Correctly diagnosing and managing a wound that harbors a soft tissue foreign body can be challenging, especially when the foreign body is radiolucent. To further complicate matters, wounds that harbor foreign bodies often occur in the hand or foot where the likelihood of iatrogenic injury from blind wound exploration and the potential for subsequent infectious complications is high. Although usually located in superficial soft tissues, foreign bodies may cause no symptoms initially and can easily be overlooked. In a retrospective review of 200 patients referred for retained foreign bodies, 38% were misdiagnosed on the index visit.32 Even with a high index of suspicion, liberal use of radiography, and exploration, a soft tissue foreign body can be missed. The possible infectious and medico-legal consequences of this missed diagnosis can be unfortunate for the patient and the provider alike. Missed foreign bodies have been reported to be one of the most common causes of malpractice claims against emergency physicians.33
While metal and glass are radio-opaque and usually apparent on standard two-view radiographs, other commonly encountered foreign bodies, particularly organic material such as wood or thorns, are nearly always radiolucent. Plastic also is frequently radiolucent. CT or MRI may be useful in the assessment of suspected foreign bodies, but these modalities are expensive, time-consuming, and not always readily obtainable. Furthermore, the sensitivity of CT for the detection of wooden foreign bodies is low, reported to range from 0 to 60%.34, 35 Ultrasound offers some decided advantages in this setting. For detecting wooden foreign bodies—nearly always missed with plain radiography— ultrasound is 79–95% sensitive, and 86–97% specific.34, 36, 37 In a case where a radiolucent foreign body is suspected, ultrasound equipment can be brought to the bedside at the time of the examination, functioning as an extension of the physical examination. With ultrasound equipment now readily available in many emergency departments, bedside sonographic assessment of wounds suspected of harboring a radiolucent foreign body should become increasingly common.
Whether radio-opaque or radiolucent, once a soft tissue foreign body has been identified, the next issue faced by the clinician is how best to remove it. As most experienced clinicians will confirm, removal of a subcutaneous foreign body can be enormously frustrating. Ultrasound can additionally be used to provide precise preoperative localization of the foreign body, or, if desired, the foreign body may be retrieved under direct sonographic guidance.
The clinical indications for the use of ultrasound in the management of a suspected soft tissue foreign body include
- detection of a radiolucent foreign body,
- localization of a radiolucent or radio-opaque foreign body, and
- foreign body removal.
The literature on sonographic detection of soft tissue foreign bodies encompasses a wide range of specialties and methodologies.34, 36–49 The types of foreign bodies that have been studied include metal, wood, plastic, gravel, sand, and thorns or cactus spines. Sonographers with varying levels of skill perform the ultrasound examinations in these studies, ranging from emergency physicians with no prior formal training, credentialed sonographers, and radiologists specially trained in musculoskeletal ultrasound. The ultrasound machines and transducers are different in nearly every study. While this literature is therefore somewhat difficult to synthesize, a number of useful conclusions can be drawn.
Success in detecting foreign bodies varies widely in the experimental literature, depending in part on the tissue model employed and foreign body type. Using a homogenous beef cube as a tissue model, ultrasound was 98% sensitive and specific in identifying a variety of embedded foreign bodies in one report,39 whereas another study using a chicken thigh model (a model that more closely mimics the human hand) reported an overall sensitivity of only 79% for detecting a wooden foreign body.36 In studies involving “freshly thawed” cadaver feet and hands, diagnostic sensitivities and specificities ranged from 90 to 94% and from 90 to 97%, respectively.42, 47 In contrast to such excellent results, another investigation that used ultrasound for detecting foreign bodies in chicken thighs reported an overall sensitivity and specificity of only 43% and 70%, respectively, with a sensitivity of only 50% for detecting a 1-cm-long piece of wood.41 Review of the methods employed in this study revealed that the chicken thighs were incised and systematically opened with a hemostat prior to foreign body placement. Such tissue disruption with the likely introduction of subcutaneous air probably exceeds that which occurs in natural wounding and may have made subsequent sonography difficult. Of note, vigorous wound irrigation itself can introduce subcutaneous gas bubbles that interfere with subsequent attempts to locate small glass fragments with ultrasound. However, in another study in which air was purposely injected into turkey breasts containing glass, metal, and bone, the soft tissue gas did not appear to diminish the ability to locate the foreign bodies.50
Success in soft tissue foreign body detection also depends on foreign body size. The test characteristics of ultrasound reported among various studies must therefore be interpreted with an awareness of the size of the experimental foreign body being imaged. Small glass fragments and cactus thorns were difficult to detect in one report, and may have exceeded the limits of the ultrasound transducer's resolution.39 Variations in detection rates with two differing lengths of wooden toothpicks inserted into freshly thawed cadaver feet were reported. Sensitivity decreased from 93% for detecting a 5.0-mm-long fragment to 87% for detecting one that was 2.5 mm long.42 Specificities were uniformly high across studies, indicating that it is uncommon to falsely identify a foreign body when none is present.
While it might appear intuitive that sonographer experience and expertise would be a crucial determinant of success for foreign body localization, there is little experimental evidence to support this assumption. Only one study directly compared the ability of various types of sonographers to locate foreign bodies with ultrasound.36 It found no statistically significant difference in accuracy between a board-certified radiologist whose practice was limited to ultrasonography, two ultrasound technologists, and three emergency medicine residents. Sensitivity was 74% in the hands of the emergency physicians compared to 83% and 85%, respectively, for the radiologist and technologists.
In clinical case series, wood is the most common radiolucent material reported, hand and foot injuries predominate, and most foreign bodies are found to be superficial in location.34, 37, 51 One series of 50 patients evaluated for radiolucent foreign bodies noted that 45 of the 50 injuries involved the hand or foot.37 All of the 21 foreign bodies retrieved at surgery were found less than 2 cm from the skin surface in this report. In another case series of patients evaluated for suspected wooden foreign bodies in the feet, all 10 of the wooden foreign bodies discovered with ultrasound were located between 0.4 and 1.4 cm from the skin surface.34
Since hand and foot wounds are the most common injuries that may harbor a subcutaneous foreign body, a thorough familiarity with the anatomy of the hands and feet is essential for the clinician scanning these regions. Given the relatively shallow depth of the soft tissues in these anatomically intricate regions and the multiple acoustic interfaces present, clinicians should practice scanning on normal hands and feet to gain familiarity with the normal sonographic appearance of these commonly injured areas. The utility of examining the contralateral, uninjured extremity for comparison when a confusing sonographic finding is encountered cannot be overemphasized.
Technique and Normal Ultrasound Findings
The highest frequency linear array transducer available should be used when searching for subcutaneous foreign bodies since most will be found located within 2 cm of the skin surface. A linear array transducer in the 7.5–10.0 MHz range is generally recommended. A 7.5 MHz curvilinear transducer, such as an endocavitary probe, may also function adequately for this application and has the added advantage of a smaller, rounded skin contact footprint for scanning in web spaces.38, 52 A 5.0 MHz transducer may be useful when searching for a deep foreign body. Higher frequency small parts transducers (typically in the 10–13.0 MHz range) offer the ability to discern very small foreign bodies; a 12.0 MHz transducer can reportedly detect a 1–2 mm foreign body. In recent years, small parts transducers have become a more common addition to ultrasound equipment purchased for use in the ED. The combination of high-image resolution and a small skin contact footprint make these transducers useful for imaging digits and Web spaces for foreign bodies, in addition to ocular imaging, and selected procedural and musculoskeletal applications.
Use of an acoustic stand-off pad may be necessary with some transducers to adequately image the superficial soft tissues. Stand-off pads provide a sonolucent acoustic window, raise the transducer 1–2 cm above the skin surface, and move the subcutaneous region of interest beyond the extreme near field (and beyond the transducer's “dead zone”) into a more suitable focal zone. Although incorporating the use of a stand-off pad into the ultrasound examination requires additional technical agility and some practice, the effort can be amply rewarded with improved near-field image quality. Inexpensive, commercially available gel pads are available just for this purpose and smaller chunks can be cut off for single patient use and then discarded. Other options include the use of a water- or gel-filled glove or glove finger. When using a water-filled glove, it is essential to exclude any air bubbles that may impede subsequent imaging. In the case of a large or gaping wound, copious sterile surgical gel can be applied onto the wound; after the ultrasound examination is completed, the wound should be thoroughly irrigated. A water bath technique, in which the affected extremity is submerged in a basin of water during scanning, represents an alternative to the use of a stand-off pad or the copious use of gel. Compared to direct contact with gel, the water-bath technique was easier to perform and provided superior images of tendons and foreign bodies in one report.53, 54 The water bath technique is also reported to cause less patient discomfort, since images may be obtained without direct contact between the patient and the transducer.55 The sonographer should ensure that only sealed portions of a transducer are immersed in the water bath.
