The research, published as a proof-of-concept pilot study in the International Journal of Legal Medicine, compared two walking scenarios. In the first, a volunteer stepped into a standardized 20-milliliter pool of fresh human blood and then walked ten consecutive steps at a sustained pace, mimicking someone moving steadily away from a blood source. In the second, the walker paused for two to three minutes after the first step, allowing the blood film on the outsole to partially dry before resuming at a slower pace, simulating a person who lingers near a blood pool before leaving. Each condition was run twice, producing four complete depletion sequences, with every single footprint photographed under fixed, calibrated conditions.

The numbers tell a dramatic story of rapid loss. In the continuous walking sequence, the initial blood pool covered roughly 160 square centimeters, but the very first footprint already carried only 60 to 70 square centimeters of blood. By the second step, that figure had collapsed to 20 to 30 square centimeters, and by the third it fell to 10 to 20 square centimeters. More than 80 percent of the transferable blood vanished within the first three contacts, a pattern the authors describe as consistent with a negative exponential decay model, in which the bulk of available blood is shed almost immediately and subsequent steps leave progressively smaller residues.

Yet quantity is not the whole story. Despite this steep quantitative collapse, the overall geometry of the shoe, its tread pattern and outline, remained recognizable for a surprisingly large number of consecutive steps in the continuous sequence. The researchers analyzed each image using custom Python and OpenCV algorithms, converting photographs into HSV color space to isolate blood pixels, then computing a battery of indices: a Fragmentation Density Index tracking how broken up the pattern became, a Spatial Dispersion Index measuring how far residual deposits spread from the footprint’s core, and a Transfer Ratio comparing each step’s blood area to the first. As blood area shrank, fragmentation relentlessly increased, with the highest values appearing in the terminal stages when only isolated microdeposits remained.

The interrupted partial-drying sequence produced an unexpected twist. Counterintuitively, letting the blood partially dry did not preserve the footprint longer. Instead, interpretability declined sharply around the fourth step, the point at which the overall tread geometry could no longer be reliably reconstructed. What partial drying changed was not longevity but geography: residual deposits became more localized and concentrated rather than diffusely scattered. The authors caution that this Step 4 transition is specific to their experiment, not a universal forensic threshold, and that their design cannot separate the individual effects of the pause, the drying itself, and the slower walking pace that followed.

Perhaps the most striking discovery concerns the mismatch between what can be seen and what can be detected chemically. Using Combur Test strips, a tetramethylbenzidine-based reagent that responds to the pseudoperoxidase activity of hemoglobin, the team tested traces at each stage. Fresh deposits reacted strongly, but after roughly 24 hours of drying, direct testing of the traces was predominantly negative, even where blood was still visible to the eye. The crucial move came next: rehydrating the dried traces with physiological saline restored positive reactions, even in advanced stages of depletion and even on the shoe’s outsole itself after a full day.

This dissociation between morphological persistence and biochemical persistence carries real weight for criminal investigations. A footprint that has degraded beyond recognition, reduced to scattered punctiform flecks with no reconstructable tread pattern, may still harbor biologically reactive blood material. The study suggests that apparently worthless, fragmented impressions could retain forensically relevant evidence, including material potentially suitable for confirmatory testing, long after their value for pattern comparison has evaporated. Investigators who dismiss heavily depleted traces as uninformative may be discarding evidence that a combined morphological and biochemical approach could recover.

The authors are careful to frame their conclusions as hypothesis-generating rather than definitive. The study involved a single 59-kilogram subject, one worn athletic shoe in EU size 43, one blood source collected in EDTA tubes, and one smooth non-porous surface resembling linoleum. Blood viscosity, water loss, film thickness, and plantar pressure were not directly measured, and the image segmentation retained an analyst-dependent component without formal inter-rater validation. Combur itself is only a presumptive test, indicating heme reactivity but not confirming that a stain is human blood. Real crime scenes introduce countless variables, from carpet and concrete to rain and scrubbing, that this controlled pilot could not address.

Even so, the work opens a productive line of inquiry at the intersection of fluid physics, biomechanics, and forensic science. Blood is a complex non-Newtonian fluid whose wetting, spreading, and drying behavior on surfaces has only recently attracted systematic study, and previous research has shown that DNA can survive on shoe soles for hundreds of steps. The Italian team’s integrated approach, fusing digital morphometry with presumptive chemistry, offers a template for future studies incorporating multiple subjects, diverse footwear, varied substrates, and gait analysis. The broader message is already clear: the footprint you can see and the blood you can detect follow two different clocks, and forensic science needs to read both.

Subject of Research: Depletion dynamics of blood-bearing footwear impressions in forensic bloodstain pattern analysis

Article Title: Blood footwear impression depletion: comparative morphological and biochemical analysis of continuous and interrupted partial-drying transfer sequences

Article References: Aquila, I., Gualtieri, S., Grimaldi, G., & Sacco, M. A. (2026). Blood footwear impression depletion: comparative morphological and biochemical analysis of continuous and interrupted partial-drying transfer sequences. International Journal of Legal Medicine. https://doi.org/10.1007/s00414-026-04035-1

Image Credits: AI Generated

DOI: 10.1007/s00414-026-04035-1

Keywords: forensic science, bloodstain pattern analysis, footwear impressions, blood transfer, hemoglobin detection, Combur Test, morphometric analysis, crime scene reconstruction, partial drying, exponential decay, legal medicine, trace evidence

The research, led by Siqi Chen and Lihua Hou of Central South University in Changsha, China, together with colleagues at the Institute of Forensic Science in Beijing and Hebei Medical University, focused on a specific and stubborn forensic problem: pulp-free teeth. When a tooth loses its pulp, whether through decay, treatment, or postmortem degradation, investigators lose the soft tissue that normally yields abundant, high-quality DNA. What remains are the hard mineralized tissues, dentin and cementum, which preserve DNA remarkably well over time but have been difficult to exploit for age prediction. The team set out to determine whether DNA methylation patterns in these hard tissues could still encode a person’s chronological age with usable precision.

DNA methylation is a chemical modification in which methyl groups attach to cytosine bases at specific positions in the genome, known as CpG sites. Across a person’s lifetime, the pattern of these modifications changes in a systematic, tissue-dependent way, a phenomenon that underlies the so-called epigenetic clock first described by Steve Horvath in 2013. Because methylation is chemically stable and can be measured even in degraded DNA, it has become one of the most promising molecular approaches to forensic age estimation. Previous studies have built methylation-based age predictors from blood, saliva, bone, and dental pulp, but pulp-free teeth remained a gap in the forensic toolkit.

