High-frequency ultrasound in clinical dermatology: a review

Dermatology is among the rare specialties of medicine in which the organ of interest can be readily examined and biopsied. As a consequence, histology has become the gold standard for many dermatologic diagnoses. It has also allowed for clinicopathologic correlation to guide therapeutics and advance prognostics [1]. In the last few decades, a number of imaging modalities have been introduced to augment clinical exam or obviate the need for histology [2]. Some examples include total body digital photography (TBDP), dermoscopy, reflectance confocal microscopy (RCM), optical coherence tomography (OCT), and ultrasound. In this review, we discuss how ultrasound fits into the landscape of skin imaging. We provide a brief history of its introduction to dermatology, explain key principles of ultrasonography, and review its use in characterizing normal skin, common neoplasms of the skin, dermatologic diseases and cosmetic dermatology.

Sonography dates back to the late 1700s when two Italian zoologists discovered the acoustic orientation sense of bats [3]. Nearly a century later, the Curie brothers presented the piezoelectric effect of sound waves on natural quartz [4]. It was their observation that a force applied to the crystal produced an electric discharge. Conversely, an alternating electric current applied to a crystal produces vibrations that generate high-frequency sound waves called ultrasound [4, 5]. Over a half-century later, the technique was adapted by both the American military as SONAR and by an American Neurologist, Karl Dussik, who used ultrasound to detect brain tumors in patients [5].

Ultrasound emerged in the practice of dermatology around 1980 when two independent groups used the technique to measure skin thickness in normal and diseased skin [6]. Since then, the practice has gained popularity in several European and South American countries that incorporate its use in resident training and where its application is reimbursable [7, 8]. As the technology has improved, so has its utility. In an early study of over 4300 skin lesions, Wortsman et al. demonstrated the ability of higher frequency ultrasound (7–15 MHz) to significantly improve diagnostic accuracy. Compared to clinical examination alone, ultrasound corrected the referring diagnosis in 17% of cases and detected 3 missed malignancies [9]. The application of ultrasound in dermatology is continuously evolving and increasingly studied [10, 11].

Ultrasound allows for high-resolution imaging of the skin from the stratum corneum down to the deep fascia [12]. Newer ultrasound technologies, including high-frequency ultrasound (HFUS, between 20 and 30 MHz) and ultra high-frequency ultrasound (UHFUS, > 30 MHz), have further enhanced imaging clarity and expanded upon the myriad applications for use of ultrasound in dermatology [11, 13, 14]. With fast acquisition times, ultrasound images are produced in real-time, allowing for physical adjustments and optimal imaging. Many devices include color Doppler analysis for characterization of blood flow and vessel morphology. It is readily available and can be easily incorporated into clinical practice [10, 11, 15].

The main component of the ultrasound device is the transducer—typically a handheld device containing thousands of piezoelectric crystals. These crystals, when stimulated by an electric voltage, generate millions of acoustic waves that pass through tissue and are reflected back to the transducer. As the waves return to the transducer, they are converted back into electrical energy interpreted by a computer. The computer then generates a one-dimensional line graph in amplitude mode (A-mode) that can be used to interpret echogenicity at various distances from the probe. In brightness mode (B-mode) an image is generated with different intensities of brightness.

These varying intensities are dependent upon the ultrasound waves’ ability to pass through the tissue. Bright or hyperechoic regions represent an abrupt increase in tissue density while dark or hypoechoic areas represent a decrease or no change in tissue density. A black anechoic region represents an area of tissue from which virtually no waves are reflected [5]. Although dense structures reflect a greater number of waves, resulting in a hyperechoic ultrasonographic structures, they may also generate an artifactual acoustic shadow due to attenuation of the ultrasound waves at the interface between two tissues of varying densities [13]. Another common artifact, acoustic enhancement, can appear as a result of relatively unimpeded passage of ultrasound waves through a preceding area of low acoustic impedance. This results in the appearance of a brighter structure deep to the area of low impedance [16, 17].

