An Observational, Prospective Cohort Pilot Study to Compare the Use of Subepidermal Moisture Measurements Versus Ultrasound and Visual Skin Assessments for Early Detection of Pressure Injury

Pressure ulcers (PUs) are detected by visual skin assessment (VSA). Evidence suggests ultrasound (US) and subepidermal moisture (SEM) scanner technology can measure tissue damage before it is visible. Purpose: A pilot study was conducted to evaluate consistency between SEM and US examinations of suspected deep tissue injury (sDTI).Method: Using an observational, prospective cohort study design, patients >55 years of age were recruited. VSA, SEM, and US assessments were performed daily for a minimum of 3 and maximum of 10 consecutive days following enrollment. US results were considered indicative of sDTI if hypoechoic lesions were present. SEM readings were considered abnormal when ∆ ≥0.6 was noted for at least 2 consecutive days. Boolean analysis was utilized to systematically determine consistency between US and SEM where sDTI was the clinical judgment. Results: Among the 15 participants (10 women, mean age 74 ± 10.9 years), there was consistent agreement between SEM and US when sDTIs existed. For 1 patient who developed a heel sDTI during the study, SEM readings were abnormal 2 days before VSA indicated tissue damage and 3 days before the appearance of a hypoechoic lesion in the US. Conclusion: US and SEM results were similar, and in an evolving sDTI case, SEM detected a lesion earlier than US.

One of the greatest gaps in preventing pressure ulcers (PUs) is, in the authors’ opinion, the available and implemented technology. In nearly every field of modern medicine, clinician assessments are routinely supported by basic technological aids that effectively screen and diagnose (eg, electrocardiography and blood pressure measurements in cardiology). However, in wound prevention and care, clinicians traditionally depend mostly on visual assessment, because potentially helpful, high-end technologies such as bedside imaging examinations or immunoassays are not yet part of the regular diagnostic protocol. All currently used risk assessment tools include a visual skin assessment (VSA) or are conducted in conjunction with a VSA. Moreover, VSAs are conducted as part of the usual care if a patient has been determined to be at risk for PUs. 

In the last 2 decades, PU research has made considerable progress in understanding wound etiology and specifically revealed that PUs may develop internally under intact skin.¹ Clinical practice has evolved accordingly, redefining PU classifications and adding suspected deep tissue injury (sDTI) to American and international classification systems.¹ 

This new understanding and the global consensus reflected in the current literature that damage may occur in deep tissues and progress toward more superficial layers until eventually presenting on the skin create a difficult situation. In the case of an existing (or a progressing) sDTI, assuming clinical signs such as the typical red, maroon, or purple local discoloration of the skin (as per the definition of a sDTI) are in evidence, a clinician performing a VSA will document existing tissue death. In other words, even successful detection of these clinical signs via routine VSAs will, by definition (of a DTI), document existing subdermal tissue damage rather than prevent it. 

This dilemma originates from the lack of technology to effectively and cost-beneficially detect cell and tissue damage under intact skin in clinical practice. Much the way cardiologists need more than the naked eye to detect a cardiovascular disorder, nurses and other wound care professionals deserve and require bioengineering technologies to examine tissue viability under the skin. Therefore, the field warrants technological breakthroughs that will provide information on pathophysiological phenomena that occur in deep tissue and cannot be detected visually and timely. 

Since the identification of the inside-out damage evolution pathway and the subsequent inclusion of DTIs in international PU classification systems, ultrasound (US) has been the modality of choice to identify subdermal pathoanatomical changes that may point to tissue damage in PU care, as reported in multiple case series.^2-6 For example, in the case series by Aoi et al,² 12 patients (ages 16 to 92 years) who showed DTI-related abnormal findings on ultrasonography at the first examination, were analyzed and followed-up until the PU reached a final stage, which revealed the effectiveness of US in detecting and monitoring the injury. In a retrospective review of patient examinations, Higashino el al³ used US in conjunction with infrared thermography in 28 early-stage PUs (21 patients). The authors concluded heterogeneous hypoechoic area findings on the US assessment and high skin temperature pointed to an evolving DTI.³ In a prospective, descriptive pilot study, Scheiner et al⁵ studied a hospital convenience sample of 33 individuals determined by means of Braden score (<18) to be at risk for PUs; all the DTIs that later became purple skin DTIs were able to be detected using US before clinical signs were visible on the skin. 