To close that gap, the researchers prioritized ten previously reported age-associated CpG loci that had already been validated in skeletal and dental tissues. They then constructed their study around 128 sectioned, pulp-free teeth, treating the cut samples as simulated forensic specimens. Before building any predictive model, the team asked a critical methodological question: do the type of tooth or the sex of the individual distort the relationship between methylation and age? Using analysis of variance-based interaction testing, they found that neither factor significantly modified the methylation-age relationship, with P values of 0.340 for tooth type and 0.468 for sex. Adding interaction terms for either variable also failed to reduce prediction error meaningfully, with P values of 0.210 and 0.156 respectively.

That negative result carries real practical weight. If methylation patterns behaved differently in incisors than in molars, or in men than in women, forensic laboratories would need separate models for each category, multiplying complexity and fragmenting limited sample sizes. Instead, the findings support a unified model that can be applied to pulp-free teeth regardless of type or the sex of the donor, simplifying workflows in casework where every scrap of information matters.

With the marker set established, the team compared four statistical approaches for converting methylation values into age estimates: partial least squares regression, ridge regression, elastic net regression, and support vector regression with a radial basis function kernel. Each model’s hyperparameters were optimized through ten-fold cross-validation, a technique that repeatedly splits the data into training and testing subsets to prevent overfitting and to ensure that the model generalizes to samples it has never seen. Overfitting is a persistent danger in forensic model building, where datasets are modest and the temptation to squeeze out every last percentage of accuracy can produce models that collapse on real casework.

The verdict was clear. Ridge regression, a method that deliberately constrains model complexity by shrinking regression coefficients, provided the best balance between fitting the training data and generalizing to new samples. On the test set, it achieved an R-squared of 0.749, meaning it explained roughly three-quarters of the variance in chronological age, with a mean absolute error of 6.112 years. In forensic practice, an average error of about six years is considered useful, particularly for cases where no morphological indicators survive. It is not precise enough to identify an individual, but it can narrow an unidentified person’s age range substantially and corroborate or challenge other evidence.

The study also probed whether removing the pulp itself changes the methylation signal in the remaining hard tissues. In a paired analysis comparing pulp-retained and pulp-free teeth from the same individuals, the researchers found no significant difference in overall prediction error between the two specimen types. One marker, cg22507023, did differ significantly between the groups, a nuance that may merit attention in future marker selection. More broadly, a supplementary validation on paired teeth from different positions in the mouths of twelve randomly selected individuals confirmed that CpG methylation levels and predicted ages did not differ meaningfully within a single person. In other words, it should not matter much which tooth an investigator recovers; the epigenetic signal is consistent across the dentition.

The implications extend beyond routine casework. Teeth are famously durable, surviving fire, decomposition, and millennia of burial better than almost any other human tissue. They are often the only identifiable material recovered from disaster scenes, historical burials, and archaeological contexts. A method that extracts age information from pulp-free dental hard tissues could therefore assist not only forensic identification but also the study of ancient populations, provided the methylation signal survives in sufficiently old material. The technique also complements existing approaches such as Raman spectroscopy of teeth and amino acid racemization, giving investigators a growing menu of molecular options when morphology fails.

Challenges remain before the method becomes standard practice. The study was conducted on a specific population, and methylation-age relationships can vary across ancestries and environments, so independent validation in other cohorts will be essential. The mean absolute error of six years, while useful, still leaves wide uncertainty bands for very young or very old individuals, where age-related methylation changes can plateau. Laboratory costs and processing time must also fall before the approach can be deployed at scale. Nevertheless, the demonstration that sectioned, pulp-free teeth, specimens once considered nearly worthless for molecular age estimation, can yield predictions within about six years of the truth marks a genuine advance. The study was funded by the National Natural Science Foundation of China and regional science foundations, and was approved by the Ethics Committee of Basic Medical Sciences at Central South University. As epigenetic forensics matures, the humble tooth, even one emptied of everything soft and alive, is proving to be a stubborn witness that keeps telling time long after the body has fallen silent.

Subject of Research: DNA methylation-based forensic age estimation from pulp-free teeth

Article Title: Age estimation from DNA methylation in sectioned pulp-free teeth

Article References: Chen, S., Hou, L., Liu, L., Wen, D., Liang, X., Zhang, Y., Chen, B., Zhao, L., Liu, Y., & Zha, L. (2026). Age estimation from DNA methylation in sectioned pulp-free teeth. International Journal of Legal Medicine. https://doi.org/10.1007/s00414-026-04026-2

Image Credits: AI Generated

DOI: 10.1007/s00414-026-04026-2

Keywords: DNA methylation, forensic science, age estimation, teeth, epigenetic clock, CpG sites, ridge regression, forensic anthropology, dental tissue, machine learning, legal medicine, estimation

The research, conducted at the Department of Forensic Medicine in Hradec Králové in the Czech Republic, is the first systematic, statistically validated investigation of reactive splenic contraction in human hypothermia deaths. The finding itself is not new to physiology. The spleen acts as a rapidly mobilizable blood reservoir, capable of sequestering up to 200 to 250 milliliters of densely packed erythrocytes. When the sympathetic nervous system fires—during exercise, hypoxia, hemorrhage, or severe stress—the splenic capsule and vessels contract, squeezing stored red blood cells into the circulation and boosting oxygen-carrying capacity within moments. What had never been done before was to rigorously test whether this dramatic physiological response leaves a recognizable and diagnostically reliable footprint at autopsy in people who died of hypothermia.

The study team, led by Lenka Zátopková and Štěpánka Pohlová Kučerová together with senior author Petr Hejna, examined 54 fatalities attributed to accidental hypothermia and compared them against 170 randomly selected control cases representing both natural and non-natural deaths, including heart failure, cardiopulmonary failure, hanging, and suffocation. The researchers defined reactive splenic contraction by a strict macroscopic triad: an apparent reduction in splenic volume or weight relative to the decedent’s body size, wrinkling of the splenic capsule, and anemization of the red pulp on the cut surface, typically accompanied by increased firmness of the tissue. All three components had to be present simultaneously, a deliberately conservative requirement designed to maximize specificity and prevent misclassification based on isolated, nonspecific splenic changes.

