In Vivo Assessment and Monitoring of Burn Wounds Using a Handheld Terahertz Hyperspectral Scanner

Abstract

The accuracy of clinical assessment techniques in diagnosing partial-thickness burn injuries has remained as low as 50-76%. Depending on the burn depth and environmental factors in the wound, such as reactive oxygen species, inflammation, and autophagy, partial-thickness burns can heal spontaneously or require surgical intervention. Herein, it is demonstrated that terahertz time-domain spectral imaging (THz-TDSI) is a promising tool for in vivo quantitative assessment and monitoring of partial-thickness burn injuries in large animals. We used a novel handheld THz-TDSI scanner to characterize burn injuries in a porcine scald model with histopathological controls. Statistical analysis (n= 40) indicates that the THz-TDSI modality can accurately differentiate between partial-thickness and full-thickness burn injuries (1-way ANOVA, p< 0.05). THz-TDSI has the potential to improve burn care outcomes by helping surgeons in making objective decisions for early excision of the wound.

Keywords: burn imaging; handheld THz-TDS scanner; histology; partial-thickness burn characterization; skin tissue hyperspectral imaging; terahertz time-domain spectroscopy.

Figure 1.

a–c) Visual images, d–f) THz images, and g–i) mean of the deconvolved THz spectra are shown for representative tissue samples from each burn condition on Days 0 and 4. The quantity shown in the color bar is SA. In (g–i), the red traces and red error bars (mean ± SD) correspond to the white ROIs drawn in the burned tissue in (d–f), while the black traces and black error bars (mean ± SD) correspond to the black ROIs drawn in the healthy skin regions. (Scale bar = 5 mm).

Figure 2.

a) A representative THz amplitude reflectivity, R(f), is plotted as an example to demonstrate the calculation of spectral slopes, SS, and spectral area, SA. The red area shows SA within the f_lower and f_upper frequencies. The dotted line shows the linear fit between f_lower and f_Rmax to calculate SS_lower. The dashed line shows the linear fit between f_Rmax and f_upper to calculate SS_upper. b) The ability of the Z parameter to differentiate between burn wounds is statistically tested (1-way ANOVA, p = 0.0016) against histological assessment of the burn depth. The THz measurements and biopsies were both obtained on Day 4 postburn. The burns were grouped into two categories: those with depth of damage greater or less than 50% of the dermis layer.

Figure 3.

The daily longitudinal monitoring of the spectral area, SA, and the Z-parameter values (mean ± SD in the ROI subsets) are shown for four representative burn injuries: a,b) superficial partial thickness injury with 22% burn depth, c,d) superficial partial-thickness injury with 37% burn depth, e,f) deep-partial thickness injury with 65% burn depth, and g,h) full thickness injury with 100% burn depth. ROIs in (a, c, e, g) were drawn to be in the same general area as the previous day. Pixel size is 1 × 1 mm². By including the THz data from all burns in this study, a trend similar to representative examples in (a-h) was observed. The Z parameter increased over time in burns with depth of damage less than 50% of the total dermal thickness (i), but remained relatively constant for burns with damage greater than 50% of the total dermal thickness (j). A one-way ANOVA between the <50% and >50% groups shows that p < 0.05 on Day 4 (represented by *) and p < 0.1 on Days 3 and 2 (represented by †).

Figure 4.

a) The schematic of the scald burn induction device is shown, where 95 °C water would come in direct contact with the skin. b) Scald condition locations are illustrated where the numbers in each circle represent the heat exposure duration in seconds. c) Dermal burn depth percentage (normalized by total dermal thickness) are plotted for each experimental condition, as determined by a histopathologist using Day 4 H&E-stained sections.

Figure 5.

Images of H&E-stained skin samples are shown for each scald condition. Images were captured using a Nikon mm-60 microscope (Nikon Instruments Inc. Tokyo, Japan) with a 5 ×objective lens. Yellow arrows indicate cells marked by nuclear pyknosis. (Scale bar = 0.2 mm).

Figure 6.

a) The optical schematic for the THz PHASR scanner. b) A visual image of the handheld scanner in use.

Figure 7.

a) Peak dynamic range shows a linear relationship with acquisition time per pixel. Values for peak dynamic range were calculated as the ratio of the maximum frequency domain amplitude to the noise floor level. b) The calculated bandwidth shows a monotonically increasing relationship with the data acquisition time or increasing the number of averages. Marker shapes correspond to the number of averages and marker colors represent the ASOPS difference frequency. Large black arrows represent the optimal operation parameters selected for this study (Δf = 100 Hz, 20 Avgs).

Figure 8.

The signal processing flowchart is shown with selected steps depicted below and noted where analysis is performed in the time or frequency domain. Vertical magenta dotted lines represent the f_lower and the f_upper.