To evaluate the construct validity of 2 proposed measures (the Ultrasound Disease Activity [U-DA] and the Tissue Thickness Score [TTS]) for evaluating sonographic differences in juvenile localized scleroderma skin lesions.
To evaluate the construct validity of 2 proposed measures (the Ultrasound Disease Activity [U-DA] and the Tissue Thickness Score [TTS]) for evaluating sonographic differences in juvenile localized scleroderma skin lesions.
We conducted a retrospective review of juvenile localized scleroderma patients who had ultrasound scans of their skin lesions between October 2005 and February 2009. Imaged lesions were classified as active or inactive based upon clinical assessment. Lesions had to have been imaged within 1 month of a clinic visit or have the same clinical assessment during both the visit before and the visit after the scan. Two physicians scored the scans using the U-DA, which scores for differences in lesion echogenicity and vascularity compared with normal tissue. Tissue thickness differences were evaluated by percent differences and by using the TTS. Wilcoxon's rank sum test was performed to assess differences.
We studied 52 scans from 21 patients, 32 scans of active skin lesions and 20 scans of inactive skin lesions. Features reported by clinicians as indicative of active disease included erythema, warmth, violaceous color, new lesion, expansion of lesion, and induration. The U-DA was significantly different between active and inactive skin lesions (P = 0.0010) with significant differences found for the parameters of total echogenicity, hypodermis echogenicity, and deep tissue layer vascularity (P = 0.0014, P = 0.0023, and P = 0.0374, respectively). No significant differences were found for tissue layer thickness or TTS.
The U-DA may be a useful tool in the identification of localized scleroderma activity. Further study is needed to prospectively evaluate the validity, reliability, and sensitivity of this potential monitoring tool.
Localized scleroderma is the most common form of scleroderma in the pediatric population and can present as several different subtypes, the most common being linear scleroderma and circumscribed morphea (plaque lesions) (1). Children commonly have disease that affects deep tissue layers and are therefore at risk of major morbidity, including limb-length discrepancy, joint contractures, and neurologic problems (2, 3). Optimal therapy is not known, with treatment ranging from topical agents to phototherapy to systemic immunosuppressive medications (1, 4). Evaluation of treatment efficacy has been hampered by the paucity of sensitive disease assessment measures (2, 5). Clinical assessment of activity is often limited, based on observations that may fade early (i.e., erythema), or on findings that signify disease spread (i.e., extension of existing lesion or development of a new lesion). Moreover, activity in deeper tissues may progress before clinical signs become apparent (6).
Ultrasound is a readily available and noninvasive imaging modality that has great potential to aid monitoring of localized scleroderma disease state since it allows for the evaluation of both superficial and deep soft tissues. Several groups have used ultrasound to monitor adult localized scleroderma. Initially, adult skin lesions show dermal thickening and hypoechogenicity; with treatment, the dermis thins and the hypoechogenicity decreases (7–9). Decreased echogenicity has been shown by combined histology and sonographic studies to represent edema, an early finding in active lesions (10). Hyperechogenicity and abnormal patterns of echogenicity have also been seen in the dermis, hypodermis, and deep tissue layers of localized scleroderma lesions. Studies of the active indurative phase of scleroderma have shown the presence of inflammatory cells and collagen (11, 12), both of which have been found in combined histologic and sonographic studies to be associated with increased echogenicity (10, 13). Normalization of dermal echogenicity correlates with clinical improvement (7, 8); the same may be true for the deeper tissue layers (14, 15). In some patients, increased vascularity as detected by color Doppler may be another sign of disease activity (6, 15). Increased blood flow has been found in active localized scleroderma skin lesions by laser Doppler flowmetry and is implied by the increased temperature found in active lesions by thermography (16, 17).
