changes in ultrasound system settings, such as increased gain, can also give muscles a whiter appearance that can be mistaken for pathologically increased echogenicity. For these reasons, visual evaluation of muscle ultrasound has shown a relatively low interobserver agreement (kappa 0.53) that further deteriorated when an inexperienced observer interpreted the images.³⁰ To overcome this problem, computer-aided techniques were introduced to help image interpretation. Quantification of muscle echogenicity can be achieved with simple gray scale analysis (Fig. 3.15).^23,30–32 Quantitative echogenicity analysis is a robust clinical technique with a high interobserver agreement (kappa 0.86).³⁰ It is also suitable for research purposes.

Fig. 3.15 Transverse ultrasound image of a normal left quadriceps muscle (A) and of a patient with Duchenne’s muscular dystrophy (B). Both are 3.5 years of age. The rectus femoris muscle is encircled in both images. In this region the mean echogenicity is measured, as shown in the histograms (inserts) (scale: black = 0; white = 256). The rectus femoris of the Duchenne’s muscular dystrophy patient has an increased muscle echogenicity, with the corresponding histogram displaced to the right. Note the fine granular pattern of echogenicity, homogeneously spread among the muscle with attenuation of the ultrasound beam. Fascia within the muscle, for example, the central fascia in the anterior part of the rectus femoris (single arrow), is more difficult to recognize in the Duchenne’s muscular dystrophy patient. Also, note the reduced visualization of the bone edge of the femur in the Duchenne’s muscular dystrophy patient. The double arrow indicates subcutaneous tissue. The quadriceps muscle was measured halfway along the line from the anterior superior iliac spine to the patella. F, Femur; VI, vastus intermedius; VL, vastus lateralis; VM, vastus medialis.

An often mentioned objection to quantitative gray scale analysis is that it is more time consuming. However, with current special software, quantification of echogenicity can be done quickly, and evaluation of a standard screening protocol (comprising four muscles) takes less than 5 minutes.

Next to calculation of mean echogenicity, other quantitative techniques such as backscatter analysis have been introduced in muscle ultrasound because this technology is less influenced by system settings.³³ This technique compares measured echo intensity with phantom measurements, to make it easier to compare echogenicity measurements made using different ultrasound devices. However, some hardware differences, especially differences in transducers, can still be a problem when using normal values established in other centers. Moreover, this technique has not yet been validated in a large prospective study.

Even with quantitative echogenicity analysis, visual evaluation of the images still has its value. Focal changes in a muscle that has a normal overall muscle echogenicity can easily be detected visually. Visual evaluation also provides information on the distribution of echogenicity within a muscle, that is, whether it is homogeneous or inhomogeneous. New developments in digital image analysis, such as texture analysis, are being designed to quantify these alterations within muscle tissue.^25,34 These techniques are already used for detecting focal changes in other tissues, such as the breast and prostate.^35–36

Interpretation of Muscle Thickness

Muscle atrophy is often found in neuromuscular disorders. However, patients with symptoms such as hypotonia and delay in motor development without a neuromuscular disorder often show muscle atrophy, which is explained by disuse. Exercise and body composition also influence muscle thickness and should be taken into account when interpreting the results. For example, atrophy of the quadriceps muscle in wheelchair-bound patients is not surprising and will therefore not help in determining whether a neuromuscular disorder is present. True hypertrophy can also be found in both healthy subjects (as a training effect) and patients with a neuromuscular disorder, such as myotonia congenita or Brody’s disease. Increased muscle thickness also can be caused by pseudohypertrophy, such as that found in the calves of Duchenne’s muscular dystrophy patients. Muscle thickness will therefore have to be interpreted in combination with the results of quantitative echogenicity measurements. If atrophy is present, an increased muscle echogenicity indicates that the atrophy is probably caused by a neuromuscular disorder, whereas a normal echogenicity is suggestive of disuse atrophy (Table 3.3).

