Novel multispectral imaging to predict disease progression in pediatric morphea

Morphea, or localized scleroderma, is an inflammatory, fibrosing skin disorder that can be progressive and debilitating. Infrared thermography frequently has false positive results. The aim of this study was to assess the ability of multispectral imaging to predict disease progression in children with morphea.


| INTRODUCTION
Morphea, also known as localized scleroderma, is an inflammatory, fibrosing skin disorder that affects 3.4 children per million per year in Ireland and the United Kingdom. 1 Morphea can be progressive and debilitating, and a key prognostic factor is early introduction of potent anti-inflammatory therapy. 2 Assessment of disease activity is difficult and distinction between early-inflamed lesions and late sclerotic lesions can lead to either undertreatment and therefore persistent disease, or overtreatment, with side effects from therapy potentially outweighing benefits. 3The Localized Scleroderma Assessment Tool (LoSCAT) is a clinical score with two domains, an activity index (LoSAI) and a damage index (LoSDI), across 18 anatomic sites. 4The LoSAI for each area includes the sum of three separate activity scores -new or enlarging lesions, erythema, and induration, each scored from 0 to 3. The LoSDI for each area includes dermal atrophy, subcutaneous atrophy, and dyspigmentation, each scored from 0 to 3. Scores for each site are based on the most severe score for each parameter.In order to minimize intersubject variability, all skin changes are compared with the contralateral skin area.LoSCAT is easy to use, is validated in pediatric morphea, and has shown excellent reliability. 5However it is liable to subjectivity and inter-observer variability.Infrared thermography has been commonly used as an objective monitoring tool of choice to assess disease activity in morphea. 6Activity is defined as a 0.5 C discrepancy between affected skin and either surrounding or contralateral unaffected skin.
However, because damaged sclerotic tissue often has significant subcutaneous atrophy, false positive results are common due to enhanced heat transfer from subcutaneous blood vessels. 6In addition, thermography is often not available outside of expert centers, and requires specialist equipment and trained technicians.Multispectral imaging, capturing image data within a wide wavelength range across the electromagnetic spectrum, can provide a dense, information-rich color dataset and can provide absorption data at ultraviolet and infrared wavelengths. 7The Antera™ 3D camera, based on advanced optical technology developed at Trinity College Dublin, is a handheld portable device that easily, noninvasively, and painlessly views the skin in up to three dimensions as well as performing multispectral analysis. 8The parameters it reports on are hemoglobin, elevation, melanin, and depression, which correlate with the LoSCAT parameters of erythema, induration, dyspigmentation, and atrophy, respectively.
The aim of this study was to assess the ability of multispectral imaging to predict disease progression, defined as new or enlarging lesions within the subsequent 3 months, in children with morphea.