As part of the multicenter multidisciplinary Localized Scleroderma Clinical and Ultrasound Study Group (LOCUS), we have proposed a preliminary Ultrasound Disease Activity measure (U-DA) (18). The U-DA evaluates echogenicity and vascularity differences in each tissue layer of the lesion compared with the corresponding normal tissue layer. Both decreases and increases in these parameters are scored, with 0 representing no difference compared with the normal site. A major intention of the U-DA is to facilitate standardization of localized scleroderma sonographic interpretation; scoring levels were set based upon the observed range of differences. A measure for scoring tissue layer thickness differences (Tissue Thickness Score [TTS]) was also proposed (18). We now report on our initial evaluation of the ability of these measures to identify disease activity in a set of previously collected patient scans. Skin lesions were categorized as “active” or “inactive” based on the treating physician's clinical judgment, as is the current standard of care. Although clinical judgment can be flawed, physician assessment of disease state has commonly been used to evaluate the validity of new rheumatic measures, such as the Localized Scleroderma Severity Index, a measure for evaluating severity features of skin lesions and their extent (19), and multiple systemic lupus erythematosus measures (Systemic Lupus Erythematosus Disease Activity Index, British Isles Lupus Assessment Group, Systemic Lupus Activity Measure, and others) (20). Significant differences were found between clinically active and inactive skin lesions for mean total U-DA score, but not for TTS. Our results suggest that the U-DA may be a useful tool in standardizing measurements of disease activity in children. Moreover, it may reveal subclinical activity, particularly in deeper tissues.
We performed a retrospective analysis of all pediatric localized scleroderma patients who were followed by the Pediatric Rheumatology division at Hackensack University Medical Center (HUMC) and had ultrasound scans of their localized scleroderma skin lesions acquired between October 2005 and February 2009. Patients were followed by 1 of 3 pediatric rheumatologists (SCL, KAH, and JEW), all of whom had 6–24 years of postfellowship experience in clinical practice. Each pediatric rheumatologist was asked to categorize the patient's imaged skin lesions as active or inactive based upon a chart review of clinical features present at the time of the scan. Since there are no published consensus criteria for defining active localized scleroderma skin lesions, we did not specify any criteria or definitions. Instead, clinicians were asked to specify which features present in the active study lesions indicated activity to them.
To ensure that clinical assessment accurately reflected disease state at the time of the scan, scans had to have been acquired within 1 month of a clinic visit, or the lesion had to have the same clinical assessment in both the visit before and the visit after the scan. Only 1 ultrasound imaging time point was studied for each patient; the first imaging date that met the above criteria was used.
Each patient was assigned a study identification number. These numbers, together with the anatomic location of imaged skin lesions, were used to uniquely identify each scan. Charts were reviewed to collect demographics, clinical features, and treatment history. We followed the Padua Preliminary Classification for classifying lesion subtype (1). Data were de-identified and entered into Excel spreadsheets (Microsoft). This study was approved by the HUMC Institutional Review Board, and is part of a Childhood Arthritis and Rheumatology Research Alliance– approved localized scleroderma project.
All scans were acquired by 1 radiologist (MSL) using an Acuson Sequoia 512 machine, frequencies 8–14 MHz. To determine the numerical value of the U-DA, skin lesion images were evaluated in comparison with the control (normal) site, which was usually the contralateral side. Both the lesion and control site images were simultaneously displayed as a dual image during acquisition and scoring (18).
Based upon lesion appearance and location, the clinician and radiologist decided which skin lesions to scan; some sites, such as those on the scalp or near the eye, could not be readily imaged. To evaluate very large skin lesions, we divided them into separate imaging sections following the anatomic sites specified for the modified Rodnan skin thickness score (21). For example, a linear scleroderma lesion of the entire upper extremity could be imaged in the hand, forearm, and upper arm. Different anatomic sites were treated as separate lesion scans for this study.
