New perspectives of forensic pathology through machine learning approach on autopsy data: a pilot study

Area of Science:

• Forensic Pathology
• Medical Informatics
• Computational Biology

Background:

• Traditional forensic pathology relies on expert interpretation of complex data.
• Experience-based judgment is crucial but can be enhanced by objective analytical tools.
• This study explores artificial intelligence (AI) for improved data collection and analysis in forensic pathology.

Purpose of the Study:

• To analyze autopsy data using machine learning (ML) for forensic case evaluation.
• To introduce a novel research direction in forensic pathology leveraging AI.
• To determine the predictive value of solid organ weights in determining cause of death.

Main Methods:

• Analysis of judicial autopsies conducted between 01/01/2020 and 31/12/2021.
• Inclusion of medical records review, autopsy findings, histological examination, and toxicological analysis.
• Application of a random forest regression model, a machine learning technique, to assess organ weight importance.

Main Results:

• Routine autopsy data can be effectively processed using machine learning techniques.
• AI identifies key procedural elements that enhance the prediction of cause of death.
• The study highlights the significance of organ weight variables in predicting mortality.

Conclusions:

• This research is among the first to investigate organ weight's role in predicting cause of death.
• Artificial intelligence offers an optimal solution for complex forensic challenges.
• Machine learning integration into forensic analysis provides valuable insights for cause of death determination.