Optimizing depth and focus settings is particularly important when using ultrasound to search for a small, subcutaneous foreign body. The transducer should be held perpendicular to the skin surface and area of interest systematically scanned in two orthogonal imaging planes. Best visualization of a foreign body will occur when the transducer is aligned such that the long-axis of the ultrasound beam is parallel to the long-axis of the foreign body. Small objects can easily be missed when scanned in short-axis orientation alone. With a small wooden foreign body, however, it is sometimes the prominent posterior acoustic shadow on the short-axis view that alerts the sonographer to its presence. Because wounds containing foreign bodies can occur on any part of the body, a wide array of normal sonographic findings is therefore possible depending on the anatomic region being scanned. Most wounds suspected of harboring a foreign body occur in the hands and feet, however, where numerous anatomic structures and interfaces, each with a distinct sonographic appearance, will be encountered. The skin surface is the most superficial echogenic structure encountered, and is seen adjacent to the transducer surface (or, if an acoustic stand-off is used, adjacent to the distant side of the anechoic stand-off pad). In the hands and especially the feet, this layer is notably thicker than elsewhere on the body. Subcutaneous fat appears hypoechoic with a reticular pattern of echogenic connective tissue seen between the fat lobules. The thickness of this layer varies considerably with body location and habitus. Fascial planes appear as thin, echogenic, usually horizontal lines above the muscle layer immediately below. Muscle tissue appears relatively hypoechoic with regular internal striations (linear or pennate in long-axis, speckled in short-axis relative to the orientation of the muscle fibers). Tendons are moderately echogenic, appear ovoid and finely speckled in short-axis, and rectangular with a characteristic fibrillar sonographic pattern in long-axis orientation. Interestingly, tendons will appear considerably more hypoechoic when imaged obliquely; this characteristic of tendon imaging is known as anisotropy and is discussed in greater detail in the musculotendon portion of this chapter. Tendon movement can be observed real-time when the corresponding joints are moved. Bone appears brightly echogenic on the cortical surface closest to the transducer, with prominent posterior acoustic shadowing. Joint spaces can be readily identified by a V-shaped discontinuity in the bright cortical echo of adjacent bones. Blood vessels are anechoic and have a circular or tubular profile when scanned in short or long-axis, respectively. They can be further characterized with color flow Doppler, if necessary. In general, veins will compress easily with transducer pressure whereas arteries will remain pulsatile. Sonograms of the thenar eminence of a normal hand and a chicken thigh (commonly used as a tissue model for foreign body imaging) are shown in Figures 16-46 and 16-47.
Sonogram of the thenar eminence of a normal hand using a 7.5 MHz annular array transducer and an acoustic stand-off pad. The anechoic stand-off appears first, then the hyperechoic skin surface, followed by the hypoechoic thenar eminence muscles below. The flexor pollicis longus tendon is seen in cross section as a hyperechoic circle in the middle of the image. The hyperechoic surfaces of the first and second metacarpals are seen in the far field with associated posterior acoustic shadowing.
Sonogram of a chicken thigh using a 7.5 MHz annular array transducer and an acoustic stand-off pad (a similar technique is used for all the experimental foreign body images that follow). Note the tissue thickness and appearance is similar to that of the hand. The skin has been removed, thigh muscle tissue appears hypoechoic, the thigh bone on the left of the image appears hyperechoic with posterior acoustic shadowing, and an echogenic horizontally oriented fascial plane is seen in the far field.
Common and Emergent Abnormalities
Soft tissue foreign bodies exhibit a variety of sonographic patterns depending on the material involved, the size of the foreign body, and the length of time the foreign body has been present in the tissue. Common materials such as wood, glass, metal, plastic, and gravel will generally appear hyperechoic with variable amounts of posterior acoustic shadowing and associated artifacts that are material and shape dependent. A wooden foreign body typically casts a hypoechoic posterior acoustic shadow that often facilitates its discovery (Figures 16-48, 16-49, 16-50, and 16-51). Linear metallic foreign bodies will typically display a reverberation artifact with bright, regularly spaced parallel lines seen distal to the actual object (Figures 16-52, 16-53, and 16-54). Metal objects that are small, or rounded, may display a comet-tail artifact (Figure 16-55). The acoustic profile of glass is less consistent, however, and acoustic shadowing, reverberation artifact, or diffuse beam scattering may all be encountered during scanning (Figures 16-56 and 16-57). Foreign bodies retained for longer than 24 hours are frequently surrounded by a hypoechoic “halo” resulting from edema, pus, or granulation tissue. This hypoechoic region around the foreign body facilitates identification and localization of the foreign body (see Case Study). In a similar fashion, a local anesthetic injected adjacent to a foreign body may improve the ability to visualize it. Finally, sonograms of less commonly encountered foreign bodies such as plastic (with a prominent reverberation artifact) and gravel (with a prominent posterior acoustic shadow similar to a gallstone) are seen in Figures 16-58 and 16-59.
Short-axis sonogram of a wooden toothpick embedded in a chicken thigh. The hyperechoic wood fragment is seen in the near field. The posterior acoustic shadow draws the eye up to the location of the foreign body (arrow).
Long-axis sonogram of a wooden toothpick in a chicken thigh. The hyperechoic surface of the 2-cm wood fragment is seen in the near field in the center of the image (arrow points to center of toothpick); a posterior acoustic shadow is seen below.
Long-axis sonogram of a wooden foreign body in a patient's foot. A 7.5 MHz annular array transducer and an acoustic stand-off pad were used to obtain the image; the skin surface and immediate subcutaneous tissue appear hyperechoic. The wood fragment appears in the near field as a hyperechoic linear structure that slants to the right; a prominent posterior acoustic shadow is seen beneath the wood fragment.
Short-axis sonogram of the wooden foreign body in Figure 16-50. The hyperechoic wood fragment appears in the near field on the right side of the image (arrow). The transducer is no longer entirely in contact with the foot because of the location being scanned. A prominent posterior acoustic shadow is again seen beneath the foreign body. More distally and in the center of the sonogram, the first metatarsal bone and its posterior acoustic shadow are seen in cross section.
Long-axis sonogram of a needle in a chicken thigh. The needle appears hyperechoic (arrow) with a characteristic reverberation artifact seen below.
Long-axis sonogram of a broken needle fragment in the arm of an injection drug user. The needle appears hyperechoic (arrow). Although not appreciated on the sonogram, a fine reverberation artifact was seen on real-time imaging.
Clinically stable appearing victim of a gunshot wound to the right chest with a bullet seen on the left side in the chest radiograph. A mass was palpable beneath the skin on the left chest wall. A sonogram of the mass (the bullet) is notable for a reverberation artifact and posterior acoustic shadowing. The bullet was superficial in location and outside of the chest cavity (confirmed by CT).
Sonogram of a BB in a chicken thigh. A prominent comet-tail artifact is seen.
Sonogram of a linear glass shard embedded in a chicken thigh. The glass fragment in the center or the image appears hyperechoic with an associated reverberation artifact.
Sonogram of a piece of a broken glass bottle embedded in a chicken thigh. Although hyperechoic, the glass fragment in this image is indistinct with some dirty shadowing possibly due to air pockets surrounding the fragment.
Long-axis sonogram of a plastic toothpick in a chicken thigh. The plastic surface of the toothpick appears hyperechoic with a prominent reverberation artifact seen below.
Sonogram of a piece of gravel in a chicken thigh. A prominent posterior acoustic shadow is seen beneath the hyperechoic surface of the gravel fragment.
Common Variants and Selected Abnormalities
Various wound characteristics can complicate sonographic evaluation for a soft tissue foreign body. Air introduced into the wound by the injury itself, from the process of wound exploration or wound irrigation, or from bubbles inadvertently administered with the anesthetic agent can cause imaging difficulties. Air bubbles and associated artifacts may obscure the foreign body or may be mistaken for a foreign body when none is actually present. Air bubbles introduced during wound irrigation can complicate subsequent attempts to locate small glass fragments. Air pockets can sometimes be obliterated by compression with the transducer, thereby improving image quality. Ultrasound examination of large open wounds may be difficult because of bleeding, associated tissue distortion, or patient discomfort.
A 42-year-old man presented with a 10-day-old puncture wound of the finger with associated soft tissue swelling. The patient reported getting a splinter in his hand while sitting on a park bench. He thought he had removed the entire splinter at the time of the injury.
On physical examination there was swelling over the volar proximal third phalanx area but no definite fluctuance (Figure 16-60a). Distal neurocirculatory examination was intact, and there were no signs of lymphangitis or tenosynovitis.