Two investigators (MSL and SCL) jointly scored all scans using the U-DA, which scores for the greatest level of difference in echogenicity and vascularity in each lesion tissue layer (dermis, hypodermis, and deep tissue layer [most commonly muscle or glandular tissue]), in comparison with the unaffected site (18). Echogenicity scores are determined from gray-scale images, vascularity scores from color Doppler images, and differences must be seen in at least 2 separate dual images.
Hyperechogenicity was more commonly seen than hypoechogenicity. No echogenicity difference between the lesion and normal tissue layer is scored as 0, while decreased echogenicity is scored as −1. Hyperechogenicity scoring ranges were set based upon the observed range of differences: dermis is scored 1, deep tissue layer from 1–2, and hypodermis from 1–3. A score of 1 means a mild increase, while a score of 3 means that lesion hypodermis echogenicity is greater than or equal to normal dermis echogenicity. Normally, the dermis is more echogenic than the hypodermis. We do not know if scores of −1 and +1 are equivalent; both could be associated with active lesions (i.e., edema versus inflammatory cells or early collagen deposition ). We gave them different signage to facilitate the tracking of the findings.
Vascularity is scored the same for all tissue layers, with −1 representing a decrease in lesion vascularity compared with the normal site, and scores of 1–3 representing progressive increases in level of hyperemia. Hyperemia was found more commonly than hypovascularity. A score of +1 represents up to 2-fold more color Doppler signals or 1 more larger-sized color Doppler signal in the lesion than the normal site. A score of 3 requires at least 4 more larger color Doppler signals, or that ≥20% more of the lesion layer has color Doppler signal. Further study is needed to evaluate if −1 and +1 are equivalent; it may be that hypovascularity indicates damage rather than activity. The absolute values of the individual parameters are summed to determine the U-DA score, which ranges from 0–15. Examples of the scoring levels can be found at http://www.ped-rheum.com/content/pdf/1546-0096-8-14.pdf (18).
Lesion tissue thickness differences were evaluated in comparison with normal skin tissue layer thickness by percent differences ([normal tissue layer thickness − lesion tissue thickness × 100]/normal tissue layer thickness) and TTS proposed by LOCUS (18). The TTS establishes scoring values for different percent differences in the tissue thickness of lesions compared with the normal layer. No difference is scored as 0, and an increase (thickening) is scored as −1. Thinning of the dermis is scored as +1, thinning of the deep tissue layer as +1 for <20% thinning, and as +2 for ≥20% thinning, and thinning of the hypodermis is scored from 1–3. A hypodermis score of 3 represents ≥90% thinning or the complete loss of subcutaneous fat (18). Signage was kept for individual scores.
Each skin lesion was considered an independent observation. Due to small sample size and lack of normal distribution, statistical analyses were carried out by Wilcoxon's rank sum tests. Wilcoxon's rank sum tests were performed on total U-DA, echogenicity, and vascularity scores for each tissue layer, tissue thickness percent differences, and TTS for active and inactive lesions. Significance tests are 2-sided normal approximation; P values less than 0.05 were considered significant. All analyses were completed using the SAS software, version 9.1.
We identified 21 localized scleroderma patients: 9 patients (A1–A9) with clinically active skin lesions, 4 patients (M10–M13) with both clinically active and inactive skin lesions, and 8 patients (I14–I21) with clinically inactive skin lesions at the time of their scan. There were a total of 32 (61.5%) ultrasound scans of active skin lesions and 20 (38.5%) scans of inactive skin lesions. All active skin lesion scans were acquired within 1 month of a clinic visit, with 25 scans acquired on the day of their clinic visit. Sixteen inactive scans were acquired within 1 month of a clinic visit, the other 4 between 31 and 44 days from a clinic visit.
