Post mortem computed tomography meets radiomics: a case series on fractal analysis of post mortem changes in the brain.

Estimating the post-mortem interval is a fundamental, albeit challenging task in forensic sciences. To this aim, forensic practitioners need to assess post-mortem changes through a plethora of different methods, most of which are inherently qualitative, thus providing broad time intervals rather than precise determinations. This challenging problem is further complicated by the influence of environmental factors, which modify the temporal dynamics of post-mortem changes, sometimes in a rather unpredictable fashion. In this context, the search for quantitative and objective descriptors of post-mortem changes is highly demanded. In this study, we used computed tomography (CT) to assess the post-mortem anatomical modifications occurring in the time interval 0-4 days after death in the brain of four corpses. Our results show that fractal analysis of CT brain slices provides a set of quantitative descriptors able to map post-mortem changes over time throughout the whole brain. Although incapable of producing a direct estimation of the PMI, these descriptors could be used in combination with other more established methods to improve the accuracy and reliability of PMI determination.

Estimation of the time of death by measuring the variation of lateral cerebral ventricle volume and cerebrospinal fluid radiodensity using postmortem computed tomography.

Using postmortem CT (PMCT), changes in the volume of the lateral cerebral ventricles (LCVs) and modifications of the radiodensity of cerebrospinal fluid (CSF) have been examined to identify a possible relationship between these changes and the time of death. Subsequent periodical CT scans termed "sequential scans" for ten corpses at known time of death were obtained, and a 3D segmentation of the entire LCV was carried out to measure its volume and radiodensity over time from ~ 5.5- h up to 273-h postmortem. A linear decrease of the LCV volume for all the cases was observed in the investigated time range, together with an overall logarithmic increase of radiodensity. Although a larger sampling should be performed to improve the result reliability, our finding suggests that the postmortem variation of CSF radiodensity can be a potentially useful tool in determining postmortem interval, a finding that is worthy of further investigation.