Case study. Photo of swollen digit (A). Radiograph of the affected digit is notable only for some soft tissue swelling (B). Longitudinal sonogram on the volar surface of the third digit over the PIP joint shows a hyperechoic foreign body (arrow) in cross section (C). Note associated hypoechoic surrounding inflammatory response. Long-axis view of 1-cm splinter (arrow) (D). Wood fragment removed (E).
A radiograph of the affected digit was obtained and notable only for some soft tissue swelling (Figure 16-60b). An ultrasound was performed using a 7.5 MHz annular array transducer and an acoustic stand-off to obtain the images in both longitudinal (Figure 16-60c) and transverse (Figure 16-60d) planes. A hyperechoic foreign body with an associated hypoechoic surrounding inflammatory response was seen above the PIP joint. The wood fragment was easily removed in toto with a superficial skin incision along the PIP crease (Figure 16-60e).
This case emphasizes the advantage of sonography over plain radiography for identification of wooden (and other) foreign bodies. Foreign bodies retained for longer than 24 hours are frequently surrounded by a hypoechoic “halo” resulting from edema, or early infection around the foreign body. This “halo” can facilitate identification and successful removal of a foreign body.
Technological advancements over the past decade have led to the increased use of diagnostic ultrasound for the evaluation of a wide range of musculotendinous and rheumatologic conditions. Smaller, more portable ultrasound units, high-frequency small parts transducers with resolutions to a fraction of a millimeter, tissue harmonics, compound imaging technology, and extended field of view features have helped promote ultrasound as a diagnostic tool for evaluating musculotendinous complaints, whether in the hands of radiologists, rheumatologists, or emergency physicians. Ultrasound has emerged as “a powerful extension of the physical examination”56 and nowhere more so than in the realm of musculotendinous assessment. Imaging of tendons, joints, and muscles can help the provider make a correct diagnosis for a host of painful musculoskeletal conditions; armed with this knowledge, optimal treatment and management can follow. Self-teaching programs in musculoskeletal ultrasound have been promulgated for rheumatologists with excellent results reported after 24 hours of active scanning and 8–9 hours reviewing images with tutors.57
A myriad of musculoskeletal applications for diagnostic ultrasound have been described in the ultrasound literature. These include (1) evaluation of suspected partial or complete tendon tears (rotator cuff, triceps tendon, distal biceps tendon, Achilles tendon, quadriceps tendon, patellar ligament, and flexor tendons of the hand), (2) evaluation for muscle tears (specifically, the rectus femoris and gastrocnemius muscles), (3) diagnosis of occult ganglion cysts58 (wrist or fingers), (4) dynamic evaluation of flexor tendons of the hand (for evaluation of annular pulley ligament disruption or assessment of tendon location after flexor tendon avulsion), and (5) assessment of selected nerves for suspected entrapment or compression (e.g., the ulnar and median nerves in cubital and carpal tunnel syndromes). Some of these applications require expertise in ultrasound beyond that of most sonographers, or they may require specialized transducers that are not generally available in the emergency or urgent care settings. With the now common presence of linear array transducers available for compact ultrasound units, however, and the increasing appearance of small parts transducers available on newer equipment packages, a host of selected musculotendinous applications can easily be added to the diagnostic armamentarium of the emergency sonographer.
Clinical indications for performing a musculotendinous ultrasound examination may include
- assessment of suspected partial or complete tendon tears,
- evaluation of suspected tenosynovitis and selected tendinopathies,
- assessment of joint involvement in chronic arthritis, and
- precise guidance for aspiration and soft tissue injection procedures involving tendons and bursae.
A number of clinical studies have investigated the use of ultrasound to image rotator cuff tears with variable success.59–61 In an attempt to explain the wide range of reported accuracy for diagnosing rotator cuff tears with ultrasound (60–95%), a number of contributing factors were identified. The technical difficulty involved, the considerable experience required to perform the examination, the complex anatomy of the shoulder, and the occurrence of prominent beam propagation artifacts in the shoulder all combine to make sonographic evaluation of rotator cuff injuries a challenge.62 Because of these difficulties, only a few centers routinely use ultrasound to evaluate the shoulder for a suspected rotator cuff injury. MRI, with its excellent image quality and lack of operator dependence, has become the diagnostic imaging technique most commonly used for these injuries. In general, sonography of the rotator cuff has little utility for the emergency evaluation of a shoulder injury since the clinical examination and radiography will usually adequately guide the direction of clinical care. Similarly, ultrasound is generally considered unsuitable for evaluation of meniscal or other ligamentous injuries of the knee; MRI has become the imaging technique of choice for these injuries as well.
Complete disruptions of the rotator cuff, the biceps, triceps, quadriceps, or Achilles tendons are usually reliably diagnosed clinically and imaging is usually not required. If desired, however, sonography can be used to rapidly demonstrate the site of the specific tendon disruption at the bedside. Partial tendon tears present more of a diagnostic challenge. These may be difficult to diagnose on clinical grounds alone, and may therefore be misdiagnosed altogether. It is in this clinical scenario that ultrasound can ably assist the provider and help clarify the nature and extent of the suspected tendon injury. The Achilles tendon, quadriceps tendon, patellar ligament, and the triceps tendon all lend themselves to ready sonographic evaluation when a partial tendon tear is being considered. Each of these tendons is located in the superficial soft tissues and can therefore be examined in detail with a high-frequency transducer. The ability to both visualize the substance of the tendon and perform dynamic assessment of its function and integrity in real-time can offer important diagnostic information that may not be appreciated on physical examination or with plain radiography alone.
Tenosynovitis is also readily diagnosed with bedside ultrasound and its characteristic sonographic signature (hypoechoic fluid surrounding the tendon or tendons) makes arriving at this diagnosis simple. In a patient with a swollen and painful hand or foot, the ability to rapidly and confidently make a diagnosis of tenosynovitis improves patient care and assists with the implementation appropriate therapy. Similarly, selected tendinopathies can be diagnosed with a sonographic evaluation that reveals a characteristic focal area of hypoechogenicity of the tendon being assessed. A suspected ganglion cyst can similarly be quickly diagnosed with the visualization of an anechoic sac-like structure with a thin pedicle connecting it to a joint space or a tendon sheath. Use of ultrasound for the assessment of a host of painful arthritic joint conditions is discussed in detail in the section on ultrasound guided arthrocentesis (Chapter 20).
Finally, ultrasound guidance may be used to advantage for precise percutaneous injections of painful conditions of the musculotendinous system, specifically, injections of tendon sheaths, joints, bursae, and peri-fascial injections for the treatment of common tendinopathies, bursitidies, arthritidies, and fasciitis.63 Inadvertent administration of corticosteroids within the substance of a tendon has long been known to put the patient at risk of tendon rupture. Real-time ultrasound imaging allows for continuous observation of both needle placement and medication delivery and is therefore ideally suited for precise deposition of corticosteroids and local anesthesia used for treating these conditions. A specific example is the improved management of patients with subacromial bursitis. In a report investigating the treatment effectiveness of ultrasound-guided injections, a group of 40 patients with sonographically confirmed subacromial bursitis were divided into standard blind injection and ultrasound-guided injection groups.64 The outcome measure assessed was shoulder abduction range of motion preinjection compared with 1-week postinjection. No statistical differences were noted in shoulder range of motion at 1 week in the blind injection group, whereas a statistically significant difference was reported in the group that received the ultrasound-guided injections. The study authors concluded that ultrasound could be used to advantage in guiding the injection needle accurately into the inflamed synovial bursa with significant therapeutic benefits as a result.
Nearly all of the tendinous structures that are likely to be evaluated in an ED setting are superficial in location. Tendons consist of parallel fascicles of collagen fibers and may form a single homogenous bundle, or they may be composed of multiple bundles or laminae in which case they are referred to as complex tendons. A thin layer of connective tissue, the peritenon, surrounds all tendons. A peritendinous synovial sheath usually surrounds tendons that take a more curvilinear course. This sheath contains a thin film of synovial fluid that helps reduce tendon friction and abrasions. Peritendinous bursae may additionally reduce friction between a tendon and adjacent bones during movements. Tendons are sparsely vascularized and receive their nourishment primarily from segmental vessels arising from the surrounding peritenon or through vinculae.65
Technique and Normal Ultrasound Findings
Musculotendinous ultrasound imaging should, in general, be performed with high-frequency linear array or small parts transducers in the 7.5–13.0 MHz range. Sector transducers are generally not recommended for this application because the diverging beam of sector transducers can create undesirable beam scattering artifacts that leave only a small central portion of the image unaffected.62 This drawback can be somewhat compensated for by narrowing the sector angle and using a stand-off pad. When imaging very superficial structures, such as the Achilles tendon, patellar ligament, or structures within the finger, the use of an acoustic stand-off pad may be helpful. Attention should be paid to proper frequency, focus, and depth settings to obtain optimal images. A split screen set up may be useful for comparing corresponding images from each side of the body.