Table 1 shows the study patients' demographics. There were no significant differences between active and inactive groups for sex, age at diagnosis, age at time of scan, localized scleroderma subtype, treatment, or disease duration. The disease activity state was specified by the clinician and based upon her evaluation of the appearance of the patient's lesion(s). In Table 1, inactive patients are those that only had inactive lesions at the time of the scan; the 4 patients that had both active and inactive lesions at the time of the scan (M10–M13) are listed in the active patient group. Most patients (n = 12) had linear scleroderma. At the time of their scan, 11 of the patients with active and 2 of the patients with inactive lesions had extracutaneous manifestations, most commonly musculoskeletal. Eight patients with active skin lesions had joint limitation (n = 6), arthralgia (n = 4), myalgia (n = 3), and/or arthritis (n = 2), and 9 patients had bone size differences, primarily of the extremities. Other extracutaneous manifestations in the active skin lesion group included headache (only neurologic manifestation), decreased diffusing capacity for carbon monoxide in 2 patients (1 of these also had pulmonary hypertension), stomach aches, eye pain (associated with headache), eyelash abnormalities (in patient with Parry-Romberg syndrome), and cold hands. Two of the patients with inactive skin lesions had musculoskeletal manifestations at the time of their scans, arthralgia in one and joint limitation and bony size difference in the other. Three patients with active skin lesions had laboratory abnormalities at the time of their scans (1 patient with elevated C-reactive protein level and 2 with elevated IgE). None of the patients had eosinophilia.
|Age at time of scan, mean ± SD (range) years||12.2 ± 3.7 (4.6–18)||12.0 ± 4.0 (4.6–16.80)||12.5 ± 3.5 (8.0–18)|
|Age at disease onset, mean ± SD (range) years||7.6 ± 4.0 (2.1–13.9)||7.8 ± 4.5 (2.5–13.9)||6.9 ± 3.6 (2.1–11.7)|
|Disease duration at time of scan, mean ± SD (range) years||4.6 ± 4.0 (0.42–13.5)||4.1 ± 4.3 (0.42–13.5)||5.6 ± 3.6 (0.67–12)|
|Localized scleroderma subtype|
|Linear of extremity/trunk||5||4||1|
|Linear of face/scalp||2||0||2|
|Treatment at time of scan|
|Bone size difference||9||1|
|ANA positive||5 of 12||2 of 6|
|RF positive||1 of 3||0 of 2|
|Elevated ESR or CRP level||1||0|
For our initial evaluation of the U-DA, we evaluated only 1 lesion scan per patient, selecting the skin lesion that the clinician considered to be most active for the active group. The 13 active skin lesions, from patients A1–A9 and M10–M13, consisted of 7 linear scleroderma lesions, 5 circumscribed morphea lesions, and 1 generalized morphea lesion. The 8 inactive skin lesions, from patients I14–I21, consisted of 3 linear scleroderma lesions, 3 circumscribed morphea lesions, 1 generalized morphea lesion, and 1 pansclerotic morphea lesion.
Table 2 shows the U-DA and individual parameter scores for these lesions. Active lesions showed varying patterns of sonographic differences, with most showing both echogenicity and vascularity differences (A1–A3, A5, A6, A9, M11, M13), some showing only echogenicity differences (A7, A8, M10, M12), and 1 showing no sonographic difference (A4). Lesions also varied in the location of these differences. Echogenicity differences were either found in all tissue layers (A6, M10, M13), or one (A9) or both deeper layers (A1, A2, A7, A8, M11, M12). Hyperemia was localized to one (A2, M9, M11, M13) or both deeper tissue layers (A1, A6) (Table 2).