No formal validity, reliability, or sensitivity studies have been conducted to assess US as a diagnostic tool for PUs. Despite that, the literature suggests US technology can help detect tissue integrity issues that are not apparent in a VSA.⁴ Furthermore, US technology is becoming less costly, is often available at the bedside, and recently has been miniaturized to operate on portable devices. For example, the Lumify system (Philips Medical Co, Amsterdam, the Netherlands) utilizes a tablet or cellular phone as its computer platform; a clinician connects a US probe and downloads the appropriate app in order to perform US scans. Nevertheless, interpretation of US images still requires substantial training and expertise and is not quantitative; hence, data may be subject to interrater differences. Because it is not feasible to train all nurses and other relevant health care professionals to read and decipher US data, a fundamentally different, clinically feasible approach is required for cost-effective subdermal tissue viability evaluations. Moreover, any pathology that is identifiable by means of US is, by definition, already at the stage of macroscopic damage and has already affected tissue structures. In other words, US is unable to detect damage while it is still microscopic and limited to small groups of cells, at which time the damage could still be repairable by the body systems and be fully or partially reversible, pending a timely and adequate intervention. 

An emerging technology that appears to successfully bridge the above gaps is the subepidermal moisture (SEM) scanner. The SEM scanner is a hand-held device that measures capacitance of tissues at a depth of several millimeters under the skin (depending upon the specific anatomical site, version of the device, and examination protocol).⁷ Briefly, tissue capacitance rises when the extracellular water content (called SEM) increases, because a localized inflammatory response is triggered when the first cells in a tissue die. The SEM scanner determines tissue health status at the subepidermal layers (per manufacturer guidelines, the skin needs to be dry and clean to eliminate the influence of perspiration or incontinence). Specifically, the pathophysiological mechanism for the increase in extracellular water content is activated by the death of these first cells, which triggers recruitment of immune system cells from the bloodstream through release of signaling molecules in order to dispose the cell debris.⁸ These signaling molecules also cause blood vessel walls to be more permeable, enabling immune cells in the blood to cross the walls (a process called extravasation) and reach the site of cell death.⁸ The elevated vessel wall permeability then causes leakage of plasma fluids into the extracellular space, which eventually builds up to the clinically evident edema. Noteworthy is the gradual formation of edema and its initiation as a localized, microscopic event (ie, an increase in SEM). Accordingly, the SEM scanner targets this early phase of cell death rather than exploring macroscopic signs of tissue destruction as occurs in US examinations. 

The SEM scanner has been rigorously evaluated in large-scale clinical trials conducted primarily by the Bates-Jensen group^9-13 in nursing home settings and in a prospective, observational study that was performed in a spinal cord injury care facility and a residential care facility.¹⁴ The Bates-Jensen group conducted VSAs of heel PUs among 417 nursing home residents in 19 United States’ facilities over 16 weeks.¹⁰ The elderly study population was ethnically diverse; the authors found abnormal SEM readings were associated with concurrent damage and damage 1 week later in generalized multinomial logistic models adjusting for age, diabetes, and function.¹⁰ It also was found that elevated SEM values co-occurred with skin damage at the sacral region in generalized multinomial logistic models, adjusting for age and risk.¹¹ Greater SEM values were associated with visual skin damage 1 week later using similar logistic models.¹¹ The aforementioned work was conducted based on earlier clinical studies from the same research group and indicated the SEM scanner was able to differentiate between erythema and Stage 1 PUs in nursing home residents (N = 31 in 2 centers)¹² as well as in an elderly population with dark skin tones (N = 66 in 4 centers).¹³ Additionally, the ability of the SEM scanner to differentiate between Stage 3/Stage 4 PU tissues and intact skin was demonstrated in a pilot study among persons with spinal cord injury.¹⁵ 

Overall, studies conducted to date show the SEM scanner to be effective in detecting DTIs before they present on the skin and differentiating those that resolve, remain, or deteriorate, as well as predicting the occurrence of visual skin damage approximately 1 week later.^9-13 However, the SEM scanner has never been directly compared with US and VSAs in the same cohort. 