The resolution and penetration of an ultrasound image depends on the frequency of the acoustic waves produced. Traditional lower-frequency ultrasound devices are ideal for evaluation of internal organs like the liver and lungs due to their penetration ability. HFUS and UHFUS waves, on the other hand, are unable to travel easily through tissue, but provide higher-resolution, superficial detail. This makes HFUS and UHFUS ideal for visualizing a superficial organ like the skin [5, 10, 11, 15, 18, 19]. Given the diverse range of probe frequencies and the variation of dermal thickness in different regions of the body, specific frequencies should be considered for optimal visualization of lesions in various anatomic locations [15].

In the study of melanoma, ultrasound demonstrates a homogenous hypoechoic lesion. The shape of the lesion often correlates to the histologic subtype of the melanoma. For example, nodular melanoma will typically present with a spherical hypoechoic lesion, while a superficial spreading melanoma will present with a thinner, lenticular, hypoechoic lesion subjacent to the epidermal entry echo [13]. Although no distinct ultrasonographic features of melanoma have been described, one important strength of ultrasound is its ability to predict Breslow thickness of a melanoma [48, 56]. This correlation with histologic thickness, the current gold standard on which melanoma staging and surgical planning rests, is predicted to be over 92% [57]. When compared to other techniques like OCT, ultrasound has superior correlation with histology, repeatability and inter-rater reliability [58]. Where OCT tends to underestimate Breslow thickness, ultrasound has been shown to modestly overestimate tumor thickness [58]. The cause of this overestimation is thought to result from the technology’s inability to distinguish an inflammatory infiltrate from melanocytic proliferation and neovascularization. Another contributing factor may be the desiccation of tissue as it is processed for histologic review. Nevertheless, its ability to resolve well beyond the dermis renders HFUS superior to other imaging modalities in predicting Breslow thickness [59].

Though melanomas have been distinguished from benign melanocytic proliferations with specificities as low as 30%, the addition of color Doppler sharply improves ultrasound as a diagnostic tool [43]. Melanomas have been shown to demonstrate more dense vascularization than benign nevi and exhibit low-flow arterioles. In another study of pigmented lesions, the direction of intralesional blood flow demonstrated a 100% specificity in distinguishing melanoma from non-melanoma tumors [46, 60].

In a single-center experience with a 50-MHz device, the Longport, Inc. EPISCAN, the authors examined ultra-high-frequency ultrasound biomicrographs of nearly 70 cutaneous lesions as well as normal skin and adnexae. Transducer power gain and time-gain compensation (TGC) were set to the device’s default settings—100% and 10%, respectively, with a TGC start of 0%. Skin position, or distance between transducer and impermeable membrane barrier, or skin, was 7.5 mm. Scanning depth ranged from 3.8 mm to 15.0 mm, with a scan length of 15 mm at a rate of 1 scan per frame per second. A standard acquisition protocol was followed. The protocol included a horizontal (cranial–caudal) scan and perpendicular scan in the center of each lesion or anatomical area with and without pressure. Then, 3D-views were simulated using two sweeps in both planes over 3–8 s for a total of three to eight 2D-slices in each plane.

Normal skin of the non-sun exposed medial upper arm of a middle-aged man demonstrated a hyperechoic epidermal entry echo and hyperechoic dermis with horizontally arrayed collagen bundles of the reticular dermis. These superficial structures are followed by the hypoechoic subcutis transected by a hyperechoic band of superficial fascia and the hypoechoic biceps muscle invested by the hyperechoic deep fascia (Fig. 1). The thickened stratum corneum over the thenar eminence demonstrated a hyperechoic bilaminar entry echo (Fig. 2). Ultrasound of a first digit nail also demonstrated a bilaminar appearance with hyperechoic dorsal and ventral plates separated by a hypoechoic band (Fig. 3). Finally, a basal cell carcinoma on the medial mandibular cheek of an elderly man demonstrated a hypoechoic lesion in the superficial dermis with subtle posterior acoustic enhancement (Fig. 4).