|Dermis||Hypo-dermis||Deep tissue||Dermis||Hypo-dermis||Deep tissue|
|A1||Linear scleroderma, extremity/trunk||Lower leg||0||2||1||0||1||2||6|
|A2||Circumscribed deep morphea||Thigh||0||1||1||0||0||3||5|
|A3||Circumscribed deep morphea||Face, cheek||1||2||NP||0||2||NP||5|
|A4||Circumscribed super morphea||Breast||0||0||0||0||0||0||0|
|A5||Linear scleroderma, face/scalp||Temple||1||2||NP||0||1||NP||4|
|A6||Generalized morphea||Mid back||1||2||1||0||1||1||6|
|A7||Linear scleroderma, extremity/trunk||Thigh||0||1||1||0||0||0||2|
|A8||Linear scleroderma, extremity/trunk||Forearm||0||3||2||0||0||0||5|
|A9||Linear scleroderma, extremity/trunk||Upper arm||0||1||0||0||0||2||3|
|M10||Circumscribed super morphea||Breast||1||2||2||0||0||0||5|
|M11||Linear scleroderma, extremity/trunk||Scapula||0||3||1||0||0||1||5|
|M12||Linear scleroderma, extremity/trunk||Thigh||0||2||1||0||0||0||3|
|M13||Linear scleroderma, extremity/trunk||Chest||1||1||2||0||0||1||5|
|I14||Linear scleroderma, extremity/trunk||Lower leg||0||0||0||0||0||0||0|
|I15||Circumscribed deep morphea||Forehead||0||2||0||0||0||0||2|
|I16||Linear scleroderma, face/scalp||Forehead||0||0||NP||0||0||NP||0|
|I17||Linear scleroderma, face/scalp||Forehead||1||1||NP||0||0||NP||2|
|I19||Circumscribed super morphea||Back||1||1||NP||0||2||NP||4|
|I21||Circumscribed super morphea||Foot||0||0||NP||0||0||NP||0|
The 1 clinically active A4 skin lesion with a U-DA score of 0 (A4) was classified as circumscribed superficial morphea. One of the other 3 circumscribed superficial morphea lesions also had a U-DA score of 0 (I21, inactive), while the other 2 circumscribed superficial morphea lesions had U-DA scores of 5 (M10, active) and 4 (I19, inactive). Both M10 and I19 showed sonographic differences in the hypodermis; M10 also had deep tissue layer differences.
For these 21 skin lesions, significant differences were found for mean U-DA scores between active and inactive skin lesions (P = 0.0045). The mean ± SD U-DA score was 4.15 ± 1.70 for active lesions and 1.38 ± 1.60 for inactive skin lesions (Table 3). Significant differences between active and inactive skin lesions were found for the U-DA parameters hypodermis echogenicity (P = 0.0081) and deep tissue echogenicity (P = 0.0124). Deep tissue vascularity approached significance (P = 0.0885). No significant differences were found for dermis echogenicity or vascularity (Table 3).
|Set 1†||Set 2‡|
|U-DA score||4.15 ± 1.70||1.38 ± 1.60||0.0045||3.78 ± 1.83||1.85 ± 2.08||0.0010|
|Total||2.85 ± 1.46||0.75 ± 1.04||0.0042||2.66 ± 1.26||1.10 ± 1.41||0.0014|
|Dermis||0.38 ± 0.51||0.25 ± 0.46||NS||0.47 ± 0.51||0.30 ± 0.47||NS|
|Hypodermis||1.69 ± 0.85||0.50 ± 0.76||0.0081||1.44 ± 0.88||0.60 ± 0.75||0.0023|
|Deep tissue layer||1.09 ± 0.71||0 ± 0||0.0124||0.85 ± 0.66||0.40 ± 0.52||0.095|
|Dermis||0 ± 0||0 ± 0||NS||0 ± 0||0 ± 0||NS|
|Hypodermis||0.38 ± 0.65||0.50 ± 0.93||NS||0.47 ± 0.72||0.70 ± 1.03||NS|
|Deep tissue layer||0.91 ± 1.04||0 ± 0||0.0885||0.78 ± 0.93||0.10 ± 0.32||0.0374|
|Dermal||0.54 ± 0.88||−0.13 ± 0.83||NS||0.38 ± 0.83||−0.056 ± 0.94||0.0697|
|Hypodermis||1.62 ± 1.04||0.88 ± 1.64||NS||1.67 ± 0.96||1.12 ± 1.41||NS|
|Deep tissue layer||0.50 ± 2.12||1.33 ± 1.15||NS||1.14 ± 1.07||0.57 ± 1.13||NS|
We next analyzed U-DA scores for all scans acquired from these 21 patients on their acquisition date to evaluate a larger number of lesions and the variation in U-DA scores between different skin lesions from a given patient. Nine patients had scans acquired of multiple distinct lesions; 3 of circumscribed (I18, I19, I21), 4 of both linear scleroderma and circumscribed (A5, M10, M11, M13), 1 of 2 different linear scleroderma (M12), and 1 of different generalized morphea lesions (A6). Eight patients had scans acquired at 2 or more portions of a large lesion. Seven patients (A7–A9, M10–M13) had linear scleroderma involving the entire extremity. The other patient had pansclerotic morphea of the arm (I20).