| MATERIALS AND METHODS
Ethical approval was granted by our clinical research and ethics committee in January 2016 (ECM 4 [y] January 01, 2016].Children with morphea were recruited from clinic and multispectral images of affected and matched contralateral unaffected sites were obtained using the handheld portable Antera™ 3D camera (Figure 1A).The Antera™ 3D is a camera that acquires images and analyses skin by illuminating the surface from different angles and using the differences between these images to digitally reconstruct the surface in three dimensions.Antera™ 3D uses reflectance mapping of seven different light wavelengths spanning the entire visible spectrum, allowing for a precise analysis of skin colorimetric properties, which are mostly determined by two dominant chromophores: melanin and hemoglobin.This camera measures an area of 3136 mm 2 (56 Â 56 mm 2 ). 8Analysis of hemoglobin, melanin, depth, and volume was performed using the Antera Clinical Studies software, with values corresponding to the average value across the field of the camera image.Pilot studies showed that individual hemoglobin variation scores are very responsive to ambient temperature.Therefore, the differential in hemoglobin variation between affected and matched contralateral unaffected skin was selected as the most likely best surrogate marker for clinical erythema, and therefore disease activity.Clinical assessment was performed using the LoSCAT and the Children's F I G U R E 1 Scatter plot showing hemoglobin gradient divided according to development of progression within 3 months (in blue) or no progression (in orange).Numbers on the x-axis refers to individual patient scores representing the average hemoglobin gradient between affected and unaffected sites, according to progression or not.
Dermatology Life Quality Index (CDLQI).Children were followed up every 3 months for repeat imaging and clinical review and were informed to contact the study team if they had new or enlarging lesions.
Progression was defined as an absolute increase of at least two points or a relative increase of at least 20% in the LoSCAT activity index, or a new or enlarging lesion at the time of imaging or within the interval 3 months.
The average LoSCAT score was 20.6 (range 5-73).Consistent with a cohort with moderately severe disease.The average activity component was moderate at 6.6 (range 0-39) and the average damage score was moderate at 15.2 (range 3-43) using previously defined cut-off scores. 5The average CDLQI was 3.7 (range 0-14), suggesting a small effect on quality of life. 9n of the 17 patients (58.9%) were deemed to have progression at one or more study visits during the duration of the study.All 17 patients had at least one study visit where their disease was considered to be stable.The average hemoglobin gradient between affected and matched contralateral unaffected skin was four times higher in patients at the time of disease progression (n = 10), compared to the hemoglobin gradient in patients at times of disease stability (n = 17), with an average index differential of 0.3% (range 0.1-0.4) in the progression group versus 0.08% (range 0.02-0.15) in the group who did not progress.Using a cut-off of a 0.18% hemoglobin gradient between affected and matched contralateral unaffected skin, patients could easily be differentiated into those who subsequently had progression and those who did not (Figure 2A and   2B).With only one patient who flared having a gradient below this cut off, and none of the children who did not progress having gradients above this cut off, the sensitivity of multispectral imaging in pediatric morphea is 90% with specificity of 100%.This gives a positive predictive value of 100%, and a negative predictive value of 95%.
There was no statistically significant relationship between other characteristics analyzed by the Antera™ 3D camera (melanin, depth, and volume) and disease progression.The melanin hyperconcentration measurement uses an au (arbitrary unit), with lower values signifiying lower concentrations of melanin and higher levels suggesting higher concentrations of melanin.
Morphea is a relatively uncommon dermatosis, but can be associated with significant morbidity, and is complicated by relatively poor methods for assessing disease activity.The lack of robust clinical tools to determine disease activity has recently been highlighted. 10 has not previously been studied in morphea, where subcutaneous atrophy is common, and the internal calculations may not be as accurate in the setting of significant tissue atrophy.The Antera™ 3D camera is also less helpful than the LoSCAT score in providing a global assessment of activity if there is extensive disease, as the average hemoglobin gradient will include both active and inactive areas.For this reason, the hemoglobin gradient of each individual lesion should be assessed separately, which can be time consuming if there is widespread involvement.Furthermore, while this device is promising in terms of assessing disease activity in patients with deeply pigmented skin, there are theoretical concerns about interactions Example of multispectral analysis from the Antera software in a patient with darkly pigmented skin, showing melanin level and variation, and hemoglobin hyperconcentration without melanin analysis, highlighting the utility of the assessment tool in patients with darkly pigmented skin.The circle area is randomly selected and is used to match areas according to the parameter selected.
between melanin and hemoglobin due to spectral overlap. 11However the Antera 3D camera derives the chromophore concentration from the spatial and spectral analysis of the acquired image data using seven different light wavelengths spanning the entire visible spectrum, allowing more reliable discrimination between skin erythema and melanin. 8 conclusion, this is a novel assessment tool with promising accuracy in predicting progression as an adjunct to clinical assessment.It is suitable for lightly and darkly pigmented skin and can help to guide therapy across the spectrum of disease from subtle to severe.Further research should examine its performance against thermography.

F
I G U R E 2 Example of multispectral analysis in patient with darkly pigmented skin, showing melanin level and variation, and hemoglobin gradient with melanin subtracted.T A B L E 1 Demographics of cohort of children with morphea recruited to the study.
Multispectral imaging has the benefit of specifically examining the hemoglobin concentration in tissue, and when compared to matched contralateral tissue, can indicate local inflammation and therefore activity.Specific cases where the camera was of particular use included subtle disease, progression in the context of severe damage, and darkly pigmented skin.In subtle or limited disease, progression might be underestimated because of minimally apparent inflammation or a small affected area.In severe damage, erythema is difficult to appreciate in new or enlarging lesions and thermography can be difficult to interpret.As multispectral imaging accounts for tissue volume and depth, it reduces the risk of false positive results as seen in thermography when dermal and subcutaneous atrophy result in increased thermal transmission.In darkly pigmented skin, the hemoglobin component can be analyzed alone, without the melanin component, with makes the hemoglobin gradient, and consequently erythema, much easier to appreciate (Figure3).Strengths of this study include the use of a novel innovative tool to solve a clinical problem, the comparison with validated clinical assessment tools, a good sample size for an uncommon condition, the range of severity sampled, and the acquisition of multiple datapoints per patient, permitting longitudinal evaluation.Limitations include missing data related to the COVID-19 pandemic, the single-center nature of the study, and the lack of direct comparison with thermography.Moreover, the Antera™ 3D camera