DiatomNet v1.0: A novel approach for automatic diatom testing for drowning diagnosis in forensically biomedical application.

BACKGROUND AND OBJECTIVE: Diatom testing is supportive for drowning diagnosis in forensic medicine. However, it is very time-consuming and labor-intensive for technicians to identify microscopically a handful of diatoms in sample smears, especially under complex observable backgrounds. Recently, we successfully developed a software, named DiatomNet v1.0 intended to automatically identify diatom frustules in a whole slide under a clear background. Here, we introduced this new software and performed a validation study to elucidate how DiatomNet v1.0 improved its performance with the influence of visible impurities. METHODS: DiatomNet v1.0 has an intuitive, user-friendly and easy-to-learn graphical user interface (GUI) built in the Drupal and its core architecture for slide analysis including a convolutional neural network (CNN) is written in Python language. The build-in CNN model was evaluated for diatom identification under very complex observable backgrounds with mixtures of common impurities, including carbon pigments and sand sediments. Compared to the original model, the enhanced model following optimization with limited new datasets was evaluated systematically by independent testing and random control trials (RCTs). RESULTS: In independent testing, the original DiatomNet v1.0 was moderately affected, especially when higher densities of impurities existed, and achieved a low recall of 0.817 and F1 score of 0.858 but good precision of 0.905. Following transfer learning with limited new datasets, the enhanced version had better results, with recall and F1 score values of 0.968. A comparative study on real slides showed that the upgraded DiatomNet v1.0 obtained F1 scores of 0.86 and 0.84 for carbon pigment and sand sediment, respectively, slightly worse than manual identification (carbon pigment: 0.91; sand sediment: 0.86), but much less time was needed. CONCLUSIONS: The study verified that forensic diatom testing with aid of DiatomNet v1.0 is much more efficient than traditionally manual identification even under complex observable backgrounds. In terms of forensic diatom testing, we proposed a suggested standard on build-in model optimization and evaluation to strengthen the software's generalization in potentially complex conditions.

Forensic Pathological Diagnosis of Acute and Old Myocardial Infarction Using Fourier Transform Infrared Spectroscopy.

OBJECTIVES: Fourier transform infrared spectroscopy (FTIR) was used to analyze myocardial infarction tissues at different stages of pathological change to achieve the forensic pathology diagnosis of acute and old myocardial infarction. METHODS: FTIR spectra data of early ischemic myocardium, necrotic myocardium, and myocardial fibrous tissue in the left ventricular anterior wall of the sudden death group of atherosclerotic heart disease and the myocardium of the normal control group were collected using hematoxylin-eosin (HE) and immunohistochemistry (IHC) staining as a reference, and the data were analyzed using multivariate statistical analysis. RESULTS: The mean normalized spectra of control myocardium, early ischemic myocardium and necrotic myocardium were relatively similar, but the mean second derivative spectra were significantly different. The peak intensity of secondary structure of proteins in early ischemic myocardium was significantly higher than in other types of myocardium, and the peak intensity of the α-helix in necrotic myocardium was the lowest. The peaks of amide Ⅰ and amide Ⅱ in the mean normalized spectra of myocardial fibrous tissue significantly shifted towards higher wave numbers, the peak intensities of amide Ⅱ and amide Ⅲ were higher than those of other types of myocardium, and the peak intensities at 1 338, 1 284, 1 238 and 1 204 cm-1 in the mean second derivative spectra were significantly enhanced. Principal component analysis (PCA) and partial least square-discriminant analysis (PLS-DA) showed that FTIR could distinguish different types of myocardium. CONCLUSIONS: FTIR technique has the potential to diagnose acute and old myocardial infarction, and provides a new basis for the analysis of the causes of sudden cardiac death.

Evaluating Subtle Pathological Changes in Early Myocardial Ischemia Using Spectral Histopathology.