Given the fundamentally different characteristics examined by US versus the SEM scanner, the current authors hypothesized that the SEM scanner would be able to detect damage earlier than US. Earlier detection of cell and tissue damage sets the stage for prevention of further tissue damage (either more widespread or deeper). The ability to detect tissue damage earlier using the SEM scanner, possibly before damage becomes irreversible, potentially alerts caregivers to prevent deterioration.

Accordingly, the aim of this pilot work was to determine, through an observational, prospective cohort study conducted among elderly persons in a postacute care setting, whether SEM scanner measurements are consistent with US examinations in assessing subdermal tissue damage, and, if possible, further test the hypothesis that SEM scanner measurements not only precede VSAs in alerting caregivers to the onset of PUs and sDTIs, as previously reported,^9-14 but also that SEM may precede US in doing so.

Patient recruitment. All patients at Kindred Healthcare (Virginia Beach, VA), a postacute care center with a high proportion of elderly patients, were recruited through an observational, prospective cohort study design, as follows. A nurse practitioner (NP) and the principal investigator, a board-certified physical medicine and rehabilitation physician, apprised the nursing staff of the study and helped identify potential participants (ie, participants who had a PU or sDTI). The physician met with each potential participant (and typically with a family member) and reviewed the study protocol. If the patient was interested in participating, informed consent was obtained. The patient was enrolled only after informed consent was obtained and he/she was deemed eligible to participate in the study. 

Study groups. Patients were recruited from December 2016 through February 2017 into 1 of 4 groups (inclusion/exclusion criteria varied per group): group 1 included patients at risk of developing PUs/sDTIs but with no evidence of these wounds at the time of recruitment; group 2 included patients with an existing, diagnosed sacral Stage 1 PU; group 3 included patients with sDTI; and group 4 included patients without wounds who were not at risk. Inclusion criteria were age >55 years and, for groups 1, 2 and 3, risk of developing PUs at the time of enrollment (as indicated by Braden score <13), poor mobility (Braden mobility and activity subscore ≤2 or clinically observed limited movement such as chair- or bedbound), or medical procedures involving inability to change position for >4 hours due to, for example, surgery or imaging exam. Group 4 (the control group) consisted of age-matched individuals who were not at risk for PUs as indicated by their Braden score. In the 2 groups of patients that did not have PUs (group 1 and group 4), recruitment was based on risk factors or lack thereof and a willingness to participate in the study. All participants were evaluated by the study team for a minimum of 3 and a maximum of 10 consecutive days, and all underwent the SEM and US scanning procedures (described as follows) daily. Patients with modesty concerns and physical, mental, or other limitations preventing performance or compliance with the US and/or SEM scanning assessments such as suspected or actual injury that might prevent turning, as well patients with active inflammatory skin diseases (eg, eczema, psoriasis), moisture lesions, incontinence-associated dermatitis at the sacrum, skin infection, open wounds or scarring over the scan areas, or broken skin at the sacrum and both heels (SEM scanner guidelines forbid use on broken skin) were excluded. However, possible assessment at only 1 or 2 locations was not grounds for exclusion. 

Ethical considerations. The study was approved by a medical-ethical review committee (Quorum IRB #QR32052/1) and informed consent was obtained from all participants. Classification of PUs into stages/categories was done according to the up-to-date International Pressure Ulcer Prevention and Treatment Guidelines 2014.¹ 

Procedure. Daily skin assessments were conducted by the NP from the day of enrollment until study exit. The NP conducted the VSAs to obtain data on tissue characteristics, the presence or absence of a PU, and characteristics of the PU if present. As per best practice, the NP examined the patient for changes in skin color (particularly redness/erythema), dry/flaky skin, bruising, callous, scab, cut/abrasion, cyst, rash, scar, texture change, and sensation of the skin surface through visual observation (see Table 1). In individuals with darkly pigmented skin, the NP observed and documented for persistent erythema, nonblanching hyperemia, blisters, discoloration (purple/blue localized areas), localized heat (replaced by coolness as tissue is damaged), localized edema, and localized induration. The physician took the SEM readings (photographs also were taken at the time when SEM readings were performed). In addition, the physician reviewed the inclusion/exclusion criteria and collected the demographic information and the relevant medical history data. The variables collected in the study and the respective assessment days are listed in Table 1. An US radiology specialist assessed the USs. 