The results were unambiguous. The complete contraction triad was identified in 48 of the 54 hypothermia fatalities, a prevalence of 88.9 percent, but in only 2 of the 170 controls, or 1.2 percent. This translates into a sensitivity of 88.9 percent and a specificity of 98.8 percent, with a positive predictive value of 96.0 percent, a negative predictive value of 96.6 percent, and an overall diagnostic accuracy of 96.4 percent. The statistical association between splenic contraction and fatal hypothermia was extraordinarily strong, with an odds ratio of 672.0 and a 95 percent confidence interval stretching from 131.4 to 3437.2, far beyond the threshold of statistical significance. Histological examination of the spleen in every hypothermia case confirmed the macroscopic impression of red-pulp anemization, revealing a depletion of erythrocytes within the red pulp and empty vascular lumina consistent with the wholesale mobilization of the splenic blood store.

Perhaps the most intriguing result concerns how splenic contraction relates to the established markers of hypothermia. Among the classical findings, only Wischnewsky spots showed a statistically significant association with splenic contraction. No significant relationships emerged with cold erythema, the inner knee sign, hemorrhages within core skeletal muscles, or fatty degeneration of renal tubular epithelial cells. In practical terms, this means that splenic contraction does not simply ride along with the other tissue changes of cold death. Among contraction-positive fatalities, cold erythema was absent in more than a quarter of cases, and roughly a third lacked either muscle hemorrhages or renal fatty degeneration; in six cases, both of those deeper findings were missing entirely. Splenic contraction, in other words, appears to carry independent diagnostic information rather than redundant confirmation of what other signs already show.

This independence points toward a compelling pathophysiological interpretation. Hemorrhagic and degenerative lesions such as muscle hemorrhages and renal tubular fatty change generally require prolonged survival and sustained exposure to critical cold before they develop. Splenic contraction, by contrast, reflects an immediate physiological adaptation: acute sympathetic activation, circulatory centralization shunting blood toward vital organs, and the rapid expulsion of the splenic erythrocyte reservoir into the systemic circulation. The authors therefore propose that a contracted spleen may represent one of the earliest morphofunctional responses to lethal cold stress, visible at autopsy even when the slower-developing hemorrhagic and degenerative manifestations have not yet had time to appear. Supporting this view, the mean splenic weight in contraction-positive hypothermia cases was just 76.1 grams, compared with 132.5 grams in contraction-negative hypothermia cases and 175.4 grams in controls, and the difference could not be explained by age differences between the groups.

The scientific groundwork for this conclusion was laid decades ago, though it was largely forgotten. In 1944, the German researcher Wendt demonstrated experimentally in rabbits that whole-body cooling reduced spleen volume by up to one third, with the most pronounced contraction occurring at core temperatures of 24 to 26 degrees Celsius, and that the organ regained its normal size after rewarming. Subsequent animal studies by Varman and Bernshtein corroborated the spleen’s reservoir role under hypothermic conditions, showing decreases in splenic weight and volume accompanied by rising circulating hemoglobin concentrations. Isolated human case reports followed, including Horn’s account of a 49-year-old sailor who died of cold exposure after falling overboard, and retrospective series by Gillner and Waltz and by Schneider and Klug that noted contracted spleens in a minority of hypothermia deaths. Two doctoral dissertations touched on the finding, but neither provided the prospective, controlled evidence needed to establish its diagnostic value, and major forensic textbooks continued to relegate splenic contraction to quotation marks and question marks—or to omit it entirely.

The practical implications for forensic practice are considerable. Hypothermia deaths are frequently misinterpreted at first glance, particularly when the scene suggests sudden natural death, acute intoxication, or even violence. Cold erythema and ecchymosis-like discolorations on the limbs can mimic traumatic injuries, occasionally raising suspicion of assault. In such ambiguous situations, the conditions that mimic hypothermia typically produce a congested, blood-filled spleen, whereas a contracted, pale, wrinkled spleen points decisively in the opposite direction. Because the contracted spleen is readily visible the moment the abdominal cavity is opened—sometimes alongside reverse Wischnewsky spots visible through the gastric wall—its early recognition can redirect the differential diagnosis toward hypothermia and prompt targeted dissection of the back and core musculature, inspection of the knee joints, and supplementary histological and toxicological analyses. The authors also suggest that the finding could support minimally invasive postmortem approaches, such as laparoscopic inspection of body cavities combined with postmortem imaging, where early visual cues carry exceptional value.

Modern imaging may soon extend this diagnostic reach even further. Preliminary work with postmortem computed tomography has reported elevated splenic radiodensity in hypothermia deaths, consistent with a denser, more compact parenchyma produced by contraction, and the present study’s exploratory imaging data showed compatible tendencies, though the small number of cases precluded statistical analysis and will be reported separately. The researchers are careful to emphasize the limits of their work: the study was conducted at a single institution, the macroscopic assessment was not blinded, interobserver agreement was not formally measured, and histological examination of the spleen was performed only in the hypothermia group. Splenic contraction is not pathognomonic and must never be interpreted in isolation. Nevertheless, with nearly nine out of ten hypothermia deaths showing the triad and fewer than two percent of controls doing so, the evidence strongly supports adding routine splenic assessment to the forensic autopsy protocol—transforming an organ long dismissed as morphologically unremarkable into a quiet but eloquent witness to death by cold.

Subject of Research: Validation of reactive splenic contraction as a diagnostic autopsy sign of fatal hypothermia

Article Title: Spleen contraction verified as a supportive autopsy sign of fatal hypothermia

Article References: Zátopková, L., Kučerová, Š. P., Kovařík, D., Pažinová, J., Pitková, R., Bačinová, P., Janík, M., & Hejna, P. (2026). Spleen contraction verified as a supportive autopsy sign of fatal hypothermia. International Journal of Legal Medicine. https://doi.org/10.1007/s00414-026-03984-x

Image Credits: AI Generated

DOI: 10.1007/s00414-026-03984-x

Keywords: hypothermia, forensic pathology, spleen, splenic contraction, autopsy, Wischnewsky spots, cold erythema, postmortem diagnosis, forensic medicine, blood reservoir, postmortem imaging, legal medicine

The research team conducted a narrative review with systematic elements, searching PubMed in November and December 2025 with combinations of pediatric terms such as child, infant, neonate, newborn, and adolescent and maltreatment terms such as abuse, neglect, non-accidental, and battered child, crossed with cold-related terms including hypothermia, cold injury, low body temperature, and frostbite. A supplementary non-systematic search of German-language publications in Google Scholar was added to catch literature not indexed in PubMed. Two reviewers independently screened abstracts against inclusion and exclusion criteria, excluding studies focused exclusively on adults, on clinical or therapeutic hypothermia, and on accidental cold exposure such as winter sports accidents. Out of 192 search results plus four additional German-language publications, only 22 papers ultimately made the cut, a strikingly small body of evidence for a phenomenon the authors describe as a significant threat to affected children.