Early myocardial ischemia (EMI) is morphologically challenging, and the results from conventional histological staining may be subjective, imprecise, or even silent. The size of myocardial necrosis determines the acute and long-term mortality of EMI. The precise diagnosis of myocardial ischemia is critical for both clinical management and forensic investigation. Fourier transform infrared (FTIR) spectroscopic imaging is a highly sensitive tool for detecting protein conformations and imaging protein profiles. The aim of this study was to evaluate the application of FTIR imaging with multivariate analysis to detect biochemical changes in the protein conformation in the early phase of myocardial ischemia and to visually classify different disease states. The spectra and curve fitting results revealed that the total protein content decreased significantly in the EMI group and that the α-helix content of the secondary protein structure continuously decreased as ischemia progressed, while the ß-sheet content increased. Differences in the control and EMI groups and perfused and ischemic myocardium were confirmed using principal component analysis and partial least squares discriminant analysis. Next, two support vector machine classifiers were effectively created. The accuracy, recall, and precision were 99.98, 99.96, and 100.00%, respectively, to differentiate the EMI group from the control group and 99.25, 98.95, and 99.54%, respectively, to differentiate perfused and ischemic myocardium. Ultimately, high EMI diagnostic accuracy was achieved with 100.00% recall and 100.00% precision, and ischemic myocardium diagnostic accuracy was achieved with 99.30% recall and 99.53% precision for the test set. This pilot study demonstrated that FTIR imaging is a powerful automated quantitative analysis tool to detect EMI without morphological changes and will improve diagnostic accuracy and patient prognosis.

Research Progress of Automatic Diatom Test by Artificial Intelligence.

Diatom test is the main laboratory test method in the diagnosis of drowning in forensic medicine. It plays an important role in differentiating the antemortem drowning from the postmortem drowning and inferring drowning site. Artificial intelligence (AI) automatic diatom test is a technological innovation in forensic drowning diagnosis which is based on morphological characteristics of diatom, the application of AI algorithm to automatic identification and classification of diatom in tissues and organs. This paper discusses the morphological diatom test methods and reviews the research progress of automatic diatom recognition and classification involving AI algorithms. AI deep learning algorithm can assist diatom testing to obtain objective, accurate, and efficient qualitative and quantitative analysis results, which is expected to become a new direction of diatom testing research in the drowning of forensic medicine in the future.

Comparison among Four Deep Learning Image Classification Algorithms in AI-based Diatom Test.

OBJECTIVES: To select four algorithms with relatively balanced complexity and accuracy among deep learning image classification algorithms for automatic diatom recognition, and to explore the most suitable classification algorithm for diatom recognition to provide data reference for automatic diatom testing research in forensic medicine. METHODS: The "diatom" and "background" small sample size data set (20 000 images) of digestive fluid smear of corpse lung tissue in water were built to train, validate and test four convolutional neural network (CNN) models, including VGG16, ResNet50, InceptionV3 and Inception-ResNet-V2. The receiver operating characteristic curve (ROC) of subjects and confusion matrixes were drawn, recall rate, precision rate, specificity, accuracy rate and F1 score were calculated, and the performance of each model was systematically evaluated. RESULTS: The InceptionV3 model achieved much better results than the other three models with a balanced recall rate of 89.80%, a precision rate of 92.58%. The VGG16 and Inception-ResNet-V2 had similar diatom recognition performance. Although the performance of diatom recall and precision detection could not be balanced, the recognition ability was acceptable. ResNet50 had the lowest diatom recognition performance, with a recall rate of 55.35%. In terms of feature extraction, the four models all extracted the features of diatom and background and mainly focused on diatom region as the main identification basis. CONCLUSIONS: Including the Inception-dependent model, which has stronger directivity and targeting in feature extraction of diatom. The InceptionV3 achieved the best performance on diatom identification and feature extraction compared to the other three models. The InceptionV3 is more suitable for daily forensic diatom examination.

Non/mini-invasive monitoring of diabetes-induced myocardial damage by Fourier transform infrared spectroscopy: Evidence from biofluids.

Early identification of diabetic cardiomyopathy (DCM) can help clinicians develop targeted treatment plans and forensic pathologists make accurate postmortem diagnoses. In the present study, diabetes-induced metabolic abnormalities in the myocardium and biofluids (plasma, urine, and saliva) of db/db mice of various ages (7, 12, and 21 weeks) were investigated by attenuated total reflection (ATR)-Fourier transform infrared (FTIR) spectroscopy. The results indicated that the diabetic and control groups had significantly different changes in the function groups of lipids, phosphate macromolecules (mostly nucleic acids), protein compositions and conformations, and carbohydrates (primarily glucose) in the myocardium and biofluids. The prediction model for quantifying DCM severity was developed on db/db mice's myocardial spectra using a genetic algorithm (GA)-partial least squares (PLS) regression method. Following that, the linear correlations between the predicted values for DCM severity and spectra for db/db biofluids were evaluated using the GA-PLS regression algorithm. The results showed there were good linear correlations between the predicted values for DCM severity and spectra for plasma (R2 = 0.929), saliva (R2 = 0.967), urine (R2 = 0.954), and combination of plasma and saliva (R2 = 0.980). This study provides a novel perspective on detecting diabetes-related biofluid and cardiac metabolic abnormalities and demonstrates the potential of biofluid infrared spectro-diagnostic models for non/mini-invasive assessment of DCM.

