Accurate Identification of Traumatic Lung Injury (TLI) by ATR-FTIR Spectroscopy Combined with Chemometrics.
Xinggong Liang,Gongji Wang,Zefeng Li,Run Chen,Hao Wu,Huiyu Li,Chen Shen,Mingyan Deng,Zeyi Hao,Shuo Wu,Kai Yu,Xin Wei,Ruina Liu,Kai Zhang,Qinru Sun,Zhenyuan Wang
DOI: https://doi.org/10.1016/j.saa.2022.122186
IF: 4.831
2022-01-01
Spectrochimica Acta Part A Molecular and Biomolecular Spectroscopy
Abstract:Traumatic lung injury (TLI), which is a common mechanical injury, is receiving increasing attention because of its serious hazards. In forensic practices, accurately identifying TLI is of great importance for investigations and case trials. The main goal of this research was to identify TLI utilizing attenuated total reflection-Fourier transform infrared (ATR-FTIR) spectroscopy in combination with chemometrics. The macroscopic appearance of lung tissue showed that identifying TLI in lung tissue at the decomposition stage is not feasible by only visualization, and significant pulmonary hypostasis was observed in the lungs regardless of whether the lung tissue was injured. Average spectra and principal component analysis (PCA) suggested that the biochemical difference between injured lung tissue samples from the TLI group and noninjured lung tissue samples from the negative control group was mainly attributed to the different structures and contents of proteins. Partial least squares discriminant analysis (PLS-DA) was then utilized to identify TLI with an accuracy of 96.4% and 98.6% based on the training set and the test set, respectively. Next, we focused on samples that were misclassified in the model and proposed that the misclassification could be caused by the pulmonary hypostasis effect. Therefore, two additional PCA and PLS-DA models were created to identify the pulmonary hypostatic areas between the TLI group and the negative control group and the nonpulmonary hypostatic areas between the TLI group and the negative control group. The PCA results indicated that the biochemical difference between the two groups was still associated with proteins, and the two PLS-DA models achieved 100% accuracy based on both the training and test sets. This result indicated that when pulmonary hypostasis was considered and the lung tissue was divided into pulmonary hypostatic areas and nonpulmonary hypostatic areas for separate comparisons, TLI identification was achieved with a greater accuracy than that obtained when the two areas were combined. This research confirms that the combined application of ATR-FTIR spectroscopy and chemometrics can be utilized to accurately identify TLI.