The SEM readings began on the day of enrollment and on each subsequent day until exit from the study. Every day, 6 readings were obtained at the sacrum and 4 at the heels according to the study protocol. The US readings were performed by the US specialist on the same day but at a different time. The readings were obtained per the protocol; images were saved to the US machine as well as to an external data drive. 

The NP who conducted the VSAs and the physician were wound specialists, proficient in the up-to-date international guidelines, as specified in Table 1.¹ Group 1 (at-risk) patients were assessed once daily by the clinical study team for a minimum of 3 and maximum of 10 (consecutive) days. These clinicians then provided the standard care preventions (listed in Table 1) and recorded any interventions and notable observations of patient health status (including response to interventions). Group 2 and group 3 patients with confirmed PUs or sDTIs were similarly assessed by specialists daily and had daily photographs of the (suspected) damage sites taken; deterioration or healing of their wounds was digitally recorded. Group 4 (control) patients also underwent daily US and SEM assessments of their sacrum and heels, similar to the experimental groups, for 3 days to verify that all diagnostic means (SEM, US, VSA) consistently provided benign results. Following informed consent, the physician reviewed the health status of each patient and the NP reviewed the Braden score. 

Scanning instrumentation. All participants were assessed using a portable US system (M7; Mindray, Mahwah, NJ) that featured a 6–14 MHz linear array transducer allowing depth of field of 2 cm to 39 cm, as well as by using the SEM Scanner (Point of Care 200; Bruin Biometrics LLC, Los Angeles, CA). 

The SEM scanner directly measures the steady state capacitance (in Farads) of a volume of tissue. The sensor is fundamentally a 2-electrode capacitor that has been unwrapped so that the electrodes are coplanar. The electric field created between the electrodes projects into the tissue when the electrodes are placed in contact with a patient’s skin. The shape of the field is constant and based on the geometry of the electrodes^.16 With the excitation voltage used in the SEM scanner, the sensor is sensitive to a depth of approximately 0.15 inch (3.8 mm). The capacitance of this unwrapped capacitor is dependent upon the effective dielectric constant of the tissue within the field. Water has a high dielectric constant of 80 compared to a dielectric constant of 4 for the dry collagen, which is the major structural component of the extracellular matrix (ECM).¹⁷ As localized edema builds up due to the pathophysiology of inflammation, the effective dielectric constant of the tissue region affected by the PU rises toward that of water. Although the mixing rule for the effective dielectric constant for a 2-phase composite, such as a biological tissue, cannot be analytically predicted,¹⁸ testing with the SEM scanner indicates the effective capacitance is linearly related to the percentage of water in the tissue. Assuming, as an example, the ECM normally has a volumetric collagen content of 50% and the remaining content is water, the dielectric constant of normal ECM is (0.5 x 80) + (0.5 x 4) = 42. A change in the water content of the ECM from 50% to 55% in a small subsurface region of tissue may not be visually detectable on the skin surface, but it will change the effective dielectric constant of that region of tissue to (0.55 x 80) + (0.45 x 4) ≅ 46, a 9% increase that can be detected by the SEM scanner. 

The SEM scanner is CE-marked and pending United States Food and Drug Administration decision and not available for sale in that country. Participants were examined using US and the SEM scanner at the same anatomical locations (ie, the sacrum and left and right heels according to the study design as detailed above), with follow-up periods detailed in Table 1. 

Outcome measures. Generally, heterogeneous hypoechoic area findings on the US assessment are an indication for subdermal tissue damage.^2-5 An abnormal US exam exhibits inconsistent tissue structures (unclear layered structure, discontinuous fascia) or hypoechoic lesions. Hypoechoic lesions, in particular, were considered to be a typical finding for sDTI that may indicate localized tissue decomposition, as reported by Aoi et al.² 