Investigation of early biochemical alterations in myocardia of the diabetic db/db mice by FTIR microspectroscopy combined with machine learning.

Diabetic cardiomyopathy (DbCM) is a serious complication of diabetes that affects about 12% of the diabetic population. Sensitive detection of diabetes-induced biochemical changes in the heart before symptoms appear can assist clinicians in developing targeted treatment plans and forensic pathologists in making accurate postmortem diagnoses. The Fourier transform infrared (FTIR) spectroscopy-based approach allows for the analysis of the sample biomolecular composition and variations. In the current study, the myocardial tissues of mouse models of type 2 diabetes mellitus (T2DM) at various ages (7, 12, and 21 weeks) were analyzed using FTIR microspectroscopy (FTIRM) in combination with machine learning algorithms. The carbonyl esters, olefinic=CH and CH2 groups of lipids, total lipids, saccharides, and ß-sheet to α-helix conformational transition in proteins increased significantly in diabetic mice myocardial tissues compared to healthy mice. Furthermore, partial least-squares discriminant analysis and random forest-guided partial least-squares discriminant analysis revealed the time-dependent progression of the spectral lipidomic profiles during the development of DbCM. Finally, a random forest classifier was developed for diagnosing DbCM, with 97.1% accuracy. This study demonstrates that FTIRM is a novel method for monitoring early biochemical changes in the myocardia of mice with T2DM.

Label-Free Diagnosis of Pulmonary Fat Embolism Using Fourier Transform Infrared (FT-IR) Spectroscopic Imaging.

The diagnosis of pulmonary fat embolism (PFE) is of great significance in the field of forensic medicine because it can be considered a major cause of death or a vital reaction. Conventional histological analysis of lung tissue specimens is a widely used method for PFE diagnosis. However, variable and labor-intensive tissue staining procedures impede the validity and informativeness of histological image analysis. To obtain complete information from tissues, a method based on infrared imaging of unlabeled tissue sections was developed to identify pulmonary fat emboli in the present study. We selected 15 PFE-positive lung samples and 15 PFE-negative samples from real cases. Oil red O (ORO) staining and infrared spectral imaging collection were both performed on all lung tissue samples. And the fatty tissue of the abdominal wall and the embolized lipid droplets in the lungs were taken for comparison. The results of the blind, evaluation by pathologists, showed good agreement between the infrared spectral imaging of the lung tissue and the standard histological stained images. Fourier transform infrared (FT-IR) spectroscopic imaging significantly simplifies the typical painstakingly laborious histological staining procedure. And we found a difference between lipid droplets embolized in abdominal wall fat and lung tissue.

A potential method for sex estimation of human skeletons using deep learning and three-dimensional surface scanning.

Deep learning based on radiological methods has attracted considerable attention in forensic anthropology because of its superior classification capacities over human experts. However, radiological instruments are limited in their nature of high cost and immobility. Here, we integrated a deep learning algorithm and three-dimensional (3D) surface scanning technique into a portable system for pelvic sex estimation. Briefly, the images of the ventral pubis (VP), dorsal pubis (DP), and greater sciatic notch (GSN) were cropped from virtual pelvic samples reconstructed from CT scans of 1000 individuals; 80% of them were used to train and internally evaluate convolutional neural networks (CNNs) that were then evaluated externally with the remaining samples. An additional 105 real pelvises were documented virtually with a handheld 3D surface scanner, and the corresponding snapshots of the VP, DP, and GSN were predicted by the trained CNN models. The CNN models achieved excellent performance in the external testing using CT-based images, with accuracies of 98.0%, 98.5%, and 94.0% for VP, DP, and GSN, respectively. When the CT-based models were applied to 3D scanning images, they obtained satisfactory accuracies above 95% on the VP and DP images compared to the GSN with 73.3%. In a single-blind trial, a multiple design that combined the three CNN models yielded a superior accuracy of 97.1% with 3D surface scanning images over two anthropologists. Our study demonstrates the great potential of deep learning and 3D surface scanning for rapid and accurate sex estimation of skeletal remains.