Thirty-two scans of active skin lesions were acquired from 13 patients (range 1–5 per patient, median 2), with 20 scans from 9 patients considered to have only active lesions and 12 from 4 patients considered to have both active and inactive skin lesions. Twenty scans were acquired of inactive skin lesions (range 1–4 per patient, median 1), with 14 scans from 8 patients considered to have only inactive lesions and 6 from 4 patients considered to have both active and inactive skin lesions. No correlation was found between specific clinical features of activity and U-DA score, or with its parameters. The average U-DA score was 3.8 for the 16 skin lesions (50%) with erythema, 4.1 for the 16 skin lesions (50%) with warmth, 4.3 for the 7 skin lesions (21.9%) with a violaceous color, 3.3 for the 6 new skin lesions (18.8%), 4.8 for the 5 skin lesions (15.6%) with increased tissue loss, and 3.6 for the 5 skin lesions (15.6%) with induration. Three skin lesions had increased pigmentation, visible veins, or thinning; their U-DA scores were 5, 5, and 3, respectively. Fifteen skin lesions (46.9%) had 2 features, most commonly warmth and erythema (n = 10, mean ± SD U-DA score 3.5 ± 2.0), and 6 skin lesions (18.8%) had 3 features (mean ± SD U-DA score 4.8 ± 2.5).
Generally different skin lesions from a given patient had similar U-DA scores (data not shown). As was seen with the first set of scans, significant differences between active and inactive scans were found for mean U-DA scores, hypodermis echogenicity, and total echogenicity in the second set (P = 0.0010, 0.0023, and 0.0014, respectively) (Table 3). Results from these 2 sets of scans differed for mean deep tissue echogenicity and vascularity, with mean deep tissue vascularity now found significantly different, whereas mean deep tissue echogenicity was not significantly different (P = 0.0374 and P = 0.095, respectively) (Table 3).
The increased power from the larger sample size likely enabled us to find deep tissue vascularity to be significantly different in the 52-scan analysis. The failure to find a significant difference between active and inactive skin lesions for deep tissue echogenicity in the larger scan set may represent inaccurate clinical assessment of activity. None of the scans from patients that were classified as only having inactive skin lesions had increased deep tissue echogenicity, while 4 of the 5 inactive skin lesion scans from patients considered to have both active and inactive skin lesions had increased deep tissue echogenicity (data not shown). These patients are discussed further in the next section.
Four patients (M10–M13) had scans acquired of both clinically active and inactive skin lesions on the same day (Table 4). Two patients (M10 and M12) showed the expected pattern of lower U-DA scores for skin lesions rated clinically inactive compared with those rated active. The other 2 patients (M11 and M13), however, showed similar or even higher scores in the skin lesions rated as inactive. Over ∼4 years the linear scleroderma of one patient (M11) progressed from her hand to scapula, and she developed new circumscribed morphea lesions on her face and chest. The active scapula, cheek, chin, and chest lesions had erythema, violaceous color, warmth, and/or were new or larger, and had U-DA scores ranging from 2–5 (Table 4). Her affected hand showed only chronic atrophy, but was imaged because of intermittent hand pain. Ultrasound imaging showed robust hyperemia of hypodermis and muscle layers, and increased echogenicity of all hand tissue layers (U-DA score of 9) (Table 4). Another patient (M13) had a 10-year history of linear scleroderma extending up her arm with recent extension to her anterior chest. The chest and upper arm portions of her lesion were erythematous, while the older lower portion and an older abdominal lesion showed only chronic changes. All had similar U-DA scores (5, 4, 3, and 4, respectively) (Table 4).
