A novel approach for estimating postmortem intervals under varying temperature conditions using pathology images and...

Area of Science:

• Forensic Science
• Computational Pathology
• Artificial Intelligence

Background:

• Accurate postmortem interval (PMI) estimation is crucial in forensic investigations.
• Traditional PMI methods face challenges due to environmental variability and subjectivity.
• There is a need for more objective and reliable PMI estimation techniques.

Purpose of the Study:

• To develop and validate a novel predictive model for PMI estimation using pathological tissue images and AI.
• To assess the model's performance under different temperature conditions (4°C, 25°C, 37°C).

Main Methods:

• Utilized a ResNet50 neural network for analyzing patch-level tissue images.
• Extracted deep learning features and integrated them with machine learning for whole slide image (WSI) classification.
• Trained and tested the model on a dataset under controlled temperature conditions.

Main Results:

• The AI model achieved high performance with micro and macro AUC values of at least 0.949 at the patch-level and 0.800 at the WSI-level.
• External validation demonstrated patch-level AUC values exceeding 0.960.
• The model showed robust performance across different temperature conditions.

Conclusions:

• AI-powered analysis of pathological tissue images offers a promising approach for accurate and efficient PMI estimation.
• This novel model has the potential to significantly enhance forensic investigations.
• Further advancements in AI can lead to more precise forensic case resolutions.