A number of vascular lesions have been studied using ultrasound, including Port-wine stains (PWS), arteriovenous (AV) malformations and spider nevi. The first study to successfully examine PWS using ultrasound was conducted in 2018 [90]. This retrospective study utilized both 10-MHz and 22-MHz transducers with and without Doppler to assess lesional depth, echogenicity, vessel density and AV flow signals. The results of this study demonstrated a clear contrast between lesional and normal skin [90]. Furthermore, the study demonstrated ultrasound could help differentiate a PWS from other vascular anomalies due to their hypoechoic, poorly vascularized appearance with venous flow, while other capillary-arteriovenous malformations appear hyperechoic and well-vascularized with mixed arterial and venous blood flow [90].

Another type of vascular lesion known as a Spider nevus has also been characterized using ultrasound [91]. Spider nevi were historically classified as low-flow capillary malformations [92]; however, a 2017 study by Alegre-Sanchez et al. found that spider nevi demonstrated pulsatile flow on dermoscopy and ultrasound. This finding was subsequently supported by the presence of high-flow arterial-type waveform as seen on Doppler-ultrasound, suggesting that spider nevi are more similar to arteriovenous malformations, or high-flow vascular lesions [91].

Ultrasound can also be used to guide and assess response to treatment [48]. For example, researchers have used its Doppler capabilities to correctly identify venous structures and aid in guiding injectable foam sclerotherapy for the treatment of venous leg ulcers, demonstrating good healing rate and a low reflux ratio [93].

Several cosmetic applications of ultrasound have been established. For example, botulinum toxin injections demonstrate increase subcutaneous echogenicity leading to a blurring of the normally discrete boundaries between the subcutis and underlying muscle [94]. Similarly, fillers can be visualized using ultrasound to determine locations of prior injections—an important application since a small but significant proportion of patients seeking injectable fillers fail to report prior procedures. The risk of granulomas and filler dermopathies associated with injections of multiple filler materials has been established [95‐97].

The efficacy of topical vitamin C therapy in increasing collagen production, reducing ultraviolet radiation induced oxidative stress and enhancing dermal remodeling has been demonstrated in several studies [98, 99]. One notable study using HFUS evaluated the skin of patients throughout a 60-day period of topical vitamin C therapy, demonstrating increased echogenicity in both the epidermis and dermis after 40 days and to a greater extent after 60 days [98]. Ultrasound at 15-MHz and 20-MHz has also been used to evaluate the efficacy of an oral micronutrient supplement on preservation of skin quality through the measurement of skin thickness and density [99, 100].

While ultrasound has been used in dermatology for more than 40 years, its uses have continued to evolve with improvements in technology. Today, there are a few dermatologic lesions with pathognomonic sonographic features. One example is the “jellyfish” sign described in some cases of dermatofibrosarcoma protuberans [101]. Other examples include the “saw-tooth” pattern of cellulite [102‐104]. Pathognomonic ultrasonographic features are likely to increase with integration of higher frequency ultrasound devices in clinical practice; for future work with HFUS, a core international task force group formed by 15 physicians from 11 countries called the DERMUS (dermatologic ultrasound) group recommends dermatologists use Doppler-capable ultrasound devices with a frequency of at least 15- to 22-MHz in all ultrasound examinations of the skin [18].

While there is support for use of ultrasound in future dermatological practice, there are a number of limitations associated with the use of this tool. One major limitation is the skill of the operator. Consequently, the DERMUS group suggests utilizing ultrasound at least 300 times per year to achieve minimum competency. Similarly, it is strongly advised that the dermatologist functions as the sonographer because of the years of training in dermatologic pathologies and ability to correlate histology with clinical presentation of cutaneous lesions [18].

As frequency advancements in ultrasonography continue, the broad applications of this imaging modality will continue to grow. Ultrasound is a fast, safe and readily available tool that can aid in diagnosing, monitoring and treating dermatologic conditions by providing more objective assessment measures.