Skin is typically seen as a thin, hyperechoic layer adjacent to the transducer or stand-off pad. Subcutaneous tissue appears hypoechoic, is of variable thickness depending on body location and habitus, and has a somewhat hyperechoic reticular pattern of connective tissue between the fat lobules. Skeletal muscle is readily identified by its characteristic hypoechoic echotexture with echogenic internal striations that appear linear or pennate (feather-like) in long-axis and speckled in short-axis. Fascial planes are brightly echogenic and follow the surface contour of the muscle being imaged. Muscle fiber movement is often visible with muscle contraction in real time. In long-axis, tendons appear as hyperechoic rectangular or linear structures that generally track on a parallel course with the skin surface. The tightly packed parallel echoes emanating from the collagen fibrils within the substance of the tendon give rise to a tendon's characteristic fibrillar echotexture. In short-axis, tendons appear round, ovoid, or flat, and will usually appear as a cluster of multiple small echogenic dots. A subtle anechoic rim will be seen in tendons with a tendon sheath. Commonly imaged normal tendons are seen in Figures 16-61, 16-62, 16-63, 16-64, and 16-65.
Long-axis sonogram of the flexor pollicis longus (FPL) tendon in the thenar eminence of the hand (A). Somewhat hyperechoic skin and hypoechoic thenar muscles appear in the near field. Immediately beneath is the FPL tendon notable for its moderately echogenic and fibrillar appearance. The hypoechogenic region at the far right edge of the tendon is due to anisotropy. Tendon motion within the tendon sheath can be observed with real-time imaging. The curved echogenic surface of the first metacarpal bone and its acoustic shadow are seen in the right mid field of the image. Short-axis sonogram of the FPL tendon (B). Skin appears in the near field and the thenar muscles appear hypoechoic with some linear echoes that represent fascial planes. The FPL tendon appears round and echogenic with multiple small dots making up the substance of the tendon. The echogenic first metacarpal and its associated posterior acoustic shadow are seen in the left mid field of the image.
Short-axis sonogram of the biceps tendon (A). Skin and deltoid muscle appear in the near field. The anterior humeral head is seen in cross section as an echogenic line with a U-shaped depression known as the bicipital groove. The moderately echogenic biceps tendon (arrow) is located within the bicipital groove. Long-axis sonogram of the biceps tendon (B). Successive layers seen in this image are skin, hypoechoic deltoid muscle scanned along the long-axis of its muscle fibers, the hyperechoic and fibrillar appearing biceps tendon (arrow), and the brightly echogenic surface of the anterior humerus.
Sonogram of the lateral epicondyle of the elbow. The common extensor tendon appears somewhat hypoechoic beneath the skin and subcutaneous tissue. The radial collateral ligament is somewhat more hyperechoic in this image and connects the lateral epicondyle on the left with the radial head on the right of the image. The origin of the common extensor tendon is located somewhat higher up on the lateral epicondyle to the left of this image (arrow).
Sonogram of the patellar ligament. Long-axis view of the proximal aspect (A). The echogenic surface of the patella is seen on the left of the image with a sharply demarcated posterior acoustic shadow. The proximal portion of the patellar ligament is seen as a 3–4 mm thick, horizontal, and somewhat fibrillar appearing band coming off the inferior pole of the patella. Hoffa's fat pad lies beneath the ligament. Long-axis view of the distal portion of the patellar ligament (B). The fibrillar ligament is seen inserting on the tibial tuberosity to the right near field of the image. A small portion of Hoffa's fat pad is seen beneath the ligament on the left. An extended field of view feature or a longer linear array transducer would allow the entire ligament to be demonstrated on one image. Short-axis view of the proximal patellar ligament (C). The ligament (arrow) is seen to be thick and generally rectangular in configuration with slightly curved edges. It sits immediately beneath the skin, extends almost the entire width of the image, and is hyperechoic relative to the hypoechoic fat pad below.
Long-axis sonogram of the Achilles tendon. The tendon is seen immediately beneath the skin, appears about 5 mm in thickness, and has a fibrillar appearance in this long-axis view (above arrows). Posterior compartment muscles lie beneath the tendon. The calcaneus lies to the right but is not seen in this image. An acoustic stand-off might help bring the tendon into a more optimal focal zone.
Tendons exhibit an optical phenomenon known as anisotropy (i.e., the image obtained from the tendon is directionally dependent on the angle of the ultrasound beam) and a thorough understanding of this concept is essential for anyone engaged in musculotendinous imaging. A tendon will appear hyperechoic only when the insonating beam is precisely perpendicular to the tendon fibers. At all other angles, there is a significant reduction in the percentage of the ultrasound beam that is reflected back to the transducer and the tendon will therefore appear hypoechoic. Anisotropy is important because if a tendon is not properly imaged its internal structure cannot be adequately assessed.65 It should be noted that most significant tendon pathology manifests itself sonographically as a hypoechogenic defect. If inadequate attention is paid to the imaging technique, a focal area of false hypoechogenicity may be interpreted as representing evidence of tendinosis or tendon rupture when, in fact, the hypoechogenicity is simply due to the insonating beam not being perfectly parallel with the tendon fibers at that region. A heel–toe imaging technique helps avoid this pitfall; the transducer is gently rocked in its long-axis plane from one end to the other while scanning long-axis over a tendon. A focal area of hypoechogenicity that disappears with this technique is due to anisotropy; a defect that persists throughout changes in the angle of insonation likely represents true tendon pathology. Anisotropy is present in both long- and short-axis planes, although it is more apparent when scanning the tendon in long-axis.
Specific Imaging Techniques
The biceps tendon is most easily found by scanning the upper arm in short-axis just above the level of the axillary crease with the arm slightly externally rotated. The normal biceps tendon will be seen as a hyperechoic circular structure resting within the echogenic bicipital groove of the proximal humerus. Once located in short-axis, a long-axis view can easily be obtained by simply rotating the transducer.
The triceps tendon is best imaged from behind the patient with the transducer oriented in a midline long-axis orientation just above the olecranon process. This view is similar to that used when evaluating the elbow joint for an effusion. The triceps tendon will be seen in the near field as a moderately echogenic fibrillar structure that inserts on the echogenic olecranon process. The hypoechoic posterior fat pad and the echogenic outline of the posterior humerus, olecranon fossa, and olecranon will be seen in the far field of the image.
The lateral and medial epicondyles of the elbow are best imaged in long-axis across the joint line with the arm and elbow in full extension, and the hand in either a “thumbs up” position for lateral epicondyle imaging or in a hypersupinated position for medial epicondyle imaging. The bony outlines of the respective epicondyles and either the radial head (lateral) or coronoid process (medial) will be noted on the sonogram. The radial and ulnar collateral ligaments are thin structures found adjacent to and crossing the joint line. The common extensor or flexor tendons will be seen lying superior to the collateral ligaments and their point of origin lies several centimeters above the joint line on their respective epicondyles.
Hand and foot tendons are quickly assessed for tenosynovitis by placing the transducer in the region of clinical interest in a short-axis orientation relative to the direction of the tendon. Individual tendons are best examined in long-axis when evaluating for tendon disruption.
The patellar ligament should be scanned in longitudinal and transverse scan planes with the knee held in 30 degrees of flexion to avoid the “false hypoechogenicity” artifact that can occur in this location. This phenomenon is due to tendon anisotropy in the region adjacent to the patella. The contralateral knee should always be scanned for comparison.62
The Achilles tendon is best evaluated with the patient prone with the foot hanging over the edge of the examining table. Use of a stand-off pad can be helpful, and the foot should be plantarflexed and dorsiflexed to observe dynamic tendon movement.