Table 3.3 Interpretation of Muscle Thickness in Correlation with Muscle Echogenicity Findings

Transducer and System Settings

The appropriate transducer to perform muscle ultrasound in children and superficial muscles in adults is a 7.5-MHz probe, and a 5-MHz probe for deeper muscles of adults. A single broadband transducer with a range of at least 12-7 MHz also can be used. It is important to select the optimal settings for the visualization of skeletal muscles. Gain and compression have to be chosen so they will result in a picture in which a normal muscle is relatively black (see Fig. 3.10). This ensures that diseased muscles will appear whiter on the ultrasound screen. Conversely, when the gain is already set too high at the start of the examination, normal muscle tissue will already appear white. If a diseased muscle is then displayed, this muscle will also be white on the screen. Therefore, less contrast between normal and diseased muscle will be available for their discrimination (Fig. 3.17), resulting in a loss of sensitivity and specificity. Focus has to be adjusted to a depth where the muscle of interest is expected. Normally, two or three focal points at 1 to 3 cm deep will be sufficient. It is important to save these settings in a default, often called “preset,” making sure that every measurement will be performed with the same system settings. Such a preset is necessary when one wants to compare measurements between different patients or subjects and makes it possible to familiarize oneself with the normal appearance of muscle and the difference with diseased muscles. If muscle echogenicity is quantified it is mandatory to keep all system settings constant, otherwise the measurements are not comparable. Another adjustment, the so-called time gain compensation, is automatically applied to the ultrasound image by the ultrasound system but can also be further adjusted manually by an operator to obtain the desired image quality on the screen. This adjustment is forbidden when measuring muscle echogenicity quantitatively. When performing quantitative muscle ultrasound measurements one has to be sure that the bars of the time gain compensation are in the same (neutral) position, to be able to compare muscle echo intensities between measurements.

Fig. 3.17 A, Ultrasound image of the quadriceps femoris muscle in a patient with spinal muscular atrophy (SMA) showing increased muscle echogenicity in the rectus femoris muscle. B, Ultrasound image showing a healthy muscle, which appears as white as the muscle of the SMA patient. Because of adjustments of the system settings, the overall gain and the time gain compensation (TGC) of the depth corresponding to the rectus femoris muscle have been increased. This figure emphasizes that system settings have to be checked before interpreting the ultrasound image. Preferably, system settings should be kept constant in a default or so-called preset, to avoid this problem. An important difference between the healthy muscle and muscle of the patient with SMA is the visibility of structures within and surrounding the difference. Because these structures are clearly visible in the healthy subject, one could have already suspected that this abnormal echogenicity is caused by system settings instead of a neuromuscular disorder. Subcutaneous layer is indicated by double arrows F, Femur; R, rectus femoris; V, vastus itermedius muscle.

Measurement Errors

The patient has to be placed in a relaxed position to avoid muscle contraction, because contraction of a muscle results in an increased muscle diameter and a decreased echogenicity (Fig. 3.18).⁴⁸ To compare muscle thickness or muscle echogenicity between subjects and over time, all measurements have to be performed at the same anatomically defined locations with the subject in a standardized position. For example, bending the knee will cause a change of the direction of the muscle fibers in the quadriceps muscle, resulting in an increased muscle echogenicity compared with measuring echogenicity with a straight leg.^1,42

Fig. 3.18. A, Ultrasound image of the upper arm showing the biceps brachii muscle (Bi, double arrow) and brachialis muscle (Br). B, Compression decreases the muscle thickness, especially of the most superficial layers, i.e., the biceps brachii muscle (double arrow). Compression can be avoided by paying attention to the outline of the skin, which has to be slightly curved as shown in AC, Contraction of the biceps brachii muscle increases its thickness (double arrow). H, Humerus.

A generous amount of contact gel between the skin and transducer should be used to ensure optimal acoustic coupling and to prevent pressure on the underlying tissues (see Fig. 3.18). Occasionally, a water bath (i.e., immersion of the limb) is necessary when imaging superficial or sharply contoured structures such as muscles of the hand and foot (Fig. 3.19). Especially when measuring in a transverse plane, it is important to place the transducer perpendicular to the underlying muscle tissue, because oblique scanning will not only overestimate muscle thickness but also result in a decrease of muscle echogenicity (see Fig. 3.4).⁴⁸ Placing the transducer in a way that results in an optimal bone echo and delineation of the surrounding fascia will further standardize the measurement, thereby avoiding differences caused by angulation of the transducer.