Exploring metabolic alterations associated with death from asphyxia and the differentiation of asphyxia from sudden cardiac death by GC-HRMS-based untargeted metabolomics.

The determination of cause of death is one of the most important tasks in forensic practice. However, asphyxia is a difficult cause of death to determine, especially when the deceased has an underlying disease that can lead to a sudden unexpected death, such as coronary atherosclerotic heart disease (CAHD, which is the leading cause of sudden cardiac death, SCD), because its determination is currently still based on an exclusion strategy. In this study, gas chromatography coupled with high-resolution mass spectrometry (GC-HRMS)-based untargeted metabolomics was employed to obtain the pulmonary metabolic profiles of rats who died from asphyxia and SCD. First, fourteen metabolites were identified to investigate the mechanism of death from asphyxia, and we proposed some explanations that may account for these metabolic alterations, including the perturbation of amino acid metabolism, lipid metabolism, and energy metabolism (TCA cycle). Second, we discovered eight potential biomarkers to differentiate between asphyxia and SCD as the cause of death. The excellent classification performances of the eight individual biomarkers and their combination in fresh lung tissue were observed. Third, we also explored the relative change in the concentration of the eight metabolites and their classification performance in decomposed tissue (at 24 h postmortem). Lactic acid, pantothenic acid, and the combination of the eight biomarkers can be recognized as perfect classifiers to discriminate asphyxia from SCD even when decomposition has occurred. Our results showed that GC-HRMS-based untargeted metabolomics can be used as a promising tool to explore the metabolic alterations of the death process and to determine the cause of death.

The Use of Gas Chromatography Coupled with High-Resolution Mass Spectrometry-Based Untargeted Metabolomics to Discover Metabolic Changes and Help in the Determination of Complex Causes of Death: A Preliminary Study.

The determination of cause of death (COD) is one of the most important tasks in forensic practice and is mainly based on macroscopical and microscopical morphological signatures. However, some CODs are hard to determine because the significant morphological signatures can be nonspecific, variable, subjective, or even absent in the real world. In this study, gas chromatography coupled with high-resolution mass spectrometry (GC-HRMS)-based untargeted metabolomics was employed to obtain plasma metabolic profiles of rats that died from anaphylactic shock (AS), mechanical asphyxia (MA), or sudden cardiac death (SCD). The metabolic alterations of each COD group compared to the control group were investigated using a principal component analysis, partial least-squares discriminant analysis, the Wilcoxon test, and fold change analysis. A range of differential features was screened, and 11, 8, and 7 differential metabolites were finally verified for the AS, MA, and SCD groups, respectively. We proposed some explanations that may account for these metabolic differences, including glucose metabolism, the tricarboxylic acid cycle, glycolysis, lipid metabolism, creatinine catabolism, and purine metabolism. Next, for each COD, we used its differential metabolites, which were obtained through comparisons of each COD group to the control group and represented the metabolic changes of the individual COD, to perform a receiver operating characteristic (ROC) analysis to preliminarily evaluate their ability to discriminate each COD group from the other COD groups. We found that creatinine in the AS group and malic acid and uric acid in the MA group might represent some specific metabolic changes for the relevant COD with high areas under the curve in the ROC curve analysis. Moreover, the combination panel for AS or MA also showed a good ability to discriminate it from the others. However, SCD had fewer metabolic signatures and was relatively harder to discriminate from the other CODs in our work. The preliminary study demonstrates the feasibility of GC-HRMS-based untargeted metabolomics as a promising tool to reveal metabolic changes in different death processes and to determine the complex CODs.

An efficient method for building a database of diatom populations for drowning site inference using a deep learning algorithm.

Seasonal or monthly databases of the diatom populations in specific bodies of water are needed to infer the drowning site of a drowned body. However, existing diatom testing methods are laborious, time-consuming, and costly and usually require specific expertise. In this study, we developed an artificial intelligence (AI)-based system as a substitute for manual morphological examination capable of identifying and classifying diatoms at the species level. Within two days, the system collected information on diatom profiles in the Huangpu and Suzhou Rivers of Shanghai, China. In an animal experiment, the similarities of diatom profiles between lung tissues and water samples were evaluated through a modified Jensen-Shannon (JS) divergence measure for drowning site inference, reaching a prediction accuracy of 92.31%. Considering its high efficiency and simplicity, our proposed method is believed to be more applicable than existing methods for seasonal or monthly water monitoring of diatom populations from sections of interconnected rivers, which would help police narrow the investigation scope to confirm the identity of an immersed body.

