(E-175) Multispectral Imaging of Cutaneous Water for Skin Tone Insensitive Inflammation Assessment

Introduction:: Dermatological inflammatory skin conditions impact an estimated 20-25% of the population and are the fourth leading cause of nonfatal disability^1,2. The primary clinical tool for diagnosis and management is qualitative visual inspection of disease-related changes in skin morphology³. Interobserver variability can result in misdiagnosis. Non-white patients face additional challenges as erythema appears more subtly against pigmented skin, resulting in systematic underdiagnosis⁴. There exists an urgent clinical need for a technology that can objectively and quantitatively evaluate cutaneous inflammation across diverse skin tones⁵. Edema, or an increase in local tissue fluid content, is a pathophysiological hallmark of inflammation. In the short-wave infrared (SWIR) electromagnetic spectrum (~900-1800 nm), water is the dominant absorber and melanin absorption is relatively low⁶. This is in contrast to the visible spectrum (VIS), where water is virtually invisible, and melanin is highly absorbent. It has been shown that spectral SWIR reflectance imaging can be used to noninvasively obtain quantitative maps of tissue water^3,6-9. This drives our central hypothesis that a SWIR multispectral imaging modality would address the need for an improved inflammation assessment tool by obtaining high contrast images of edema regardless of skin tone. Prior SWIR cutaneous water imaging studies have primarily utilized near infrared (NIR) wavelengths, built stationary research-only systems, and have not investigated the impact of skin tone. Therefore, the objective of this study was to A) build and characterize a clinic-ready visible-SWIR imaging modality and B) demonstrate the ability of the device to capture images of cutaneous water on skin of variable pigmentation.


Materials and Methods:: A novel imaging modality was constructed to enable widefield multispectral imaging. The core components were an InGaAs camera (Allied Vision) and a aberration correcting lens (Computar). A six-position filter wheel with bandpass filters centered at 540, 650, 850, 1200, 1300, and 1600 nanometers were placed in front of the lens (Thorlabs). Two 35 W broadband (400-2200 nm) halogen lamps (Amazon) were positioned on either side. All components were connected to a computer, and a MATLAB interface was written to automate image acquisition. Specifications were characterized at an imaging distance of 0.2 m.

The devices’ ability to interrogate cutaneous water on various skin tones was assessed. One Caucasian male and one Black male were recruited. Each subject's skin tone was measured using a Konica Minolta CM700d spectrophotometer and the corresponding Individual Topography Angle (ITA) was calculated¹⁰. Each subject received a .2 mL intradermal injection of saline, and the resultant wheal was imaged¹¹. Each six-image set was flatfield corrected and normalized. Pseudo red-green-blue color images, derived by stacking three multispectral channels, were generated to demonstrate the relative ability of the SWIR and VIS-NIR spectrum to capture water. The first image was constructed with a NIR wavelength, 850 nm, and two VIS bands, 650 nm and 540 nm, in the R-G-B channels respectively. The second image was constructed with SWIR wavelengths 1200, 1300, and 1600 nm in ascending channels. The Weber contrast between the saline region and a region of adjacent background skin was calculated (I_wheal-I_background/I_background) as a quantitative measure of signal intensity.


Results, Conclusions, and Discussions:: The novel modality supports portable and automatic acquisition at six spectral bands from 400-1600 nm. The device was characterized at 0.2 meters as follows: FOV 6.5x5.2 cm; Resolution .051 mm per pixel; Power Density ~158 µW/cm2; Set Acquisition Time 10.2 seconds. The device’s portability, FOV, and automated software support simplified subject imaging. The primary performance limitation is acquisition time, which will be addressed with optimization.

The device measured a significantly higher water signal in the SWIR than in the VIS-NIR spectrum on both subjects. The Caucasian subject had ventral forearm ITA of 53.0º, consistent with ‘light’ skin tone, and the Black subject had ITA of 26.8º, consistent with ‘tan’ skin tone. The six-image set was used to construct two three-channel pseudo images for each subject, shown in Figure 1. The saline wheal was practically invisible in the VIS-NIR images. The strong absorption of hemoglobin at 850 nm resulted in a prominent vein. Alternatively in the SWIR images, a darker oval is clearly distinguishable. The Weber contrast values of the lightly- and darkly-pigmented subjects were .1155 and .0272 in the VIS-NIR multispectral image, and .4595 and .5200 in the SWIR pseudo-image. The pseudo images support the strategy that SWIR derived images would provide clinicians with an easily interpretable, pigmentation insensitive, and high contrast perception of inflammation. The objective contrast quantification may permit standardized longitudinal monitoring.

This work evaluated a novel imaging modality, motivated by the need for a more objective and equitable method of inflammatory skin disease assessment. A portable VIS-SWIR imaging modality is a significant step toward collecting the necessary data for further validation. Skin pigmentation is known to make skin disease visualization more challenging by eye and impacts traditional light-based devices. Our preliminary results indicate that SWIR imaging may ensure that cutaneous imaging meets the needs of a diverse patient population. Successful translation of an objective point-of-care cutaneous inflammation evaluation tool would have a profound impact on the standardization of skin disease management. Future work will include saline injection studies across the full spectrum of skin tone, and imaging clinical cases of Atopic Dermatitis, Psoriasis, and Hidradenitis Suppurativa.

Acknowledgements (Optional): : We thank Dr. Mikhail Berezin, Dr. Christine Obrien, and the WUSTL Department of Radiology for support and discussion.

References (Optional): :

1. Adamson, A. S. The Economics Burden of Atopic Dermatitis. Management of Atopic Dermatitis. vol. 1027 79–92 (2017).


2. Karimkhani Aksut, C. Global skin disease morbidity and mortality. J. Invest. Dermatol. 137, S31 (2017).


3. Stamatas, G. N. In Vivo Monitoring of Cutaneous Edema. Biomed. Opt. vol. 126 issue 8. 1753-1760 (2006).


4. Alexis, A. F. Psoriasis in skin of color. J. Clin. Aesthetic Dermatol. 7, 16–24 (2014).


5. Liu, C.H. FASEB journal. vol. 33, 12. (2019).


6. Nachabé, R. Estimation of biological chromophores using diffuse optical spectroscopy. Biomed. Opt. Express 1, 1432 (2010).


7. Ezerskaia, A. Potential of short-wave infrared spectroscopy for quantitative depth profiling of stratum corneum lipids and water in dermatology. Biomed. Opt. Express 9, 2436 (2018).


8. Wilson, R. H. Review of short-wave infrared spectroscopy. J. Biomed. Opt. 20, 030901 (2015).


9. Shibata, S. Novel Assessment of Acute Lung Injury.. Am. J. Respir. Crit. Care Med. 160, 317–323 (1999).


10. Bino, Sandra. Relationship between skin response to ultraviolet exposure and skin color type. Pigment cell research. 19. 606 (2007).


11. Shrestha, P. Fluid absorption by skin tissue during intradermal injections through hollow microneedles. Sci. Rep. 8, 13749 (2018).