Fig. 3.19 Ultrasound image of the extensor digitorum brevis (EDB) muscle measured submerged in a water bath, which avoids air artifacts and compression of the tissue, such as the vein in the subcutaneous tissue. Cub, Cuboid; LC, lateral cuneiform bone.

All measurement errors and ways to avoid them are summarized in Table 3.4.

Table 3.4 Checklist for Evaluating the Quality of Muscle Ultrasound Measurement

Measuring Muscle Thickness

Muscle thickness is usually determined during the measurement, with a caliper function that is generally incorporated in the ultrasound device’s software. In most circumstances, assessment of muscle thickness will be sufficient to determine whether muscle atrophy or hypertrophy is present, but it is also possible to measure muscle cross-sectional area. This type of measurement has become a standard feature in newer ultrasound devices. Measurement of muscle cross-sectional area can have additional value in, for example, the estimation of total muscle mass,¹⁷ or to relate the force-generating capacity of muscle to its cross-sectional area.⁴⁹ The assessment of the cross-sectional area can be done easily in small muscles, when the whole muscle width is visible on the screen. However, larger muscles such as the rectus femoris or biceps brachii muscle often do not fit the size of the transducers (normally 4 to 5 cm wide) and can therefore not be visualized in one image except when they are severely atrophic, or in neonates. Additional image processing such as the panoramic view is then necessary to visualize the entire muscle (Fig. 3.20).⁵⁰

Fig. 3.20 A, Because the calf muscles are too wide to be imaged in one picture, additional methods have to be applied to visualize the entire medial and lateral gastrocnemius muscles. B, An example of a measurement applying a panoramic view, which is incorporated in some newer ultrasound devices. This makes it possible to visualize the entire muscle in one image. However, care has to be taken to measure distances and intensities in these pictures because they can be subject to artifacts. GL, Lateral gastrocnemius muscle; GM, medial gastrocnemius muscle; PL, peroneus longus muscle.

It is advisable to measure muscle thickness from the skin to a highly reflective structure beneath the muscle such as an underlying bone, because other structures such as the surrounding fascia can be difficult to recognize in severe neuromuscular disorders. This might imply that to measure muscle thickness it is necessary to include more than one muscle in the measurement, for example, the rectus femoris and vastus intermedius, when measuring the quadriceps. The underlying bone remains detectable in most circumstances. In some conditions such as severe neuromuscular disorders or obesity, it might be necessary to adjust the system settings only for the purpose of thickness measurement to visualize the bone (for example, by increasing time gain compensation, adjusting the focus, or using a transducer with a lower frequency) (Fig. 3.21).

Fig. 3.21 Ultrasound image of the quadriceps muscles of a 10-year-old patient with Duchenne’s muscular dystrophy. A, In severe neuromuscular disorders visualization of structures beneath muscle tissue can be difficult because of attenuation of the ultrasound beam. B, Measuring with a lower frequency and adjusting the time gain compensation and the focal zone enable visualization of the femur, which is necessary to measure muscle thickness.

Measuring Muscle Echogenicity Using Gray Scale Analysis

When quantifying echogenicity it is important to keep all settings that influence the gray value—such as the gain, compression, focus, and time gain compensation—constant throughout the measurements. Slight changes of the probe position already induce some variation in muscle echogenicity. To obtain the most reliable echogenicity it is preferable to make at least three images of each muscle and average the results of all three measurements. This reduces measurement variation. Between each measurement the subject is allowed to move and the transducer has to be repositioned for each new image. The images are then stored for offline calculation of mean muscle echogenicity.

Echogenicity can be determined with the use of computer software containing a histogram function. For this it is necessary to select a region of interest of which the mean gray value is to be calculated. Dedicated software has been developed that automatically imports images and allows easy region of interest selection and echogenicity measurements. It also contains features that automatically compare the measured gray values with age- and muscle-specific normal values. But when this software is unavailable any image processing program, such as Adobe photoshop, can do the job.