Soft tissue injection techniques. Ultrasound imaging can provide real-time guidance for delivery of therapeutic injections of corticosteroids. A free-hand technique is most commonly used for injection of tendon sheaths, joints, and bursae. The guiding principle is that the medication delivery needle (typically a 1 1/2′′ 20–25 G spinal needle) should be oriented as perpendicular to the insonating beam as possible in order that the needle will appear maximally reflective. The needle's position will be apparent by its strong reverberation or ring-down artifact and its position can be further fine-tuned by injecting a small amount of anesthetic solution and observing the corresponding echoes. When a soft tissue injection procedure is undertaken, the tendons should be oriented so that they display maximal anisotropy for best visualization. A short-axis orientation over the tendon and a long-axis orientation over the injection needle are recommended for avoiding intratendinous injections.63
Common and Emergent Abnormalities
General Ultrasound Findings
Partial tendon tears are usually seen on ultrasound as hypoechoic to anechoic regions within the substance of the tendon. Complete tendon tears will demonstrate obvious tendon discontinuity during real-time scanning and the tendon sheath may be seen to be filled with a hypoechoic hematoma at the site of disruption. Tenosynovitis will sonographically manifest as tendon thickening, tendon sheath widening, and loss of the tendon's normal fibrillar echotexture. An associated anechoic collection of inflammatory synovial fluid will be seen surrounding the tendon. Synovial fluid is typically anechoic in acute inflammatory processes. Tendonitis or tendinosis usually manifests as areas of patchy hypoechogenicity and loss of fibrillar echotexture within the tendon. Tendons without a sheath will show thickening and altered echogenicity that varies according to the duration of the process.
Specific Clinical Scenarios
DeQuervain's disease is an overuse injury involving the first dorsal compartment or the hand (specifically, the abductor pollicis longus and extensor pollicis brevis tendons). Although a positive Finkelstein's sign usually confirms the diagnosis, the tendon can be rapidly imaged in short-axis to confirm the diagnosis in ambiguous cases. Fluid within the tendon sheath, tendon thickening, and pain on “sonographic palpation” are typical (Figure 16-66). Septic or reactive tenosynovitis may be encountered in the hands or feet, most commonly in injection drug users. The affected tendons in both entities typically appear thickened and hyperechoic with variable amounts of surrounding inflammatory synovial fluid (Figure 16-67).
Short-axis sonogram of a patient with tenosynovitis of the first dorsal compartment of the hand (DeQuervain's disease). The combined abductor pollicis longus and extensor pollicis brevis tendons appear hyperechoic and thickened and fluid is seen surrounding the tendon. An adjacent vein is seen to the left of the tendon sheath. There is no skin contact with the transducer on the right because the footprint of the transducer exceeds the relatively narrow curved surface of the radial wrist. Tenderness was present on “sonographic palpation.”
Long-axis sonogram of the thenar region in an injection drug user with a palmar tenosynovitis (A). The fibrillar flexor tendons of the hand are seen surrounded by a large amount of hypoechoic fluid. Short-axis sonogram of the same case in (B). Multiple flexor tendons are seen in cross section and a large amount of surrounding hypoechoic inflammatory fluid is noted.
Ultrasound can play a useful role in the evaluation of the athlete with chronic localized knee pain suggestive of tendinopathy of the proximal patellar ligament (“jumper's knee”). In a review of 25 surgically proven cases of “jumper's knee,” ultrasound correctly identified the lesion in all patients. The authors advocated the use of ultrasound as “the method of choice for the evaluation of jumper's knee, as it is cheap, noninvasive, repeatable and accurate.”66 Sonographically, the lesion appears as a localized area of hypoechogenicity in the central portion of the patellar ligament near its insertion on the patella. The patient usually has exquisite tenderness with palpation at this site (Figure 16-68). Diagnosis and injection therapy of another common tendinopathy, lateral epicondylitis or “tennis elbow,” is discussed in detail below.
Long-axis sonogram of the proximal patellar ligament showing a hypoechoic tendon defect near the origin of the patellar tendon (A). The remainder of the tendon appears fibrillar and echogenic. A heel–toe insonating technique confirmed that a hypoechoic defect was present and that a tendinopathy (“jumper's knee”) was present. Short-axis sonogram of the same patellar ligament (B). The ligament is seen as a somewhat echogenic horizontal structure about 5 mm beneath the skin surface and about 4 mm in width. In the central portion of the tendon there is a focal area of hypoechogenicity that persists with careful imaging. This is the classic location and appearance of a “jumper's knee” or tendinopathy of the proximal patellar tendon.
Partial and Complete Tendon Tears
The final common pathway for each of these lesions is excessive stress placed on the actively contracting muscle in question. Partial rupture of the proximal patellar ligament is apparent sonographically as a characteristic cone-shaped hypoechoic lesion, exceeding 0.5 cm in length, and found close to the origin of the patellar tendon at the inferior border of the patella. This hypoechoic lesion represents a focal discontinuity of the ligament in that anatomic area and an associated hematoma. An underlying tendonitis (“jumper's knee,” above) is often associated with this injury and is identified by thickening of the tendon and overall hypoechogenicity in the area of inflammation. A complete tear of the patellar ligament is usually clinically obvious. The disruption usually occurs at the point of attachment to the patella. On sonography the discontinuity between the patellar ligament and the patella will be apparent. An associated avulsion fracture fragment originating from the inferior pole of the patella may be seen attached to the proximal end of the ligament (Figure 16-69). A partial tear of the triceps tendon will appear as a focal area of hypoechogenicity seen along the anterior insertion of the triceps tendon on the olecranon process. The hypoechoic defect represents both torn tendon and an associated hematoma (Figure 16-70). Similarly, a partial tear of the Achilles tendon will appear as an obvious hypoechoic region of tendon discontinuity and hematoma. Dynamic imaging can help further clarify the extent and location of the tendon disruption (Figure 16-71).
Long-axis sonogram of a patellar ligament rupture. The ligament is retracted and thickened with a hyperechoic avulsion fracture fragment attached to its proximal end (arrow). A high riding patella sits out of the field to the left of the image. The ligament has completely torn from its attachment at the lower pole of the patella.
Long-axis posterior sonogram of the elbow in a patient with painful ROM after a fall on an outstretched arm. Plain radiographs were normal. The insertion of the triceps tendon onto the olecranon (O) reveals a large hypoechoic defect in the substance of the triceps tendon. Hypoechoic hematoma is seen extending around the tendon posteriorly. The hyperechoic posterior fat pad is seen below the partial tendon tear; no joint effusion was detectable in the posterior recess.
Long-axis posterior sonogram of an “ankle sprain.” The lower end of the Achilles tendon is seen in the near field about 5 mm beneath the skin. A large hypoechoic defect in the substance of the tendon is apparent (arrow), and on dynamic scanning the hypoechoic gap could be seen to widen. On short-axis views, some of the tendon was noted to still be intact. The hyperechoic curve of the posterior surface of the calcaneus (C) is seen in the right far field of the image.
The Achilles tendon is surrounded anteriorly and posteriorly by two bursae near its point of insertion on the calcaneus. The superficial calcaneal bursa lies in the subcutaneous tissue posterior to the Achilles tendon, and the retrocalcaneal bursa is found between the distal Achilles tendon and the calcaneus. These bursae are not typically seen in normal subjects and a sign of bursitis is that the bursae are demonstrable by ultrasound at all (Figure 16-72).69 Injection of the retrocalcaneal bursa is sometimes performed to relieve the pain of a chronic Achilles tendon bursitis. The patient is scanned prone with the tendon and bursa in a transverse scan plane using a 7.5 or higher MHz transducer. Needle placement and steroid injection into the bursa are best accomplished with a lateral approach.
Long-axis sonogram of the posterior heel in a patient with chronic heel pain (A). The Achilles tendon is seen in the near field as a thick fibrillar structure just beneath the skin. The echogenic line beneath the tendon represents the posterior surface of the calcaneus and the site of the Achilles tendon insertion. Between the Achilles tendon and the upper border of the calcaneus is a hypoechoic sac-like structure that represents the retrocalcaneal bursa. The bursal sac extends somewhat above and around the superior edge of the calcaneus and contains some echogenic debris within it, likely thickened synovium. Light probe pressure is required to avoid collapsing the bursa. Short-axis sonogram of retrocalcaneal bursitis in the same patient (B). This is the orientation that should be used for an intrabursal injection. The retrocalcaneal bursa is again seen as a hypoechoic sac just above the echogenic posterior calcaneal surface. Echogenic thickened synovium is seen on the left side of the bursal cavity. The Achilles tendon appears somewhat ovoid in this orientation, and the tendon fibers appear as multiple echogenic dots in short-axis. The injection needle should be inserted from the lateral aspect of the ankle and the needle directed into the bursal sac under real-time guidance. The injection needle should be maximally reflective in this orientation since it will be nearly perpendicular to the insonating beam.
Common Variants and Selected Abnormalities
A patient with a palpable mass in the hand, wrist, or digit may be assessed with ultrasound for the presence of a ganglion cyst. Ganglion cysts reportedly represent 50–75% of all soft tissue masses of the hand; they commonly appear as a sonolucent cystic or ovoid structure adjacent to a tendon sheath or the wrist joint. Typically, an anechoic linear duct will be seen extending from the ganglion cyst to the tendon sheath or joint giving the structure a “tadpole” appearance (Figure 16-73).67
Long-axis sonogram of a volar index finger in a patient with atraumatic finger pain. The bony outline of the proximal phalanx and the PIP joint are apparent. The flexor tendons reveal a finely fibrillar echotexture with a focal region of hypoechogenicity that persisted even with a heel–toe imaging technique to exclude anisotropy as the cause. A small hypoechoic ganglion cyst (arrow) is seen adjacent to this focal area of tendonitis and appears as a hypoechoic sac with a thin neck connecting it to the synovial sheath.