|Patient||Lesion location||Disease activity†||Clinical feature(s)||Total echogenicity‡||Total vascularity§||U-DA score|
|M10||Arm||I||Chronic atrophy, dyspigmentation||2||0||2|
|Forearm||I||Chronic atrophy, dyspigmentation||0||0||0|
|Scapula||A||Lesion extension, warmth||3||3||6|
|M12||Foot||I||Dyspigmentation, moderate atrophy||0||0||0|
|Calf||A||Erythema, violaceous, warmth||5||4||9|
|Left thigh||A||Erythema, warmth||3||0||3|
|Right thigh||A||New lesion||3||0||3|
|M11||Hand||I||Chronic atrophy, shorter finger/hand||4||5||9|
|Scapula||A||Erythema, violaceous, lesion extension, warmth||4||1||5|
|Infraclavicular||A||New lesion, warmth||3||0||3|
|Cheek||A||New lesion, erythema, shinier||3||0||3|
|Chin||A||New lesion, violaceous||2||0||2|
|M13||Wrist||I||Chronic atrophy, dyspigmentation||3||0||3|
|Abdomen||I||Chronic atrophy, dyspigmentation||4||0||4|
|Chest||A||New lesion, erythema, warmth||4||1||5|
We examined differences in tissue layer thickness between lesion and normal site by calculating tissue layer differences as the percent difference of lesion to normal tissue layer thickness. Most active skin lesions showed thinning of the dermis (mean ± SD −3.97% ± 45.79%), while inactive skin lesions were more likely to show thickening or no difference (mean ± SD 9.39% ± 47.78%). The majority of skin lesions, active and inactive, showed thinning of the hypodermis (mean ± SD −36.13% ± 29.96% for active and mean ± SD −16.79% ± 54.87% for inactive). Only a minority of skin lesion scans had measurable deep tissue layers; 5 of the 6 active skin lesions showed thinning compared with 3 of the 7 inactive skin lesions. None of the tissue layer differences between active and inactive scans were significant.
Tissue thickness differences were also evaluated by the TTS, which scores for thinning versus thickening (negative numbers) and grades the degree of thinning (positive numbers). No significant differences were found between active and inactive skin lesions for TTS for any tissue layer (Table 3).
The U-DA was developed to standardize interpretation of ultrasound images of skin lesions in juvenile localized scleroderma. In this study, we carried out the first evaluation of the construct validity of this measure for detecting disease activity. We found that the U-DA could discriminate between clinically active and inactive localized scleroderma skin lesions (P = 0.0010). Of the different U-DA components, significant differences between active and inactive skin lesions were found for total echogenicity, hypodermis echogenicity, and deep tissue layer vascularity scores (P = 0.0014, 0.0023, and 0.0374, respectively).
Unlike prior adult localized scleroderma studies, we did not find a significant difference in dermal echogenicity between active and inactive skin lesions or a significant association between tissue thickness differences and disease activity state. Although studies of adults have reported dermal thickening in early active lesions, we found dermal thickening to be more common in inactive lesions and dermal thinning to be more common in active lesions. Both active and inactive skin lesions were more likely to have hypodermis thinning than thickening. There were too few skin lesions with measurable deep tissue layers for us to adequately assess whether there were significant differences associated with this layer.