The region of interest preferably consists of a large area within the muscle of interest to avoid sampling errors. In the case of a patchy ultrasound appearance such as spinal muscular atrophy (Fig. 3.22), sampling error can occur when a relatively black part of the muscle is selected. The surrounding fascia should be avoided and not be incorporated in the region of interest because this can falsely elevate muscle echogenicity.

Fig. 3.22 A, Quadriceps muscle of a 2-year-old boy with spinal muscular atrophy type 2, measured halfway along the line from the anterior superior iliac spine to the patella. The echogenicity is very inhomogeneously spread among the muscle, with a moth-eaten pattern, probably representing atrophic and hypertrophic muscle fibers. Normal structures within and surrounding the muscle are difficult to distinguish. Muscle thickness is diminished. B, Depicts the different structures schematically. RF, rectus femoris muscle.

(Adapted from Pillen S, Arts IMP, Zwarts MJ: Muscle ultrasound in neuromuscular disorders, Muscle Nerve 37:679-693, 2008.)

From the region of interest a histogram can be constructed from which the mean gray value can be calculated. Because most ultrasound devices produce images with an 8-bit gray scale, the gray scale analysis results in a value between 0 (black) and 255 (white).

Unfortunately, the muscle echogenicity value is highly dependent on the type of ultrasound device and the system settings used. It is therefore not possible to use normal values established on one device on another device without the help of a correction. Fortunately the mean muscle echogenicity is a solid variable, probably because it is calculated from a rather large area and mainly influenced by the macrotexture of an image that remains approximately the same with different ultrasound devices. This makes it possible to achieve a reliable conversion of normal values after using a conversion equation.⁵¹ Such a conversion is made by measuring the same tissue or phantom on the two different ultrasound devices and comparing the results. For converting echogenicity values, one should refrain from additional postprocessing such as compound imaging or other resolution enhancement settings to minimize the difference between ultrasound devices. It is also advisable to perform quality checks of the ultrasound device on occasion to ensure that previously established normal values can still reliably be used.

Visual Assessment of Muscle Ultrasound Images

Muscle ultrasound images can contain additional clues to a specific type of neuromuscular disorder. In Chapters 8 through 10, these features are described for motor neuron disease and specific myopathies. These features cannot be quantified (yet) and can therefore currently only be assessed by visual inspection of the ultrasound images.

For visual assessment, first the distribution of echogenicities within the muscle has to be evaluated: is it homogeneous or inhomogeneous? If inhomogeneous, does the image contain focal areas of increased muscle echogenicity that become especially apparent with angulation of the transducer, pointing to an inflammatory myopathy? Or is it a moth-eaten appearance containing very black areas that can be found in spinal muscular atrophy (see Fig. 3.22)? Table 3.5 summarizes several patterns of echogenicity distribution and the corresponding possible diagnoses.

Table 3.5 Clues for Specific Neuromuscular Disorders

The fascia surrounding the muscle should also be evaluated: is it thin and clearly demarcated (i.e., normal) or thickened with a fuzzy appearance suggestive of a fasciitis (Fig. 3.23)? One should further screen the entire muscle for the presence of subcutaneous and intramuscular calcifications, especially when dermatomyositis is suspected.

Fig. 3.23 This image shows a transverse measurement of a quadriceps muscle of a patient with eosinophilic fasciitis. The fascia surrounding the rectus femoris is clearly thickened (∗). The central fascia separating the anterior part of the rectus femoris muscle is also thickened and less well demarcated (arrow). The measurement was made halfway along the line from the anterior superior iliac spine to the patella. The double arrow indicates subcutaneous layer. F, Femur; R, rectus femoris; VI, vastus intermedius; VL, vastus lateralis.

(Adapted from Pillen S, Arts IMP, Zwarts MJ: Muscle ultrasound in neuromuscular disorders, Muscle Nerve 37:679-93, 2008.)