To understand why cold is so dangerous to young children, the review lays out the underlying physiology in detail. Hypothermia is generally defined as a core body temperature falling below 35 degrees Celsius, although the World Health Organization uses a threshold of 36.5 degrees for newborns. Crucially, hypothermia does not require extreme environments: it can develop at ordinary room temperature if a victim is unable to take self-rescue measures and no help arrives. In water, the danger multiplies, because heat conducts 20 to 25 times faster through water than through air. The body fights back with stress hormone release, shivering, peripheral vasoconstriction, glycogen mobilization, and, in infants, heat production from brown fat tissue. Yet children lose heat faster than adults because of their larger surface-to-volume ratio, narrower range of thermoregulatory responses, and faster metabolic exhaustion. A literature review cited in the paper documented neonatal cooling rates ranging from 1.2 to a staggering 28.5 degrees Celsius per hour.

Local cold injuries follow a different but equally grim pathophysiology. Frostbite results from rapid exposure to extreme cold, producing ice crystals both inside and outside cells, and typically strikes the extremities, nose, and ears. Severity is usually graded clinically into four levels. Non-freezing cold injury develops more slowly and involves no ice crystal formation, but both forms trigger vasoconstriction, ischemia, inflammatory cellular damage, electrolyte shifts that initiate cell death, and thrombosis. The body does mount a defense known as the hunting response, in which blood vessels dilate for five to ten minutes to protect chilled tissue, but if cold exposure continues, circulation to the affected body part is shut down to preserve the core, leading to further ischemia and necrosis. The consequences can range from lasting sensory deficits to amputation and death, and both injury types demand intensive, multidisciplinary treatment.

Against this physiological backdrop, the case reports assembled by the Düsseldorf team paint a disturbing picture of how cold exposure is weaponized or allowed to harm children. Gustavson and Levitt described four cases of abuse-related pediatric hypothermia in 1996, noting that hypothermia without a consistent history should always raise suspicion of abuse. In three of their four cases, the children had been forced into cold water or held under a cold shower, leading the authors to propose inflicted cold-water immersion or near drowning as a recognized mechanism. Other reports lean toward neglect: Bernstein and colleagues described a newborn found in a dumpster, still attached to her placenta, with an initial core temperature of just 20 degrees Celsius. Knight and colleagues, reviewing pediatric neglect deaths, reported a three-year-old boy with patterned contusions found dead at home, in whom autopsy revealed hypothermia as the cause of death.

One of the review’s clearest findings is demographic. Every case of abuse- or neglect-related hypothermia identified in the literature involved infants and children aged six or younger, and only one reported case involved a child older than four, a six-year-old boy tied to a bed with an electrical cord in a cold room, deliberately prevented from escaping. This age pattern almost certainly reflects vulnerability: very young children cannot remove themselves from cold environments, cannot communicate what happened, and depend entirely on caregivers for warmth and shelter. The authors also observed that most of the affected children showed concomitant injuries indicative of physical abuse or pronounced signs of other forms of neglect, suggesting that cold exposure tends to appear in particularly severe cases of maltreatment, where caregivers who inflict one hazard are likely violating the child’s wellbeing in multiple ways at once.

The rarity of published cases, however, may be more apparent than real. The review highlights substantial underreporting: mild hypothermia resulting from abuse or neglect likely goes unrecognized in clinical settings or is deemed too insignificant to publish, and hypothermia is notoriously difficult to detect post mortem, which may allow it to escape forensic notice entirely. Publication bias further skews the record toward the most severe cases. The true extent of unreported cases is therefore likely considerable, the authors conclude, and clear generalizations from 22 publications remain difficult. Outcomes in the documented cases nonetheless span the full spectrum, from children who recovered without lasting physical effects to those requiring amputation of extremities and those who died.

Local cold injuries tell an even sparser story. The review found only isolated reports, including a 2024 case of medical child abuse in which a mother caused what the authors called cold burn injuries by repeatedly spraying her 17-month-old child’s skin with deodorant spray. Neglect-related frostbite is similarly rare in the literature, with the closest cases being newborns left outside in the cold, though the authors note that distinguishing neglect from active abuse is often legally and clinically impossible in such scenarios. A retrospective study of 46 pediatric frostbite cases by Boles and colleagues found that lack of supervision was a major risk factor and the most common cause, with the child unsupervised in 9 of 14 cases involving children up to 12 years old. The review also flags alcohol intoxication as a risk factor for hypothermia in older children and adolescents, raising questions about supervisory neglect depending on age and circumstances.

A particularly subtle finding concerns hypothermia as a secondary phenomenon rather than a direct result of cold exposure. Hypothermia is a well-known sign in severe trauma and can indicate brain injury with impaired central thermoregulation, and it appears regularly in abusive head trauma. Kennedy and colleagues, studying 520 children with traumatic brain injury, of whom 50 had abusive head trauma, found a twofold increase in unfavorable outcomes among children with secondary hypothermia. Lee and colleagues, in a retrospective study of 156 physically abused children admitted to intensive care, found that trauma-induced hypothermia was associated with poor neurological outcomes and was the second strongest risk factor for mortality, with abusive head trauma accounting for 87.1 percent of ICU admissions. This creates a diagnostic trap: secondary hypothermia from abusive head trauma could be mistaken for additional cold-related abuse, and healthcare professionals must be aware of the phenomenon to prevent false allegations.