An abnormal SEM reading was defined as SEM delta (Δ) ≥0.6 for at least 2 consecutive days (see Table 1) according to manufacturer guidelines. The SEM Δ, the parameter of primary clinical interest, is the difference in the SEM values between nonwounded tissue and a nearby region of tissue that may have subsurface damage that affected local tissue fluid content (ie, the inflammatory response had triggered microscale edema). For example, a set of 4 to 6 measurements may be taken at the sacral or heel regions. It is almost certain that at least 1 of these measurements at a certain anatomical site will be over healthy (undamaged) tissue. The SEM scanner compares the set of SEM values to each other and calculates the Δ value as the difference between the highest SEM value and the lowest SEM value in the set. A healthy body site will have fairly uniform distribution of SEM values, because all measurement points will reflect normal tissue resulting a low Δ value (typically 0.0 to 0.2) for that area. Larger Δ values are an indication of potential subsurface tissue damage in the examined area, reflected in nonuniform tissue fluid contents due to the localized build-up of edema (this threshold has been set as Δ ≥0.6 in the present study based on manufacturer guidelines). 

In this study, abnormal US and visual/photographic findings were compared to SEM Δ values to identify markers indicating PUs/DTIs and either declining or improving tissue status (also considering clinical judgment and outcome). This was done to reveal the value of SEM measurements in determining the risk for subdermal damage in PUs and in early detection of DTIs that later become apparent by means of US and/or VSA. Given the older age and fragility of the studied patient sample and the variety of underlying background conditions, comorbidities, and possible history of PUs, the focus in this study was clinically significant, consistent tissue pathologies. Hence, a lesion in skin or subdermal tissue was defined as an abnormal finding detected through VSA (as per the categories in Table 1) and confirmed by a hypoechoic US finding as listed above, which was detectable over at least 2 consecutive days. Abnormal findings in all measurable outcomes are defined in Table 1. 

Data collection. All study variables (see Table 1) were initially noted using paper and pencil and later entered to Excel spreadsheets (Microsoft Co, Seattle, WA) while maintaining patient anonymity. The heel and sacrum data were reported separately and entered into the case report form, which was available to the skin assessment NP until the end of the study. Importantly, the clinical principal investigator performing the SEM scanner readings was blinded from the US and VSA results of each participant until patients exited from the study. 

Data analysis. All quantitative results were reported as the mean ± 1 standard deviation (SD) from the mean. Boolean analysis and rule-based classifiers programmed in spreadsheets were utilized to systematically determine if a lesion (as defined above) was consistent with abnormal SEM readings across the diagnosed PU/DTI cases in the study. False positives were not considered because the focus of the pilot study was different and was not statistically powered for sensitivity and specificity, nor was it powered for validity and reliability. Rather, the focus was on consistency between SEM and US where there were apparent clinical signs of (suspected) tissue damage by means of VSA.

Accordingly, for the 3 patients recruited to group 3 who had sDTIs according to VSAs (patients 11, 12, and 13 [see Table 2]), examination of the subdermal tissue status was justified; hence, Boolean analysis and rule-based classifiers were used to determine agreement between US and SEM findings (see Table 3). The Boolean analysis is considered positive per patient, anatomical site, and visit day according to the following rule-based classifiers: 1) SEM Δ readings were normal (below 0.6) and US examination did not identify a hypoechoic lesion, or 2) SEM Δ readings were abnormal (0.6 or above) and US identified a hypoechoic lesion.

The demographics of the recruited patients are specified in Table 2. Consent was obtained from a convenience sample of 15 patients (10 women, 5 men, mean age 74 ± 10.9 years old). Eleven (11) were white/Caucasians with Fitzpatrick classification¹⁹ for human skin color of II–III; the other 4 were black/African Americans with Fitzpatrick scale of III and above. Patients from groups 2 and 3 in Table 2 where PUs or (suspected) DTIs existed were pooled for similarity in characteristics; the number of patients assigned to the different groups were 7, 6, and 2 for groups 1, 2-3, and 4, respectively. A variety of background and acute conditions typical to the facility setting were documented in the medical records of patients, including respiratory (eg, bronchitis, chronic obstructive pulmonary disease) and cardiovascular (eg, pulmonary edema, deep vein thrombosis, and peripheral vascular) conditions, type 2 diabetes, sepsis, and renal failure, all of which may be associated with known PU etiological and risk factors. None of the patients had undergone surgery during the study period. One (1) patient was incontinent for urine, and 2 had a history of sacral PUs. Mean Braden risk assessment scores for the pooled participant groups (2-3), excluding the controls, were 11.4 ± 2.8 (range 10–16), indicating an overall high risk for the study sample, with 4 individuals at very high risk (score ≤9). The at-risk group (group 1) had an even lower mean Braden risk score of 9.7 ± 2, pointing to a high to very high risk for PUs for this specific group. In every case, the daily score obtained by the NP was the same score the patient had in the chart previously recorded by the nursing staff (before the initiation of the study).