Visual assessment and qualitative evaluation of muscle ultrasound imaging require a certain amount of experience before one can interpret these features reliably. The diagnostic accuracy of these features has not yet been investigated prospectively.

Contraction of the Entire Muscle

A normal voluntary contraction can easily be evoked and studied. It should be realized that during a normal contraction the firing of motor units is such that shortening of the muscle appears as a smooth movement. Individual motor units cannot be visualized at current ultrasound resolution. During contraction, the muscle shape changes, resulting in an increased cross-sectional area with decreased echogenicity. Smoother voluntary contractions can easily be differentiated from pathologic spontaneous muscle movements, such as fasciculations, that result in jerky, short, and local movements. Video 3.1 shows an example of a normal contracting biceps brachii muscle. During contraction of this muscle, the fascia dividing the medial, and lateral head of the biceps becomes visible.

Different involuntary movements such as tremor, cramp, myoclonus, and chorea can be visualized using ultrasound.⁵³ Tremors are repetitive, rhythmic, involuntary muscle contractions, often leading to oscillating movements of parts of the body. Tremor is associated with many neurologic disorders and can be the only clinical sign (as in essential tremor), or be part of a neurologic syndrome (e.g., Parkinson’s disease). Furthermore, tremor can also occur in healthy people as an exaggerated physiologic tremor. On muscle ultrasound, tremor is recognized as rhythmic contractions of the entire muscle. The observed movements on ultrasound may change according to the anatomic position of the extremity or depend on the posture of the patient. It has been shown that dynamic muscle ultrasound can be used to estimate the tremor frequency.⁵⁸

When studying patients with cramp, chorea, or myoclonus, the muscle contracting can be identified and visualized, but because the ultrasound appearance is essentially the same for all contractions, differentiation from voluntary contractions based on dynamic muscle ultrasound alone is not possible.⁵³

Fasciculations

Fasciculations are random, spontaneous contractions of the muscle fibers of a single motor unit. They are caused by a variety of reasons,⁵⁹ with affection of a part of the lower motor neuron as the common characteristic. Diseases in which fasciculations can be seen include motor neuron disease, various neuropathies, myelopathy, and toxic syndromes. Their occurrence and distribution can be important diagnostic clues, especially in amyotrophic lateral sclerosis.⁶⁰ Superficial fasciculations can be seen through the skin as a contraction of a small part of the muscle. If they are generated in deeper parts of the muscle, fasciculations can only be detected by EMG or muscle ultrasound.

Fasciculations can also occur as a benign phenomenon found in many healthy people, mainly in the calf and intrinsic foot muscles.^52,54,61,62 The prevalence of fasciculations increases with age and strenuous physical exercise. They usually do not occur in muscles above the knee.⁶³

On ultrasound, fasciculations appear as spontaneous–focal jerky movements that occur irregularly and randomly in different parts of the muscle (see Video 3.2). They are easily recognized, especially when screening in a transverse plane, as a focal muscle thickening in both the horizontal and vertical directions. In a longitudinal plane only a vertical thickening can be seen, which makes it slightly more difficult to identify fasciculations in longitudinal scans.

Several studies have investigated the sensitivity in detection of fasciculations by ultrasound.^52,55,62 The area of muscle that is examined during needle EMG is very small compared with the area that can be visualized using ultrasound.⁵⁵ With ultrasound the entire transverse plane of a muscle can usually be screened for fasciculations simultaneously. This is the main reason that the sensitivity for detecting fasciculations using ultrasound is higher than needle EMG or visual inspection of the skin.^52–5564 Moreover, fasciculations can still be recognized during voluntary muscle contraction, something that is not possible with EMG because of the summation of electrical potentials recorded by the needle electrode.⁵⁵ The specificity of ultrasound equals that of visual inspection (86%).⁵² Recognition is relatively easy with an interobserver agreement of 0.84 to 0.85, which is comparable with the recognition of fasciculations by EMG.^52,62

To detect any fasciculations present, ultrasound screening has to be performed for a sufficient period of time. In one study, 77% of the observed fasciculations occurred with a frequency of more than one per 2 seconds.⁵² This frequency was not related to the muscle studied. It can be calculated that screening a single muscle for 8 seconds yields a sensitivity of 95% to detect any fasciculations.⁵² Most authors, however, recommend a screening duration of 10 seconds.