The review closes with a warning about differential diagnosis and interpretation pitfalls. Cold-induced skin injuries can resemble thermal burns and must be distinguished from them, while cold aerosol burns, sometimes called frosties, are frequently self- or peer-inflicted among adolescents. Accidental hypothermia in newborns, whose limited thermoregulatory reserves put them at risk of pneumonia and sepsis, must also be considered, and the authors note that external factors such as homelessness, forced displacement, or war may prevent parents from protecting their children through no fault of their own. Even bruising demands careful reading, because hypothermia impairs coagulation and can produce more pronounced hematomas, while standard laboratory coagulation tests run at 37 degrees Celsius fail to capture the full extent of the dysregulation. The authors acknowledge that their findings, drawn largely from high-income countries, may not generalize elsewhere, and they call for studies with larger case numbers. Their central message stands: hypothermia and cold injuries, though rare on paper, are severe, potentially fatal manifestations of child maltreatment, and recognizing them could save the youngest and most vulnerable patients.

Subject of Research: Hypothermia and local cold injuries as rare but severe manifestations of child abuse and neglect

Article Title: Hypothermia and cold injuries – rare but severe manifestations of child abuse and neglect

Article References: Küppers, L., Ritz, S., & König, L. M. (2026). Hypothermia and cold injuries – rare but severe manifestations of child abuse and neglect. International Journal of Legal Medicine. https://doi.org/10.1007/s00414-026-03977-w

Image Credits: AI Generated

DOI: 10.1007/s00414-026-03977-w

Keywords: child abuse, child neglect, hypothermia, frostbite, cold injury, cold exposure, forensic medicine, pediatrics, abusive head trauma, thermoregulation, child protection, legal medicine

The research, led by Oguzhan Ekizoglu together with Marine Dario, Maxime Pastor, Catherine Cyteval, and Laurent Martrille, focused on a bone that has been comparatively neglected in the MRI-based age estimation literature: the proximal humerus, the ball-shaped top of the upper arm bone. While the hand-wrist region and the knee have been extensively mapped with MRI for forensic purposes, the shoulder has received far less systematic attention, particularly in Western European populations. The team retrospectively analyzed 346 shoulder MRI examinations drawn from a French cohort spanning the ages of 7.18 to 33.82 years, comprising 204 males and 142 females. Because the scans had been acquired for clinical reasons and were fully anonymized, the ethics committee of CHU Montpellier approved the study and waived the need for individual informed consent, an approach consistent with the Declaration of Helsinki.

The technical backbone of the study is the combined Schmeling–Kellinghaus staging system, a widely used framework originally developed for the medial end of the collarbone. In this system, ossification of the epiphysis, the secondary growth center that fuses to the main bone shaft after puberty, is graded in discrete stages, from an entirely cartilaginous growth plate through progressive partial fusion to complete union, with an optional subdivision for the earliest signs of fusion. On T1-weighted MRI sequences, the developing bone and the cartilaginous plate between epiphysis and metaphysis produce distinct signal intensities, allowing a trained observer to assign each shoulder to a stage. Two independent raters, blinded to each subject’s age and sex, staged all 346 scans, providing a direct test of how reproducible the method is when applied to the humeral head.

The results were strikingly consistent. The correlation between chronological age and ossification stage was strong and statistically significant across the whole sample, with a Spearman’s rank correlation coefficient of 0.878, and nearly identical values of 0.880 in males and 0.878 in females, all with p-values below 0.001. In practical terms, this means that as the humeral epiphysis progresses through its developmental stages, age increases in a tight, predictable relationship, and that this relationship holds equally well for both sexes. Statistical testing with the Mann–Whitney U method found no significant difference between males and females in the ages associated with any given stage, a finding that simplifies the application of the system, since a single set of stage descriptions may suffice for both sexes in this population.

Reliability, the measure of whether different observers reach the same conclusion from the same image, was equally impressive. Cohen’s kappa, a statistic that quantifies agreement beyond what would be expected by chance, reached 0.911 within observers rating their own scans twice and 0.890 between the two independent raters. Values above 0.8 are conventionally considered to indicate very good agreement, and these figures place the proximal humerus staging on par with, or better than, many established forensic imaging indicators. For a method that might one day inform legal decisions with life-changing consequences, this level of observer agreement is not a technical nicety but an essential prerequisite, because inconsistent readings between experts would undermine the credibility of any age estimate offered in court.

The most legally consequential finding concerns the age of majority, the threshold at 18 years that separates minors from adults in most jurisdictions. In this cohort, the youngest individual showing stage 4, defined by the presence of a visible epiphyseal scar after fusion, was 18.33 years in males and 18.06 years in females. In other words, once the epiphyseal scar is visible on the proximal humerus, the data suggest the person has very likely passed their eighteenth birthday. Notably, stage 5, the final stage of the amplified scheme, was never observed despite the cohort including individuals up to nearly 34 years of age, an observation the authors interpret in light of the persistent visibility of the epiphyseal scar in adult shoulders and the anatomical characteristics of trabecular bone in this region.

Beyond the conventional minimum-age approach, the team applied an ordinal probit transition analysis, a statistical framework borrowed from paleodemography that models the probability of moving from one ossification stage to the next as a function of age. Rather than relying on the single youngest individual observed at a stage, a figure that is highly sensitive to how many older or younger people happen to be in the sample, transition analysis yields sex-specific ages at which the probability of progression to the next stage reaches 50 percent. This provides a more statistically robust characterization of the timing of skeletal maturation and has been increasingly advocated in forensic age research, including recent work on third molars and other skeletal indicators. The authors are careful, however, to stress that both minimum-age and transition-age estimates derived from any single cohort should not be applied as forensic thresholds without independent, population-specific validation, a caution that reflects the well-documented variation in skeletal maturation across populations and socioeconomic conditions.

Perhaps the most original contribution of the study is the description of a previously unreported interpretive pitfall. The authors identified a trabecular hypointense pattern, a dark appearance of the internal bone architecture on T1-weighted images, that can potentially mimic a persistent gap between the epiphysis and the metaphysis, the appearance that would normally indicate an unfused, and therefore juvenile, growth plate. Mistaking this normal variant for an open growth plate could lead an examiner to underestimate a person’s age, with serious implications in asylum or criminal proceedings. The paper includes recommendations for recognizing this pattern, underscoring that even a highly reliable staging system depends on the examiner’s ability to distinguish true developmental landmarks from imaging artifacts and anatomical variants of the cancellous bone in the proximal humerus.

