Deep Learning in Postmortem Interval Estimation: A Systematic Review of Current Practices and Future Directions

Estimating the postmortem interval remains a challenge in forensic medicine due to the high environmental variability affecting decomposition. This systematic review examines the available scientific evidence on the application of deep learning techniques to improve the accuracy of postmortem interval prediction. A systematic review was conducted following the PRISMA 2020 guidelines in major biomedical databases up to September 2025. Studies applying deep learning in human or animal models for postmortem interval prediction were included. Methodological quality and risk of bias were assessed using PROBAST+AI. Of the 508 initially identified records, 8 studies met the inclusion criteria. Deep learning models, primarily artificial neural networks and convolutional neural networks, frequently demonstrated higher accuracy than traditional methods in characterizing complex postmortem degradation patterns. However, 7 studies presented a high risk of methodological bias, mainly due to small sample sizes or experimental design limitations. Deep learning holds potential for improving postmortem interval estimation, but its clinical application requires further validation in human cohorts and standardized protocols.