All at-risk patients received the usual care for PU prevention, including routine VSAs, repositioning every 2 hours, and use of repositioning aids such as pillows, wedges, and heel boots according to clinical judgment. One (1) patient from group 1 was prescribed a low-air-loss mattress; all others were positioned on standard hospital mattresses. All participants completed the study as per the protocol without adverse events. Patients were assessed for 7 ± 4 consecutive days of follow-up (across all groups). 

All participants in the at-risk group (1) exhibited elevated (Δ ≥0.6) SEM readings in at least 1 anatomical location for 2 or more consecutive days. All patients in group 1 had intact skin at all their examined anatomical locations over the study period, but 1 patient in the group developed a sDTI during the course of the study.

Four (4) of the 15 study participants (1 in the at-risk group and 3 in the confirmed/suspected PUs groups 2 and 3) already had or developed a visible sDTI within the timeframe of the study (ie, a period prevalence of 20%). All sDTI lesions developed at the heels. 

For the patient from at-risk group 1 (patient 2; see Table 2) who developed a left heel sDTI (described in the record as “soft mushy heel”) 3 days after the start of the study, SEM readings were abnormal from the first visit day onward. The VSA indicated sDTI on visit day 3. The US indicated mild structural inconsistencies (unclear, layered structures and discontinuous fascia) on the first day, but an additional hypoechoic lesion, which is a typical finding in a sDTI,² was formed and detected by the US on visit day 4. On visit day 4, the abnormal SEM Δ reading was in agreement with both the US and VSA assessments that indicated a tissue lesion. Interestingly, the sDTI appeared to resolve according to VSAs during the follow-up in visit days 9 and 10. This was likely aided by the early intervention to offload the left heel using a pillow/wedge, which was conducted immediately when the respective SEM reading was abnormal on the first day while the VSA was still normal. The abnormally elevated SEM  readings and US finding of hypoechoic lesion remained until the end of the 10-day follow-up. Thus, the initially high and increased SEM Δ reading from 1.3 to 1.7 between the first and second days, leading to the rapid, preventative intervention (heel elevation), appeared 2 days before the hypoechoic lesion US finding. The timeline of the above events for patient 2 is depicted in the Figure. 

In patients with existing DTI, an important agreement was observed between US and the SEM Δ reading in 5 different potential wounds appearing for at least 2 consecutive days. Focusing on the 3 patients comprising group 3 (patients 11, 12, and 13; see Table 2) and using a Boolean analysis and rule-based classifiers found SEM measurements always agreed with the US (see Table 3); agreement was observed both in positive and negative identification. In addition, although there always was agreement between SEM and US readings, they did not always agree with VSAs (eg, for both heels of patient 12; see Table 3), which points to the limited ability of VSAs to reveal the true status of subdermal tissues. 

The Boolean analysis and rule-based classifiers (see Table 3) indicated that if a subdermal lesion existed (based on US), SEM measurements always agreed with the US-based identification of hypoechoic lesion (likely pointing to subdermal tissue composition).² In addition, for the single patient who developed a heel DTI during the study period, the SEM measurements were predictive of tissue damage 2 days before damage was detected via VSA and 3 days before the appearance of a hypoechoic lesion in the US (see Figure). A review of the outcomes of the Bates-Jensen series of large-scale studies in nursing home residents^9-13 concluded abnormal SEM readings may precede positive VSA findings by 3 to 10 days,²⁰ which concurs with the current observation. 