The broader context of this work is a slow but decisive shift in forensic practice. European radiation protection directives and international basic safety standards from bodies such as the International Atomic Energy Agency have sharpened the ethical and legal scrutiny applied to X-ray-based age assessments in healthy young people, pushing researchers toward MRI as a non-ionizing alternative. Previous MRI studies of the proximal humeral epiphysis, including earlier work by Ekizoglu and colleagues on Turkish cohorts and studies from China and Turkey using both 1.5 and 3.0 Tesla scanners, have demonstrated the promise of the approach, but population-specific data remained sparse for Western Europe. By extending the combined Schmeling–Kellinghaus system to a French cohort for the first time, the Montpellier team has filled an important geographic gap and shown that the staging framework travels well across populations, at least in terms of correlation strength and observer reliability.

What remains before shoulder MRI can enter routine forensic use is a matter of validation and standardization rather than proof of concept. Large, independently collected cohorts from different populations will be needed to confirm the minimum ages and transition probabilities reported here, and the newly described trabecular pitfall will need to be characterized across scanner types and imaging protocols. Cost and scanner availability also remain practical constraints, since MRI is more expensive and less widely accessible than plain radiography. Nevertheless, the study offers a compelling vision of the field’s future: a radiation-free, highly reproducible, and anatomically well-characterized method for answering one of forensic science’s most sensitive questions. For the hundreds of thousands of young people whose ages are contested each year in legal systems worldwide, the humeral head, imaged safely with magnets rather than X-rays, may soon become a key witness.

Subject of Research: Non-ionizing forensic age estimation using T1-weighted MRI of the proximal humeral epiphysis in a French cohort

Article Title: T1-Weighted MRI of the proximal humerus: Toward a non-ionizing approach to forensic age estimation in a French cohort

Article References: Ekizoglu, O., Dario, M., Pastor, M., Cyteval, C., & Martrille, L. (2026). T1-Weighted MRI of the proximal humerus: Toward a non-ionizing approach to forensic age estimation in a French cohort. International Journal of Legal Medicine. https://doi.org/10.1007/s00414-026-03996-7

Image Credits: AI Generated

DOI: 10.1007/s00414-026-03996-7

Keywords: forensic age estimation, MRI, proximal humerus, epiphyseal fusion, Schmeling-Kellinghaus staging, age of majority, transition analysis, non-ionizing imaging, forensic medicine, immigration, radiation protection, French cohort

The research team built a two-stage architecture that mirrors how a forensic expert actually works. The first stage performs segmentation: a DeepLabv3+ network, a convolutional model widely used for dense pixel-level prediction, was trained to delineate six distinct injury regions on head CT scans. DeepLabv3+ is particularly suited to this task because of its atrous spatial pyramid pooling module, or ASPP, which applies convolutions at multiple dilation rates in parallel. This allows the network to capture both fine local detail, such as the edges of a small contusion, and broader contextual information, such as the spatial relationship between a hemorrhage and the surrounding skull and brain tissue. In this study, the segmentation network achieved a mean Dice coefficient of 0.87, a standard overlap metric indicating substantial agreement between the model’s injury maps and expert annotations.

The second stage of the cascade handles classification. A ResNet18 backbone, a relatively lightweight residual convolutional network, was combined with the same ASPP module to categorize each case as either an acceleration injury or a deceleration injury. Crucially, the classifier did not operate on raw images alone. Instead, it received segmentation-guided information from the first stage, meaning the features extracted by the segmentation network were fed forward to inform the classification decision. This design choice proved decisive. In an ablation study, in which components of the system are systematically removed to measure their contribution, the macro-F1 score, a balanced measure of precision and recall across classes, improved from 0.71 to 0.82 when segmentation-guided information was included. The cascaded model as a whole attained a mean area under the receiver operating characteristic curve, or AUC, of 0.94 for injury mechanism classification, a figure that places its discriminative power in the range typically associated with expert-level medical imaging systems.

The physics underlying the classification problem explains why it is so difficult. In an acceleration injury, the head is stationary and is set into motion by an impacting force, as when a fist or a blunt object strikes a stationary skull. In a deceleration injury, the head is already moving and is abruptly stopped, as when a falling person’s head strikes the ground or a vehicle occupant’s head hits the windshield. Both scenarios produce contusions, hemorrhages, and fractures, but the distributions of these lesions differ in subtle ways rooted in inertial biomechanics. Coup injuries, arising at the site of impact, and contrecoup injuries, arising on the opposite side of the brain as it rebounds against the inner table of the skull, follow patterns that depend on the direction and magnitude of the forces involved. Distinguishing these patterns reliably has traditionally required expert judgment, and even experts can disagree when the injury presentation is ambiguous.

What makes the new system scientifically interesting is not merely its accuracy but the way its errors and successes align with established biomechanical principles. The authors validated the model externally across frontal, temporal, and occipital regions of the head, and found that its performance patterns were consistent with what biomechanics would predict. This kind of consistency check matters because deep learning models are notorious for exploiting spurious correlations, learning to recognize hospital-specific imaging artifacts or demographic confounds rather than genuine pathology. A model whose regional performance tracks known mechanisms of force transmission is more likely to have learned something real about injury physics, which is essential if the tool is ever to carry weight in a courtroom.

The study also demonstrated that the system could reliably distinguish injured brains from normal images, a prerequisite for any forensic screening application. This binary detection task may seem trivial compared with mechanism classification, but in forensic practice it is far from it. Postmortem changes, varying image quality, and the enormous anatomical variability of human brains all complicate automated analysis. The fact that the cascaded architecture maintained high accuracy on this task while simultaneously performing the harder mechanism classification suggests that the segmentation stage is doing genuine work, localizing pathology in a way that generalizes across cases rather than memorizing individual examples.

The work sits within a rapidly expanding literature on artificial intelligence in forensic imaging. Previous studies have used deep learning to detect skull fractures from curved maximum intensity projections, to identify critical findings in head CT scans, and to classify fractures on the basis of their location and number in order to discriminate between falls and blows. Each of these efforts has attacked a piece of the forensic puzzle. The present study is notable for targeting the mechanism question itself, which is arguably the most legally sensitive piece of all. It also joins a growing family of cascaded and joint segmentation-classification architectures, in which one network’s output refines another’s input, a strategy that has proven effective in medical imaging tasks ranging from cell classification to cancer diagnosis.