Ultrasound-Guided Steroid Injections
The techniques employed for ultrasound-guided soft tissue injection therapy have been reviewed above. Injection of a common tendinopathy, lateral epicondylitis, is demonstrated in Figure 16-74. Medial or lateral epicondylitis is thought to occur as a result of chronic repetitive stress and microtrauma of the common flexor or common extensor tendons, respectively.68, 70 Sonographically, there may be some swelling of the extensor or flexor tendon near its origin on the epicondyle. With lateral epicondylitis, there is a reported predilection for deep tendon fiber injury. Intratendinous hyperemia may additionally be noted with color flow assessment.
Long-axis sonogram over the lateral epicondyle in a patient with a clinical diagnosis of lateral epicondylitis. An ultrasound-guided steroid injection was performed at the site of maximal tenderness, several centimeters above the joint line (the joint line appears on the far right of the image, adjacent to the curve of the distal lateral epicondyle). The injection needle (located on upper left corner of image) is passed as perpendicular to the insonating beam as possible for maximal reflectivity and is readily identified by its reverberation artifact. The injection site is immediately above the common extensor tendon to avoid intratendinous administration of steroids that can lead to tendon rupture. The deeper fibers adjacent to the bony epicondyle (arrow) reveal a focal area of hypoechogenicity that is typical for lateral epicondylitis.
Ultrasound may also play a useful role in the management of patients with subacromial bursitis. On a subacromial scanning window (patient seated, arm behind the back with the elbow flexed), subacromial bursitis will appear as a hypoechoic fluid collection between the deltoid muscle and the supraspinatus tendon.
Other Musculotendinous Sonographic Findings
With biceps femoris, rectus femoris or gastrocnemius muscle tears, a “clapper in the bell” sign may be seen on the sonogram where the retracted, ruptured upper portion of the muscle (the clapper) is surrounded by a hypoechoic hematoma (the bell). Gouty tophi will demonstrate posterior acoustic shadowing, much like gallstones. Inflamed bursae (also discussed in the Arthrocentesis section) will appear as superficial hypoechoic fluid collections. Care should be taken to limit transducer pressure on smaller bursae to avoid causing them to collapse and thereby displacing the bursal fluid out of the field of view.
The most important potentially misleading artifact in tendon sonography is the false hypoechogenicity that results from the slightest obliquity of the ultrasound beam in relation to the tendon fibers.62 Since areas of hypoechogenicity are the sonographic clues to tendon pathology (specifically, tendon disruption and tendonitis), optimal tendon imaging is paramount. The angle of insonation should ideally be as parallel to the course of the tendon fibers as possible to avoid this significant pitfall. Where areas of tendon hypoechogenicity are encountered, a heel–toe scanning technique should be employed to further clarify the true sonographic character of the area in question.
Clinical Considerations and Indications
When evaluating a patient with preauricular or submandibular swelling or tenderness, a bedside ultrasound examination of the area can rapidly clarify if the source of the problem lies within the parotid or submandibular gland. Although CT is the traditional imaging modality of choice for parotid and submandibular inflammatory conditions, and MR is considered the modality of choice for evaluation of parotid and submandibular tumors, ultrasound is considered by some to be the initial imaging modality of choice for assessment of any palpable abnormalities of the parotid gland.71 Ultrasound can clarify if an intra- or extraglandular lesion is present, provides information as to whether a focal process or diffuse glandular involvement is present, and can help guide aspiration if an abscess is found in a case of acute bacterial sialadenitis. Ultrasound is particularly useful in the evaluation of suspected sialolithiasis where it has a reported 96% diagnostic accuracy.72 Given the somewhat questionable reliability of the clinical examination (one report notes that 30% of a series of 38 preauricular masses were not intraparotid in location73) and the ready availability of portable ultrasound equipment in many emergency departments, ultrasound examination of the region of facial swelling or tenderness should be considered to be part of the bedside workup of any patient with suspected salivary gland disease in the ED.
Salivary gland disease can be broadly classified into four categories: acute sialadenitis (with viral and bacterial etiologies), chronic sialadenitis (with infective and noninfective etiologies), sialolithiasis, and tumors (benign and malignant). The sonographic features of these diseases will be discussed in further detail below. The majority of parotid tumors (85–90%) are benign pleiomorphic adenomas. They are slow growing, painless masses seen most commonly in middle-aged patients and are usually found in the superficial lobe of the parotid gland. By contrast, 50% of tumors in the submandibular gland are found to be malignant. Salivary gland calculi may occur in the parotid gland, but they are most commonly found in the submandibular gland, presumably because of the more mucinous content of submandibular gland secretions. Most salivary gland calculi occur in the submandibular duct; the remainder are found within the gland or in the ductal hilum. Multiple stones are found in 25% of patients.72
The body of the parotid gland lies in a preauricular location, its uppermost portion roughly in line with the external acoustic meatus and from there the gland extends inferiorly and posteriorly to the angle of the jaw. The parotid duct (Stenson's duct) arises from the anterior border of the gland, lies superficial to the masseter muscle and about 1–2 cm below the zygomatic arch, courses horizontally at earlobe level through the buccal fat pad, and pierces the buccinator muscle to enter the mouth at the parotid papilla opposite the upper 2nd molar. The normal duct is approximately 2–3 mm in diameter and 4–6 cm in length. On occasion, an accessory parotid gland may be seen lying anterior to and following the course of the parotid duct. The gland is broad and flattened superficially and wedge-shaped on its posterior and deep aspects. The bulk of the gland overlies the masseter and mandible. The facial nerve, the retromandibular vein, and the external carotid artery lie in a vertical orientation immediately deep to the gland. Lymph nodes may be found within the substance of the parotid gland.
The submandibular glands are found beneath the superficial subcutaneous tissues of the submandibular triangle, lateral to the anterior belly of the digastric muscle. The much larger superficial lobe and the much smaller and more posterior deep lobe are C-shaped in long-axis profile and connect where they wrap around the lateral border of the mylohyoid muscle. Intraglandular ducts drain into the submandibular (or Wharton's) duct that emerges from the hilum of the submandibular gland. The submandibular duct passes medially, then up and over the lateral border of the mylohyoid muscle where it then courses medial to the sublingual gland to a papilla adjacent to the lingual frenulum in the anterior floor of the mouth. The duct is about 5 cm in length. Unlike the parotids, no intraglandular lymph nodes are found within the submandibular gland. The named Küttner lymph node is reliably found in the space between the posterior border of the submandibular gland and the anterior border of the inferior aspect of the parotid gland.72
Sublingual glands are located below the mucous membranes of the floor of the mouth adjacent to the mandible and the genioglossus muscle. They drain via numerous small caliber ducts either directly into the floor of the mouth or into the submandibular duct.
Technique and Normal Ultrasound Findings
The parotid and submandibular glands are best imaged with a 7.5–12 MHz linear array transducer. The parotid gland is scanned in long-axis in a vertical scan plane in front of the lower ear. The superior portion of the transducer will need to be angulated somewhat anteriorly, however, to evaluate the portion of the gland that lies inferior to the ear at the angle of the jaw. The gland is predominantly superficial in location with some deeper portions of the gland hidden by the mandible. When evaluating the parotid duct in its long-axis, a transverse scan plane is employed where transducer orientation is nearly horizontal from mid earlobe level to the mid cheek. The parotid duct appears as two closely spaced parallel echogenic lines with a thin region of lucency between them, and is found superficial to the masseter muscle. Because of the fatty glandular tissue composition of the gland, the normal parotid appears fairly homogeneous with a fine granular echotexture that appears similar to a fatty liver (Figures 16-75 and 16-76). Intraparotid ducts appear as echogenic linear structures within the substance of the gland. Intraparotid nodes are common, most commonly in a preauricular location. Lymph nodes are typically elliptical in shape with a hypoechoic periphery and a hyperechoic fatty central hilum. When scanning the parotid in a long-axis orientation, the retromandibular vein and the external carotid artery will appear as two parallel hypoechoic channels beneath the gland; they are best appreciated with the use of color Doppler imaging. The facial nerve may be seen as a thin fibrillar structure overlying the more superficially located vein.