US has been evaluated against other technologies such as infrared thermography and has been identified as a successful means for early detection of PUs in multiple case series/retrospective studies. A systematic review²¹ suggests US and SEM measurements are promising in the early detection and prediction of PUs, but more studies are needed. Nevertheless, a fundamental difference exists between US and SEM, which is critically important to emphasize in the context of the present study: US detects macroscopic pockets of fluids (edema) that are visible to the radiologist, presented as hypoechoic lesions (as in the example of patient 2), whereas the SEM scanner has the sensitivity to detect the occurrence of edema while it is still microscopic and invisible to the eye. The present study is the first in the literature to directly compare US and SEM measurements within the same participant group in a clinical setting. Consistent with the above argument that SEM should detect PUs earlier, the current researchers found the SEM reading for the 1 patient who developed a DTI during the course of this study increased from the first to the second visit days, which was 2 days before the US exam clearly indicated tissue pathology through a hypoechoic lesion (macroscopic fluid pocket) finding. Although the authors did not conduct infrared thermography scans in this research, it is noteworthy that infrared thermography is limited to near-surface measurements of skin temperature and hence is unlikely to detect deep tissue damage through localized temperature changes.²¹ Accordingly, currently, it is most appropriate to compare the SEM scanner to US technology. 

Interpreting US images requires expertise and a months of training.²² According to a study analyzing 5 consecutive US courses with a total of 363 participants²³ and a systematic review,²⁴ the teacher-to-learner ratio should be 1:3 during hands-on, multiple-day US training in order to facilitate training the skills needed for even simple image acquisition. Additionally, there is the potential for interobserver disagreements in interpreting US results (which are neither objective nor quantitative); hence, sensitivity and specificity may vary among studies.²² 

Clendenin et al²⁵ conducted a study that focused on the interrater and interdevice agreement and reliability of the SEM scanner in a group of 31 volunteers who were free of PUs or broken skin at the sternum, sacrum, and heels. The authors analyzed more than 3000 SEM scanner readings and reported good interoperator agreement with mean differences ranging from -0.01 to 0.11. These researchers further reported interoperator and interdevice reliability (Pearson product correlation coefficient) exceeding 0.80 at all the above anatomical sites. Bates-Jensen et al¹¹ reported an even greater overall mean SEM reliability in their large cohort study of nursing home residents (N = 417), which was 0.92 for all skin conditions and locations. However, reliability by specific skin conditions was slightly lower and varied across conditions (normal, erythema, Stage 1 PU).¹¹ Both research groups concluded the SEM scanner is a reliable tool for assessing the presence or absence of PUs and that it shows great promise for clinical use. Overall, the aforementioned studies^11,25 demonstrate high reliability and good agreement of the SEM scanner across different operators and devices. Given the limitations of current methods to prevent and detect PUs, the above studies indicate the SEM scanner shows promise as an objective, quantitative, and reliable tool for assessing the presence of subdermal tissue damage. 

The SEM scanner has the potential to be a useful adjunct to clinical experience and judgment in early PU detection, particularly for wounds that develop under intact skin. In real life, risk assessment and early detection should be integrated on the continuum of care. Microscale tissue damage that is not yet clinically significant and can be, at a certain point in time, fully repairable and reversible by the body systems, may evolve into a macroscopic clinically significant tissue damage if the proper interventions or actions are not or cannot be taken. Hence, when damage is still microscopic — that is, if some cells in the tissue have died but the body may be able to replace them — according to clinical experience, a patient is said to be at risk (for the damage to deteriorate).²⁶ If damage then proceeds to accumulate and affect larger numbers of cells, the patient will have initial tissue damage (possibly under intact skin) that requires an early detection tool, and still the tissue damage can be self-repaired by the body if actions are adequate and timely. Hence, continuity exists between the state of being at risk and the state of being affected by early cell and tissue damage, which requires early detection. The SEM scanner has clinical potential because it detects the early phase of inflammation and can help the clinician modulate the response of the immune system and adjacent vasculature to the event of death of the first cells.²⁶ 