The practical implications for forensic practice could be substantial. Determining the manner of injury currently depends on the availability of experienced examiners, and in jurisdictions with heavy caseloads or limited forensic resources, such expertise may be scarce. An automated system that provides an objective, quantifiable assessment of injury mechanism could serve as a second opinion, flagging cases where the imaging evidence contradicts the stated history of events. The authors emphasize that the system offers a route toward objective injury mechanism determination, a phrase that carries particular weight in a field where subjective judgment has historically been the norm. Because all data underlying the findings are freely accessible to other researchers, independent groups can scrutinize and extend the work, which is exactly the kind of transparency forensic science needs.

There are, of course, caveats that temper enthusiasm. The model was trained and validated on data from a single institution, the Academy of Forensic Science in Shanghai, and although external validation across anatomical regions was performed, broader multi-center validation across different scanners, populations, and injury contexts will be needed before deployment. Legal medicine also demands more than a probability score; courts require explainable reasoning, and a 0.94 AUC does not by itself explain why a particular injury pattern indicates deceleration rather than acceleration. Future work will likely need to couple such models with interpretability techniques and with the biomechanical modeling traditions, including finite element analysis, that have long informed the field. The study was approved by the Scientific and Ethical Committee of the Academy of Forensic Science, conducted with informed consent from the families involved, and supported by Chinese national and municipal research programs, reflecting the institutional investment now flowing into computational forensics.

Nevertheless, the study marks a genuine step forward in the marriage of artificial intelligence and legal medicine. By combining a segmentation network that sees where the injury is with a classification network that reasons about how it happened, Liu and colleagues have built a system that approximates, in silicon, the diagnostic chain that forensic experts perform in their heads. The quantitative gains, from a macro-F1 of 0.71 to 0.82 with segmentation guidance, and a mean AUC of 0.94 overall, are not just incremental numbers; they demonstrate that structural information about injury location and extent is the key to unlocking mechanism inference. As such cascaded architectures mature and accumulate validation across centers and populations, the prospect of a forensic imaging tool that can tell investigators not only what happened to a brain, but how it happened, moves from aspiration toward reality, with implications for accident reconstruction, criminal investigation, and the pursuit of justice in cases where the only witness is the injured tissue itself.

Subject of Research: Deep learning classification of acceleration and deceleration craniocerebral injuries for forensic injury mechanism determination

Article Title: Classification of acceleration and deceleration craniocerebral injuries by using cascaded deep learning models

Article References: Liu, Y.-W., Tian, Z.-L., Liu, Y.-Y., Fu, E.-H., Dong, H.-W., Wan, L., Zhang, J.-H., Zou, D.-H., & Liu, N.-G. (2026). Classification of acceleration and deceleration craniocerebral injuries by using cascaded deep learning models. International Journal of Legal Medicine. https://doi.org/10.1007/s00414-026-03971-2

Image Credits: AI Generated

DOI: 10.1007/s00414-026-03971-2

Keywords: craniocerebral injury, forensic medicine, deep learning, DeepLabv3+, ResNet18, image segmentation, computed tomography, injury mechanism, acceleration injury, deceleration injury, biomechanics, legal medicine

The study, published in the International Journal of Legal Medicine, was led by Wen-tao Xia of the Academy of Forensic Science in Shanghai, together with co-first authors Ya-ning Sun and Lei Wan and their colleagues. Rather than building a system that performs a single narrow task, the team designed a multi-task learning model based on a three-dimensional residual neural network, known as 3D-ResNet18. A single network simultaneously performs three distinct jobs: it predicts the numerical age of a fracture in days, it classifies the fracture into its healing stage, and it identifies the type of fracture present. This architecture matters because the three tasks are deeply intertwined. The radiological signs that indicate a fresh fracture are the same biological processes that define its healing stage, and forcing the network to learn all three objectives at once encourages it to extract richer, more generalizable features from the image data than a single-purpose model could.

To train and validate the model, the researchers assembled a multicenter dataset of 1,848 rib fractures identified on chest CT scans. Multicenter data is a crucial ingredient in medical artificial intelligence because images acquired on different scanners, with different protocols and from different patient populations, vary in subtle ways. Models trained on data from a single hospital frequently fail when deployed elsewhere, a phenomenon that has undermined many promising AI tools. By drawing training examples from multiple sources and then testing the model on an external dataset it had never seen, the team directly addressed the generalizability question that separates laboratory demonstrations from clinically useful systems. The external test confirmed that the model’s performance held up beyond the institutions where it was developed.

The headline numbers are striking. In the fracture age prediction task, the model achieved a mean absolute error of 7.94 days, meaning that on average its estimate of how long ago a fracture occurred was off by less than eight days. In head-to-head comparison, human expert evaluation of the same fractures produced a mean absolute error of 10.65 days. The gap of more than two and a half days may sound modest, but in forensic contexts, where the difference between a fracture sustained during a documented incident and one sustained weeks earlier can determine the outcome of a case, that improvement carries real weight. For the healing stage classification task, the model reached an accuracy of 71.18 percent, and in the fracture type classification task it achieved 94.71 percent, a level of performance that approaches the reliability of specialized detection systems.

Underneath these results lies a biological story that the algorithm has effectively learned to read. When a rib breaks, the body initiates a predictable cascade of repair. In the earliest phase, hemorrhage and inflammation dominate, and the fracture line itself may be barely visible on CT. Over the following days and weeks, soft callus forms around the fracture ends, followed by hard, mineralized callus that gradually bridges the gap. Eventually the callus remodels, and the bone slowly returns toward its original architecture. Each of these stages leaves a distinct radiological signature: changes in fracture line sharpness, callus density and volume, and eventual bony union. Radiologists have long used these signs to make rough estimates of fracture age, but the transitions are gradual and their appearance varies with patient age, bone density, and health status. A convolutional neural network processing volumetric CT data can, in principle, detect patterns in these transitions that are too subtle or too distributed for consistent human recognition.

The choice of a 3D architecture is technically significant. Rib fractures are inherently three-dimensional injuries, and the diagnostic cues they present may span multiple consecutive CT slices. A two-dimensional network analyzing each slice in isolation would discard that spatial context. By feeding volumetric patches through a 3D residual network, the model can integrate information along the length of the rib and across the fracture site, capturing the full morphology of callus formation and fracture line evolution. The residual connections within ResNet18, a now-classic design that allows gradients to flow through very deep networks during training, help the model learn fine-grained image features without succumbing to the training instabilities that plagued earlier deep architectures. The study was implemented using MONAI, an open-source framework for medical deep learning, reflecting the growing standardization of tools in this field.