Some of the differences between our findings and prior adult localized scleroderma studies may reflect technical differences. The ultrasound frequencies we used provided less detailed dermis examination than those used in most adult studies (8–14 MHz versus 20 MHz) (7–9). This may have limited our ability to detect subtler differences in dermis echogenicity, although not our ability to measure dermal thickness. The higher frequencies, however, have limited depth penetration, making evaluation of the hypodermis and deeper tissues incomplete or not possible (22). Our initial studies showed that changes in these layers were common in pediatric skin lesions (15). In this study, most lesions were found to have abnormal hypodermal echogenicity levels (28 of 32 active scans and 9 of 20 inactive scans). Most active skin lesions also showed a difference in deep tissue layer echogenicity (19 of 27 scans), and almost one-half a difference in deep tissue vascularity (13 of 27 scans). Study differences may therefore reflect differences between pediatric and adult disease. The most common localized scleroderma subtype in pediatric patients is linear scleroderma, while circumscribed (plaque) morphea is the most common adult subtype (1, 23). Linear scleroderma commonly affects deep tissue layers including muscle and bone (3), while most circumscribed morphea is superficial, often involving only the dermis (24).
There was no correlation between specific clinical activity features and echogenicity or vascularity changes overall or in a specific tissue layer. There were a few discrepancies between U-DA scores and clinical assessment of disease activity. Two patients (M11 and M13) showed similar to greater levels of sonographic differences in their “clinically inactive” lesions compared with their clinically active lesions. The “inactive” hand region of patient M11 was imaged because of pain at this site. With continued treatment, the patient's hand pain resolved, tissue hyperemia disappeared, and tissue hyperechogenicity decreased (data not shown), suggesting that ultrasound may detect localized scleroderma disease activity that is not clinically apparent. Clinical assessment of disease depth was also found limited in some cases; 2 lesions classified as circumscribed superficial morphea showed sonographic changes in the hypodermis or hypodermis and deep tissue layer. However, as shown by patient A4, not all clinically apparent activity was demonstrable by our ultrasound method. No echogenicity or vascularity differences were found in this patient's warm erythematous lesion.
Our study is limited by its retrospective design and limited sample size. Activity status of the patient's skin lesion was determined from chart review, not histologic evaluation. In addition, the scorers of the U-DA were not blinded to the patients, leaving scoring open to potential bias. However, the disease activity status had been determined prior to scoring, and the score was a consensus of both evaluators in an effort to minimize bias. In addition, most patients had had multiple scans, so that recall of the specific details of any given skin lesion scan would be difficult.
We find that the U-DA can detect differences between active and inactive pediatric localized scleroderma skin lesions. These sonographic differences were located in the hypodermis and deep tissue layers. Changes in these layers, especially in the deeper portions of the hypodermis or in the deep tissue layer, would be difficult to detect by clinical examination. A combined clinical and ultrasound examination may therefore provide a more complete assessment of disease activity. Further study is needed to evaluate the histologic correlates of the sonographic differences and to evaluate the validity, reliability, and sensitivity to change of the U-DA. Additional evaluation of the correlation between tissue layer thickness differences and disease activity is also needed. Although the degree of thickness differences did not distinguish active from inactive lesions, we may find that monitoring changes in the degree of difference helps to identify disease activity. Lesions that show continued thinning or thickening may be more likely to be active than those with static thickness differences. Having a relatively objective instrument available will facilitate multisite studies to evaluate treatment regimens. This will aid in choosing the most effective treatment and thereby improve outcome in juvenile localized scleroderma.
All authors were involved in drafting the article or revising it critically for important intellectual content, and all authors approved the final version to be submitted for publication. Dr. Li had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.
Study conception and design. Li, Liebling, Haines, Weiss, Prann.
Acquisition of data. Li, Liebling, Haines, Weiss.
Analysis and interpretation of data. Li, Liebling, Haines, Weiss, Prann.