Offline analysis of the ultrasound video allows for measurements of the cross-sectional area of a fasciculation and its temporal profile.⁵⁵ The cross-sectional area of a single fasciculation was shown to be different for smaller and larger muscles, with mean areas of 5.5 ± 3.3 mm² in the first dorsal interosseous and 58 ± 18 mm² in the gastrocnemius muscle. In patients with electromyographic evidence of reinnervation, particularly in post-polio patients, the area can be up to 109 mm².⁵⁵ The duration of a single fasciculation is 500 ± 110 ms, including a contractile phase, a plateau phase, and a relaxation phase.⁵⁵

Several pitfalls have to be taken into account when screening for fasciculations. Shivering and voluntary contractions may be mistaken for fasciculations, but in contrast to fasciculations involve the entire cross-sectional area of the muscle. However, in the case of several small muscles within a muscle group, such as the forearm flexors, the differentiation between short voluntary contractions and fasciculations can be difficult. A careful observation of the patient and reevaluation of the muscle when fully relaxed will usually solve this problem. Arterial pulsations are visible as small movements within muscle that can be easily differentiated from fasciculations, because they are rhythmic and unifocal (see Video 3.3). When in doubt, color Doppler imaging will easily tell them apart. Another pitfall can occur in patients with chronic denervation and collateral reinnervation of muscle fibers, leading to very large motor units. In these patients, slight voluntary contraction leading to activation of only one or two large motor units may resemble fasciculations clinically as well as on ultrasound. This clinical phenomenon is often referred to as contraction pseudotremor or contraction fasciculations. Prolonged contraction in these cases can also appear as fibrillations (see next section; also see Video 3.4). The main differentiation between fasciculations and contraction pseudotremor can be made on the response to contraction and relaxation: fasciculations do not disappear on relaxation, whereas contraction pseudotremor does. Additionally, contraction pseudotremor causes intramuscular movements that remain in the same place, whereas fasciculations appear randomly throughout the muscle. If questions remain, surface or needle EMG may be informative.

Fibrillations

Fibrillations are spontaneous depolarizations of the membrane of individual muscle fibers that result in contraction of an individual fiber. They result from the fiber’s loss of contact with its innervating axon.⁶⁵ For this reason fibrillations are an important sign of axonal loss in peripheral nerve disorders, but they can also occur in myopathies when muscle fibers are split or inflammation is present that separates part of the fiber from the endplate zone. Because fibrillations are invisible from the outside, they can only be detected by needle EMG and appear as spontaneous single-action potentials. New developments in ultrasound techniques, resulting in better resolution and higher frame rates, have made it possible to visualize fibrillations.^56,57 Although single muscle fibers are too small to visualize individually with the current maximal ultrasound resolution, fibrillating muscle fibers cause enough displacement of their surrounding tissue to be detectable. Fibrillations appear on the ultrasound video as small, irregularly oscillating movements of muscle tissue in all directions, while the overall shape of the muscle is preserved (see Video 3.5). One study indicated that the sensitivity of ultrasound for detecting fibrillations depends on the amount of fibrillations present and, critically, on limb temperature.⁵⁷ Ultrasound was found incapable of detecting fibrillations when EMG recorded less than 5 fibrillation potentials per second, and the visualization of movements decreased steeply below 30° C. However, EMG, which is the gold standard for fibrillation detection, can also miss focal areas with fibrillations within muscle because of sampling error, and here ultrasound has the advantage of visualizing larger areas of the muscle to help avoid this problem.

The detection of fibrillations is susceptible to several pitfalls. Artifacts caused by external movements or ultrasound noise caused by scattering on bone can be mistaken for fibrillations. Table 3.7 summarizes the main pitfalls in the detection of fibrillations and the methods used to avoid them.

Table 3.7 Pitfalls in the Detection of Fibrillations and the Methods used to Avoid Them