In order to describe the potential of the SEM scanner in early detection applications, it is necessary to understand its underlying physical mode of action and how that relates to the pathophysiology of PUs. Thus, to reiterate: the SEM scanner measures the localized edema (ie, subepidermal moisture) that precedes the macroscopic edema that involves swelling and increased firmness of the affected tissues.²⁶ At the early phase of PU damage when the damage is still microscopic and limited to small numbers of cells, the damaged or dying cells release chemokines (signaling/messenger molecules also called chemoattractant molecules) that act as inflammatory signals and attract immune system cells (eg, neutrophils, macrophages, and T-cells) to the affected site.^8,22 The release of chemokines also acts to increase the permeability of capillaries, which enables extravasation (ie, infiltration of the immune system cells to the damaged region).^8,26 A key consequence of this inflammatory process is that as gaps between endothelial cells increase to facilitate the extravasation, the walls of capillaries become leaky and fluids leave the vessels and build up at the extracellular space.²⁶ Eventually, the volume of blood plasma fluids escaping from the vasculature will cause visible tissue swelling, but one should recall that this process begins microscopically and progresses over time as the immune system is recruited to deliver a sufficient number of immune cells to the damaged site.²⁶ The practical implication for diagnosis of the beginning of formation of an injury is that there will be a small increase in extracellular fluid contents very soon after the death of the first few cells.²⁵ A device that is sensitive enough to capture such slight changes in water content within soft tissues will be able to detect these cell death events. The physical principles underlying the device, its specific technology, and the sensitivity parameters derived from its engineering design together will dictate the rapidness of early detection. 

The physical theory of electrical biocapacitance of tissues²⁶ points to the prospect of achieving sensitivity to small changes in water content by means of a handheld cost-effective device. The SEM scanner utilizes the principles of biocapacitance^16,27 to detect small changes in water content of examined tissue. The scanner directly measures the steady-state capacitance (in Farads) of a volume of tissue that is associated with the local fluid contents in the tissue. The ability to detect small, local changes in the water content of ECMs caused by the early stages of edema makes biocapacitance measurement a powerful tool in the detection of early tissue damage that is a precursor to PUs. 

This is a small cohort pilot study, which limits its external validity and the ability to draw firm conclusions. Other limitations of this work are associated with the general lack of gold standard technology in the clinical practice of prevention and early diagnosis of PUs. In that context, these are still early days for this technology. SEM thresholds for detecting an invisible tissue lesion need to be tested, optimized, adjusted, and fine-tuned. Specifically, with reference to DTIs, the pathophysiology is dynamic and so are the fluid contents in the affected tissues, which change from the stage of initial cell damage and the resulting localized edema to the state of massive tissue necrosis. For example, evidence that tissue stiffness changes along this time course²⁸ makes it likely that the SEM readings and Δs will change over the evolution of a DTI as well. It is also presently unknown whether chronic conditions such as peripheral vascular disease, diabetes, or cancer (which may locally affect the pattern of edema or leakiness of the vasculature) also influence SEM Δ values. Likewise, conditions that affect the inflammatory response such as obesity, spinal cord injury, cancer, and human immunodeficiency virus may influence the SEM Δ values, so related subgroups should be studied for their SEM Δ responses. 

Using an observational, prospective cohort study design, this pilot study evaluated consistency between SEM and US examinations of PUs and sDTIs. Among the 15 participants, where lesions existed, SEM measurements always agreed with the US and VSA findings. For the 1 patient who was determined to be at risk at the beginning of the study and developed a heel sDTI during the course of the study, SEM readings were abnormal 2 days before VSA indicated tissue damage and 3 days before the appearance of a hypoechoic lesion in the US. Hence, the authors found US and SEM scanner results were similar but in the evolving sDTI case, the SEM scanner detected it earlier. Moreover, although SEM and US readings always agreed between themselves, they did not always agree with VSAs, which points to the limited capacity of VSAs to assess the status of subdermal (nonvisible) tissues. Further clinical, prospectively designed research is needed in other at-risk cohorts to evaluate the potential effect of other health conditions on the validity, reliability, sensitivity, and specificity of the SEM scanner.

Potential Conflicts of Interest: This work was supported by an unrestricted educational grant from Bruin Biometrics LLC, Los Angeles, CA. 

Dr. Gefen is a Professor of Biomedical Engineering, Department of Biomedical Engineering, Faculty of Engineering, Tel Aviv University, Tel Aviv, Israel. Dr. Gershon is the Medical Director, Gershon Pain Specialists, LLC, Virginia Beach, VA.

Please address correspondence to: Prof. Amit Gefen, Department of Biomedical Engineering, Faculty of Engineering, Tel Aviv University, Tel Aviv 69978, Israel; email: gefen@eng.tau.ac.il.