Research advances in forensic diatom testing.

In forensic practice, it is difficult to determine whether a dead body in the water resulted from drowning or from disposal after death. Diatom testing is currently an important supporting technique for the determination of death by drowning and of drowning sites, even though it is a time-consuming and laborious task. This article reviews the development of diatom testing over the decades and discusses a new method for the potential application of deep learning in diatom testing.

Determination of causes of death via spectrochemical analysis of forensic autopsies-based pulmonary edema fluid samples with deep learning algorithm.

This study investigated whether infrared spectroscopy combined with a deep learning algorithm could be a useful tool for determining causes of death by analyzing pulmonary edema fluid from forensic autopsies. A newly designed convolutional neural network-based deep learning framework, named DeepIR and eight popular machine learning algorithms, were used to construct classifiers. The prediction performances of these classifiers demonstrated that DeepIR outperformed the machine learning algorithms in establishing classifiers to determine the causes of death. Moreover, DeepIR was generally less dependent on preprocessing procedures than were the machine learning algorithms; it provided the validation accuracy with a narrow range from 0.9661 to 0.9856 and the test accuracy ranging from 0.8774 to 0.9167 on the raw pulmonary edema fluid spectral dataset and the nine preprocessing protocol-based datasets in our study. In conclusion, this study demonstrates that the deep learning-equipped Fourier transform infrared spectroscopy technique has the potential to be an effective aid for determining causes of death.

Digital whole-slide image analysis for automated diatom test in forensic cases of drowning using a convolutional neural network algorithm.

Diatom examinations have been widely used to perform drowning diagnosis in forensic practice. However, current methods for recognizing diatoms, which use light or electron microscopy, are time-consuming and laborious and often result in false positive or negative decisions. In this study, we demonstrated an artificial intelligence (AI)-based system to automatically identify diatoms in conjunction with a classical chemical digestion approach. By employing transfer learning and data augmentation methods, we trained convolutional neural network (CNN) models on thousands or tens of thousands of tiles from digital whole-slide images of diatom smears. The results showed that the trained model identified the regions containing diatoms in the tiles. In an independent test, where the slide samples were collected in forensic casework, the best CNN model demonstrated a performance competitive with those of 5 forensic pathologists with experience in diatom quantification. This pilot study paves the way for future intelligent diatom examinations; many efficient diatom extraction methods could be incorporated into our automated system.

Biochemical detection of fatal hypothermia and hyperthermia in affected rat hypothalamus tissues by Fourier transform infrared spectroscopy.

It is difficult to determinate the cause of death from exposure to fatal hypothermia and hyperthermia in forensic casework. Here, we present a state-of-the-art study that employs Fourier-transform infrared (FTIR) spectroscopy to investigate the hypothalamus tissues of fatal hypothermic, fatal hyperthermic and normothermic rats to determine forensically significant biomarkers related to fatal hypothermia and hyperthermia. Our results revealed that the spectral variations in the lipid, protein, carbohydrate and nucleic acid components are highly different for hypothalamuses after exposure to fatal hypothermic, fatal hyperthermic and normothermic conditions. In comparison with the normothermia group, the fatal hypothermia and hyperthermia groups contained higher total lipid amounts but were lower in unsaturated lipids. Additionally, their cell membranes were found to have less motional freedom. Among these three groups, the fatal hyperthermia group contained the lowest total proteins and carbohydrates and the highest aggregated and dysfunctional proteins, while the fatal hypothermia group contained the highest level of nucleic acids. In conclusion, this study demonstrates that FTIR spectroscopy has the potential to become a reliable method for the biochemical characterization of fatal hypothermia and hyperthermia hypothalamus tissues, and this could be used as a postmortem diagnostic feature in fatal hypothermia and hyperthermia deaths.

Attenuated total reflectance Fourier transform infrared (ATR-FTIR) spectral prediction of postmortem interval from vitreous humor samples.