Long-axis sonogram of a normal parotid gland. The gland demonstrates a fine homogeneous mid gray echo texture that appears similar to a fatty liver. Immediately beneath the gland lies the retromandibular vein and beneath it, the external carotid artery. Their locations are best appreciated with the use of color Doppler imaging.
Transverse sonogram of a portion of the parotid duct. The normal parotid duct (arrow) appears as two narrowly spaced echogenic lines in the near field with the masseter muscle in cross section below (seen on the left half of the image). The transducer is placed just in front of the ear in a horizontal orientation at the level of the mid earlobe.
When scanning the submandibular gland, the transducer should be placed in the submental region just medial to the mandible aiming toward the middle of the chin. From whatever side is being imaged, the orientation marker should always face left. By so doing, the chin will appear on the right side of the image when scanning the right submandibular gland, and on the left side of the image when scanning the left submandibular gland. The bulk of the gland appears posterolateral to the mylohyoid muscle and appears as a rounded and somewhat lobular structure with a finely granular echotexture identical to that of the parotid gland. The Küttner lymph node may be seen posterior and adjacent to the submandibular gland. The mylohyoid muscle appears as a horizontally oriented and somewhat striated appearing rectangular region of hypoechogenity in the near field of the image; the muscle tapers as it approaches its insertion on the symphysis menti. The inferior aspect of the symphysis menti appears as a slightly curved region of hyperechogenicity in the near field with a dense posterior acoustic shadowing beneath. The submandibular duct is somewhat narrower in caliber than the parotid duct, has a similar hyperechoic tubular appearance, and is best visualized when it is pathologically dilated. When scanning the submental region, it is important to remember that the patient's anatomy on the sonogram appears upside down; the image obtained represents a somewhat oblique sagittal section of a patient standing on his or her head.
Common and Emergent Abnormalities
A variety of sonographic patterns may be encountered with parotid and submandibular gland disease. With acute viral sialadenitis (most commonly caused by mumps) and acute bacterial sialadenitis (most commonly caused by Staphylococcus aureus and Streptococcus viridans) the gland appears enlarged, hypoechoic, and of heterogenous echotexture. With bacterial sialadenitis, air may be present within the intraglandular ducts and small hyperechoic foci with associated comet tail artifacts may be seen. If an abscess has formed, it will usually appear hypoechoic or anechoic and demonstrate posterior acoustic enhancement similar to a subcutaneous abscess. Chronic sialadenitis may occur as a result of infective or noninfective causes and has a sonographic appearance that varies with the type and stage of the disease.
Solid lesions will typically appear hypoechoic and demonstrate posterior acoustic enhancement. Cystic lesions will typically appear anechoic. A pleiomorphic adenoma is the most common benign solid parotid tumor encountered and comprises 85–90% of parotid tumors. Sonographically it appears rounded or lobular in shape, well circumscribed, and is homogeneously hypoechoic with posterior acoustic enhancement. Because the most common malignant solid tumor (mucoepidermoid carcinoma71) may initially appear sonographically similar to a pleiomorphic adenoma, all solid lesions should be referred for definitive assessment and diagnosis.
A salivary duct stone or calculus will appear as a hyperechoic focus with prominent posterior acoustic shadowing. It will most typically be found within the salivary duct of the submandibular gland, and occasionally within the ductal hilum. Less commonly, a calculus may be encountered within the parotid duct or gland. The salivary duct in such cases will typically be quite dilated and will appear in long-axis as a prominent rectangular or tubular anechoic region leading up to the stone. The gland itself may appear enlarged with dilatation of the hilum of the gland giving the gland a somewhat “hydronephrotic” appearance (Figures 16-77 and 16-78).
Submandibular gland in a patient with calculus sialadenitis. The gland is enlarged and there are hypoechoic regions within the gland and hilum caused by ductal dilatation from the distal outflow obstruction caused by the stone. The salivary duct exits from the gland and courses up and over the lateral border of the mylohyoid muscle.
Long-axis sonogram of the distal submandibular duct. The dilated submandibular duct is seen in long-axis as an anechoic channel in the far field. A large echogenic salivary duct calculus is lodged in the distal duct; posterior acoustic shadowing is noted in the far field beneath the stone. The hypoechoic and striated mylohyoid muscle is seen just beneath the skin and subcutaneous tissue at the top of the image and its insertion onto the mandible is seen on the left side of the image. The inferior and posterior edge of the mandible appears as a curved echo with posterior acoustic shadowing on the far left of the image. For orientation purposes it is important to remember that the top of the head is at the bottom of the image.
Finally, a preauricular mass or tenderness will sometimes be due to a reactive and enlarged lymph node. The lymph node has a characteristic elliptical shape with a hypoechoic periphery and an echogenic fatty central hilum (Figure 16-79).
Longitudinal sonogram of an enlarged preauricular lymph node. A tender mass was clinically palpable in front of the ear; the sonogram demonstrates an enlarged intraparotid preauricular lymph node (elliptical in long-axis with a hypoechoic periphery and a hyperechoic central hilum). The node is surrounded by a thin rim of granular mid gray echogenicity that corresponds to normal parotid tissue.
The major pitfall with ultrasound imaging of the salivary glands is failure to consider a malignant etiology as a cause for a solid lesion. Since benign and malignant lesions may have similar sonographic appearances, all solid lesions should be referred for definitive workup. The role of the bedside ultrasound examination should be to identify if a lesion is present, to establish whether it is intraglandular or not, and to identify if calculus disease is the cause of the swelling or pain. Nearly all other pathologies should be considered for referral and further evaluation with the exception of diffuse polyglandular enlargement (as seen with mumps), diffuse uniglandular enlargement and tenderness (as seen in sialadenitis), sialolithiasis, an obvious abscess cavity, and a reactive lymph node.
Rhinosinusitis is one of the 10 most common diagnoses in ambulatory practice and the fifth most common diagnosis for which antibiotics are prescribed.74 Bedside diagnosis of bacterial sinusitis, for which antibiotics are recommended, is difficult by history and physical examination alone, however. At most, only 50% of patients presenting to emergency or general care settings with sinus complaints will actually have bacterial sinusitis.75 Signs and symptoms of bacterial sinusitis are nonspecific and may be indistinguishable from the clinically similar presentation that occurs with viral rhinosinusitis. Although radiography improves diagnostic accuracy somewhat, neither plain radiography nor CT is recommended for evaluating uncomplicated sinusitis in ambulatory patients because the additional time, cost, and radiation exposure produce little additional diagnostic accuracy.76 A fluid-filled maxillary sinus can readily be identified by ultrasound, however. Ultrasound may, therefore, be employed as a rapid and safe diagnostic aid at the bedside when evaluating a patient suspected of having maxillary sinusitis. Sinusitis is also an important occult cause of fever and nosocomial pneumonia in patients undergoing prolonged mechanical ventilation.77 In the ICU setting, bedside sinus ultrasound is a well-established screening test for detecting suspected maxillary sinusitis in intubated patients.
The primary indications for sinus ultrasound in emergency and ambulatory practice are as follows:
- Confirm or exclude clinically suspected maxillary sinusitis in patients with upper respiratory symptoms
- Detect maxillary sinus fluid in intubated ICU patients to screen for suspected nosocomial sinusitis.
Although CT and MRI are sensitive imaging modalities for diagnosing sinusitis, the presence of sinus fluid or mucosal thickening alone does not necessarily indicate bacterial infection (low specificity and low positive predictive value). The true “gold standard” for diagnosing bacterial sinusitis is a positive culture from fluid obtained on sinus puncture. Ultrasound, while less sensitive for mucosal changes and small amounts of fluid, is probably more specific for clinically important disease because a positive study requires the presence of a significant amount of fluid in the sinus. Studies evaluating the diagnostic characteristics of ultrasound for bacterial sinusitis vary substantially in terms of the population studied (general practice, subspecialty, or ICU), the criterion standard employed (sinus puncture, radiography, MRI), and methodologic quality. A summary of studies conducted in the 1980s and 1990s in which sinus ultrasound was compared to sinus puncture found a weighted mean sensitivity and specificity for ultrasound of 85% and 82%, respectively. This compares favorably to the 87% sensitivity and 89% specificity found for plain radiography.78 A systematic review of these early studies concluded that operator skill and experience had a large effect on how well ultrasound performed.79
Three contemporary studies have been conducted in the ambulatory and emergency care setting comparing bedside ultrasound to radiography or MRI for diagnosing acute maxillary sinusitis. A study found that, using MRI as the gold standard, ultrasound was 64% sensitive and 95% specific compared to 73% sensitive and 100% specific for plain film radiography.80 The authors concluded that a positive ultrasound examination confirmed the diagnosis of maxillary sinusitis. If the ultrasound examination was negative and clinical suspicion was high, however, they recommended that plain-film radiographs be obtained. A study found that ultrasound was 92% sensitive and 95% specific compared to plain film radiography. In the primary care setting in which their study was conducted, the addition of bedside sinus ultrasound to the history and physical examination would have reduced antibiotic prescriptions for sinusitis by one half.81 Finally, Price and colleagues compared emergency physician–performed ultrasound to CT in 48 emergency department patients with suspected maxillary sinusitis. Sensitivity was 81% and specificity 89%.82
Maxillary Sinus Fluid in Intubated Patients
In the ICU setting, bedside ultrasound has emerged as a convenient screening test for nosocomial sinusitis, both because plain film radiographs are inaccurate assessing for sinus fluid in recumbent patients, and because obtaining CT imaging in intubated patients is particularly time and labor intensive. Ultrasound can be used at the bedside to assess for the presence of fluid in the maxillary sinus. If fluid is present, rhinoscopy or sinus puncture is subsequently performed to obtain fluid for culture. The reported prevalence of infected sinus fluid varies widely among different studies, ranging from 5% to 60% of cases.77, 83, 84 Among the numerous studies investigating the use of maxillary sinus ultrasound in the ICU setting, the three best compared ultrasound to radiography or sinus puncture (presence of fluid), and found a sensitivity of 67–100% and specificity of 86–97%.83, 85, 86 Specificity was 100% when all sinus walls were seen, which generally correlates with complete sinus opacification on radiography.
