Identification of Skin Electrical Injury Using Infrared Imaging: A Possible Complementary Tool for Histological Examination

Abstract

In forensic practice, determination of electrocution as a cause of death usually depends on the conventional histological examination of electrical mark in the body skin, but the limitation of this method includes subjective bias by different forensic pathologists, especially for identifying suspicious electrical mark. The aim of our work is to introduce Fourier transform infrared (FTIR) spectroscopy in combination with chemometrics as a complementary tool for providing an relatively objective diagnosis. The results of principle component analysis (PCA) showed that there were significant differences of protein structural profile between electrical mark and normal skin in terms of α-helix, antiparallel β-sheet and β-sheet content. Then a partial least square (PLS) model was established based on this spectral dataset and used to discriminate electrical mark from normal skin areas in independent tissue sections as revealed by color-coded digital maps, making the visualization of electrical injury more intuitively. Our pilot study demonstrates the potential of FTIR spectroscopy as a complementary tool for diagnosis of electrical mark.

Fig 1. Macroscopic and microscopic images of electrical mark and normal skin.

(A) macroscopic features of electrical mark in the hand. Microscopic features of electrical mark (B) and normal skin (C).

Fig 2. A comparison of absorbance and second derivative spectra between electrical mark and normal skin.

(A) absorbance spectra within 1800–900 cm^-1 for both categories. Superimposition of them within the Amide I region in absorbance mode (B) and second derivative mode (C). The blue and red spectra represent electrical injury and normal epidermis respectively.

Fig 3. The results of PCA based on the Amide I band.

(A) the score plot along PC 1 versus PC 2 is obtained where the blue and red circles represent the spectra from electrical injury and normal epidermis (control) respectively, and the larger circle is 95% confidence interval. (B) PCA loading plot associated with PC 1 is obtained where the negative peaks are characteristic for normal epidermis (control) and positive for electrical injury.

Fig 4. The comparison of FTIR mappings based on protein conformations between electrical mark and normal skin.

The first and second rows represent FTIR mappings in the areas of elongation cells and corresponding normal epidermis respectively, in which the peak absorbance intensities associated α-helix, β-turn, β-sheet and antiparallel β-sheet are revealed by “jet” colors with red indicating the highest relative concentration, and blue, the lowest.

Fig 5. The results of PLS classification based on spectral dataset in the calibration.

(A) PLS score plot shows that all groups representing different structures of the skin are well-separated along latent factor 1. (B) Prediction result in an additional dataset that are not included for PLS modelling shows that electrical epidermis, normal epidermis and dermis can be distinguishable from each other by using the PLS model. Red and black dot lines show classification threshold for separating the three categories respectively.

Fig 6. The identification of electrical mark by the FTIR-based PLS model in an independent tissue section.

HE stained skin tissue sections are presented here where elongation cell in electrical mark and normal skin regions are included in A and B respectively. The spectra within the black box of the HE staining images were classified by the PLS model, and the prediction results in the electrical mark (C) and normal skin areas (D) were revealed by pseudo-colored images. Electrical epidermis, normal epidermis and normal dermis appear in yellow, light blue and brown respectively.