Evaluation of postmortem interval (PMI) is of paramount importance to guide criminal investigations, especially when witnesses are not found. However, accurate PMI estimation is a challenging task in the forensic community due to the limitations of existing methods. The study aims to investigate the potential of attenuated total reflectance Fourier transform infrared (ATR-FTIR) spectroscopy for predicting PMI based on vitreous humor (VH). VH samples were collected from 72 rabbits in the range of 0-48 h postmortem at a 6-h interval. Their FTIR spectra were normalized by the extended multiplicative signal correction (RMSC) and divided into calibration and validation sets. After analysis of the absorption bands, the Bayesian ridge regression (BRR), support vector regression (SVR), and artificial neural network (ANN) methods were established by the calibration set using a 10-fold cross-validation that was further used to predict the PMI in the validation set. The validity of the models was assessed by a permutation test. The current study demonstrated that multiple macromolecules in the VH samples were reflected in a FTIR spectrum, and the spectral absorption bands at 1313 and 925 cm-1 were highly correlated with PMI. The three models allowed generalization to the validation set due to similar R2 and errors between the calibration and validation tests. The highest accuracy with R2 = 0.983 and error = 2.018 h was achieved by the ANN model in the validation test. The results suggest that ATR-FTIR spectroscopy may be useful for VH analysis in order to predict PMI in the future. Graphical abstract ᅟ.

Identification of Pulmonary Edema in Forensic Autopsy Cases of Sudden Cardiac Death Using Fourier Transform Infrared Microspectroscopy: A Pilot Study.

Many studies have proven the usefulness of biofluid-based infrared spectroscopy in the clinical domain for diagnosis and monitoring the progression of diseases. Here we present a state-of-the-art study in the forensic field that employed Fourier transform infrared microspectroscopy for postmortem diagnosis of sudden cardiac death (SCD) by in situ biochemical investigation of alveolar edema fluid in lung tissue sections. The results of amide-related spectral absorbance analysis demonstrated that the pulmonary edema fluid of the SCD group was richer in protein components than that of the neurologic catastrophe (NC) and lethal multiple injuries (LMI) groups. The complementary results of unsupervised principle component analysis (PCA) and genetic algorithm-guided partial least-squares discriminant analysis (GA-PLS-DA) further indicated different global spectral band patterns of pulmonary edema fluids between these three groups. Ultimately, a random forest (RF) classification model for postmortem diagnosis of SCD was built and achieved good sensitivity and specificity scores of 97.3% and 95.5%, respectively. Classification predictions of unknown pulmonary edema fluid collected from 16 cases were also performed by the model, resulting in 100% correct discrimination. This pilot study demonstrates that FTIR microspectroscopy in combination with chemometrics has the potential to be an effective aid for postmortem diagnosis of SCD.

Identification of pulmonary edema in forensic autopsy cases of fatal anaphylactic shock using Fourier transform infrared microspectroscopy.

Anaphylaxis is a rapid allergic reaction that may cause sudden death. Currently, postmortem diagnosis of anaphylactic shock is sometimes difficult and often achieved through exclusion. The aim of our study was to investigate whether Fourier transform infrared (FTIR) microspectroscopy combined with pattern recognition methods would be complementary to traditional methods and provide a more accurate postmortem diagnosis of fatal anaphylactic shock. First, the results of spectral peak area analysis showed that the pulmonary edema fluid of the fatal anaphylactic shock group was richer in protein components than the control group, which included mechanical asphyxia, brain injury, and acute cardiac death. Subsequently, principle component analysis (PCA) was performed and showed that the anaphylactic shock group contained more turn and α-helix protein structures as well as less tyrosine-rich proteins than the control group. Ultimately, a partial least-square discriminant analysis (PLS-DA) model combined with a variables selection method called the genetic algorithm (GA) was built and demonstrated good separation between these two groups. This pilot study demonstrates that FTIR microspectroscopy has the potential to be an effective aid for postmortem diagnosis of fatal anaphylactic shock.

MALDI-TOF MS as a Novel Tool for the Estimation of Postmortem Interval in Liver Tissue Samples.

Estimation of the postmortem interval (PMI) is a complicated task in forensic medicine, especially during homicide and unwitnessed death investigations. Many biological, chemical, and physical indicators can be used to determine the postmortem interval, but most are not accurate. Here, we present a novel matrix-assisted laser desorption/ionization time-of-flight mass spectrometry (MALDI-TOF MS) method that can be used for the estimation of PMI using molecular images and multivariate analyses. In this study, we demonstrate that both rat and human liver tissues of various PMIs (0, 2, 4, and 6days) can be discriminated using MALDI imaging and principal component analysis (PCA). Using genetic algorithm (GA), supervised neural network (SNN), and quick classifier (QC) methods, we built 6 classification models, which showed high recognition capability and good cross-validation. The histological changes in all the samples at different time points were also consistent with the changes seen in MALDI imaging. Our work suggests that MALDI-TOF MS, along with multivariate analysis, can be used to determine intermediate PMIs.