The maxillary sinuses are paired and somewhat pyramidal-shaped airspaces within the maxillary bone on either side of the nose. They are bordered by the orbital floor superiorly, the lateral nasal wall medially, the alveolar process and hard palate inferiorly, and the zygoma laterally. The maxillary sinus is typically 2–4 cm in anteroposterior depth, and fluid within the sinus renders the posterior wall of the sinus visible by ultrasound. The ethmoid sinuses lie superomedially, and the sphenoid sinus lies in the midline deeper within the skull. The frontal sinuses, which would seem to be sonographically accessible, are not well visualized by ultrasound, likely because the anterior surface of the frontal bone is quite thick.
Technique and Normal Ultrasound Findings
Maxillary sinus ultrasound may be performed with a wide array of transducers ranging from a phased array or microconvex 3.5 MHz transducer (e.g., a typical cardiac transducer) to a 3–10 MHz linear array transducer. A small skin contact footprint is preferred, however, for scanning on the curved, relatively solid surface over the anterior maxilla. Although one might think that lower frequencies would be required to adequately penetrate the anterior bony wall of the sinus, this does not appear to be the case in practice, and a fluid-filled sinus will be apparent with either transducer type. The patient should sit upright or lean slightly forward to ensure that sinus fluid, if present, layers out against the anterior wall. The sinus should be scanned in both sagittal and transverse planes, midway between the nose and the zygoma, just below the orbital rim. When using a linear array transducer, the sagittal image is easier to obtain by angling the transducer slightly superomedially, parallel to the nasolabial fold. Attention should be paid to proper depth settings (typically 5–7 cm) so that the image of the sinus fills about three fourths the screen. Comparison with the uninvolved sinus should be routine and can be useful in equivocal cases.
In the normal air-filled sinus, a prominent periodic resonance artifact will be apparent, consisting of an evenly spaced series of echogenic lines that parallel the shape of the anterior surface of the maxilla and diminish in intensity at increased depth. Beneath the distinct echo of the anterior wall of the sinus, an indistinct “snowstorm” appearance will be noted on the sonogram and the posterior wall of the sinus will not be apparent. Of note, it is important not to confuse one of the deeper periodic resonance artifacts as representing the posterior wall of the sinus (Figure 16-80).
Sagittal sonogram of a normal sinus taken with a linear array transducer. Skin and subcutaneous tissue appear diffusely hyperechoic in the near field. A bright echo is seen at the level of the anterior surface of the maxillary sinus. The series of evenly spaced horizontal echoes that diminish in intensity at increasing depth represent a periodic resonance artifact emanating from the anterior wall of the sinus. A “snowstorm” pattern of echogenicity is normally seen within the nonopacified sinus.
Common and Emergent Abnormalities
Acute viral rhinosinusitis produces abnormalities within the maxillary sinuses in up to 87% of cases, typically a thickening of the sinus mucosa along with some secretions. Occasionally, a significant amount of fluid will accumulate within the sinus and produce an air-fluid level on CT.87 An ultrasound examination of the sinus at this point would appear positive for fluid, and would represent a false-positive result for the diagnosis of bacterial maxillary sinusitis. Bacterial sinusitis is said to supervene in 1–2% of cases of viral rhinosinusitis75 and typically occurs after 5–7 days of symptoms. It may be accompanied by mucopurulent nasal discharge and signs of maxillary inflammation (such as focal, often unilateral sinus pain and tenderness). As the inflammatory process evolves, pus accumulates in the sinus, giving rise to air–fluid levels or complete opacification on radiography, decreased transillumination, and a positive sinus ultrasound examination.
In the patient with a completely fluid-filled sinus, the curved posterior wall of the sinus will be clearly apparent in the far field of the image and no periodic resonance artifact will be noted (Figure 16-81). If the sinus is only partially filled with fluid, a mixed picture will be seen. The posterior wall of the sinus will be apparent only in the lower fluid-filled portion of the sinus (right side of the image on a sagittal sonogram) and a periodic resonance artifact may be seen in the upper air-containing region of the sinus (left side of the image). An example of partial sinus opacification on CT and its corresponding sonographic representation are illustrated in Figure 16-82a,b. A continuous brightly echogenic line along the posterior wall of the sinus on sagittal imaging correlates well with complete sinus opacification on radiography. This has been referred to as a “sinusogram.”85 A discontinuous line, and some persistence of the periodic resonance artifact from the anterior wall, correlates with partial sinus opacification.83, 85
Sagittal sonogram of a patient with acute maxillary sinusitis. The brightly echogenic curve in the far field of the image represents the posterior wall of the sinus. The sinus cavity is hypoechoic and no periodic resonance artifacts are seen.
A coronal CT image of a patient with bilateral maxillary and ethmoid sinusitis (A). The left maxillary sinus is nearly completely fluid-filled; the one on the right is only partially opacified. Although this is the preferred orientation for CT imaging of the sinuses, it should be noted that the sonogram obtained when scanning the maxillary sinus will be orthogonal to this scan plane. Sagittal sonogram of the partially opacified right maxillary sinus (B). Image taken with a microconvex 3.5 MHz transducer. On the right side of the image (corresponding with the inferior fluid-filled portion of the sinus), the posterior wall of the sinus is apparent as a brightly echogenic line. On the left side of the image (corresponding with the air filled portion of the sinus) the image is indistinct and the posterior wall of the sinus is not apparent. A large amount of near field artifact is present because gain settings are too high.
The most important potential pitfall when evaluating suspected maxillary sinusitis with ultrasound is failure to consider the results of imaging in the context of the pretest likelihood of disease and other clinical indicators of severity. The sensitivity of ultrasound for maxillary sinusitis is not more than 85% and involvement of other sinuses cannot be assessed. Thus, in the patient with severe symptoms, in the elderly or diabetic patient, or when the clinical suspicion for sinusitis is high, a more sensitive imaging study such as CT should be obtained for definitive diagnosis. On the other hand, evidence of maxillary sinus fluid in what otherwise appears to be an acute viral rhinosinusitis should be treated conservatively with decongestants. It should also be noted that a number of conditions other than sinusitis will allow transmission of the insonating beam through the sinus and will result in a positive scan of the posterior wall of the sinus. Significant mucosal thickening, polyps, fluid-filled cysts, occasional solid masses, as well as blood from facial trauma and sinus fractures can all give rise to a positive “sinusogram” (Figures 16-83 and 16-84). Technical pitfalls include scanning with the patient in a supine position, failure to accurately set the image depth to a level that would include the posterior sinus, and misinterpreting a periodic resonance artifact as an image of the posterior wall of the sinus.
Sagittal sonogram of a patient with a polyp filling the lower half of the right maxillary sinus. The posterior wall of the sinus is visible in the inferior portion of the sinus (right side of the image) and posterior acoustic enhancement is evident. The upper portion of the sinus (left side of the image) has more of a “snowstorm” appearance with subtle reverberation and periodic resonance artifacts and no visualization of the posterior wall of the sinus.
Sagittal sonogram of a patient with fractures of the anterior wall of the maxilla and a blood-filled maxillary sinus. Note the marked soft tissue swelling and the cortical irregularity of the anterior wall of the maxilla in the near field. The posterior wall of the sinus is clearly seen in the far field.