The forensic implications extend well beyond rib fractures alone. The cited literature surrounding this work paints a sobering picture of why objective fracture dating is needed. Radiologists have documented imaging patterns of thoracic injuries in survivors of intimate partner violence, and separate studies have examined the radiological appearance of physical elder abuse and nonaccidental injury in the elderly. Missed rib fractures on initial trauma CT scans are a recognized problem with both clinical and forensic consequences. In cases of suspected abuse, the timing of injuries is often central to the investigation: a fracture consistent with the account given by a caregiver supports one narrative, while a fracture that predates or postdates the reported incident undermines it. Prior research has even documented self-inflicted long bone fractures in insurance fraud cases, where dating an injury could expose deception. A reproducible, quantitative tool for fracture dating gives investigators and courts evidence that does not depend on which expert happened to review the scan.

The new model also builds on a clear trajectory of prior work. Earlier studies by some of the same research community demonstrated that convolutional neural networks could detect and classify rib fractures on CT with high accuracy, and a separate multicenter study showed that a radiomics-based combined model could estimate rib fracture age from CT scans. Work from other groups has applied deep learning to estimate time-since-injury in pediatric fractures, one of the most sensitive applications in forensic radiology. The present study advances this line of research by combining age regression, healing stage classification, and fracture type identification in a single multi-task framework, and by demonstrating that the combined model outperforms manual expert evaluation on the age estimation task in external testing.

Limitations and caveats remain, as they do for any medical AI system. The raw data could not be publicly shared because of participant privacy, ethical restrictions, and legal constraints, though the authors state that de-identified data are available upon reasonable request subject to ethics approval and a data use agreement. Fracture healing is known to vary with factors such as age and bone mineral density, and the model’s performance across the full spectrum of patient populations and healing conditions will require continued scrutiny as the tool moves toward practical use. The study protocol was approved by the Ethics Committee of the Academy of Forensic Science with a waiver of informed consent, and the work was supported by China’s National Key R&D Program and several national and Shanghai-level research grants.

Still, the direction of travel is unmistakable. Artificial intelligence systems have already proven themselves at detecting rib fractures, localizing them anatomically, and flagging injuries that tired human readers miss. With this study, the same technology now demonstrates that it can read the clock embedded in healing bone more precisely than the experts who trained the field. For emergency physicians triaging trauma patients, for forensic pathologists reconstructing the sequence of events in a death investigation, and for child protection teams weighing whether an injury matches the story they have been told, a day-level estimate of fracture age delivered automatically from an existing chest CT could become a routine part of the diagnostic workflow. The rib has always kept a record of when it was broken. What has changed is that a neural network can now read that record with a precision the human eye could never consistently achieve.

Subject of Research: Deep learning estimation of rib fracture age from chest CT for forensic medicine

Article Title: A multi-task learning–based deep learning model for precise estimation of rib fracture age on chest CT

Article References: Sun, Y.-N., Wan, L., Yu, X.-Y., Wang, M.-W., Zhu, K., Li, Y.-Z., Liu, T.-A., Sheng, Y.-L., & Xia, W.-T. (2026). A multi-task learning–based deep learning model for precise estimation of rib fracture age on chest CT. International Journal of Legal Medicine. https://doi.org/10.1007/s00414-026-03962-3

Image Credits: AI Generated

DOI: 10.1007/s00414-026-03962-3

Keywords: deep learning, rib fracture, forensic medicine, chest CT, fracture age estimation, multi-task learning, 3D-ResNet18, fracture healing, artificial intelligence, legal medicine, computer vision, trauma imaging

The review’s methodology tells its own story. De-Giorgio and Benedetti searched PubMed, Scopus, and Web of Science without date restrictions, combining AFM-related terms with post-mortem and autopsy keywords, and screened references manually to catch anything the databases missed. Despite the breadth of the search, only five original research articles met their strict criteria of focusing on human post-mortem materials with forensic relevance. That number is striking. AFM has been embraced enthusiastically in materials science, nanotechnology, biology, and clinical medicine for decades, yet its application to the tissues of the deceased remains in its infancy. The five qualifying studies concentrated on just four substrates: bone, skeletal muscle, tendinous tissue, and hair. This scarcity, the authors argue, reflects a combination of high instrumentation costs, small sample sizes, and substantial methodological heterogeneity that has kept the field fragmented.

Bone emerges as the most scientifically mature substrate in the review. Building on earlier animal studies that mapped elastic and inelastic deformation behavior and visualized the lacunar-canalicular network with exceptional precision, researchers turned to human autopsy samples to probe age-related changes in bone composition. In one key investigation, Milovanovic and colleagues characterized mineral grain size in bone harvested from the lateral femoral neck of two groups of women: young women aged 20 to 40 and elderly women aged 73 to 94. The older cohort exhibited larger and more heterogeneous mineral grains, a structural condition associated with reduced mechanical resistance. Crucially, chemical analysis of calcium and phosphorus levels revealed no significant differences between the groups, ruling out a major confounder and reinforcing the validity of the purely structural observations.

A follow-up study translated those structural differences into mechanical terms using AFM-based nanoindentation. The technique works by applying a controlled force to the sample surface and recording the corresponding force-distance curves, from which local mechanical properties can be computed with precision. The results showed that the differences in mineral grain size between young and elderly women translated into distinct elastic behaviors, with significantly reduced elasticity in the elderly group, an effect closely linked to higher fracture susceptibility. Although limited by sample size, the work demonstrated a reproducible methodological framework that the review’s authors believe could be extended to contexts far more directly relevant to forensic investigation, including the characterization of fracture margins in cases of suspected trauma.

It is in forensic traumatology that the review’s most provocative proposals appear. AFM topography and force mapping can quantify nanoscale roughness, microcrack patterns, and local stiffness gradients along fracture edges, capabilities that could support the differentiation between perimortem trauma, injuries inflicted around the time of death, and post-mortem damage that occurred afterward. One high-impact translational scenario involves cardiopulmonary resuscitation. CPR is well recognized to generate rib fracture patterns that can complicate post-mortem interpretation, sometimes mimicking injuries inflicted by an assailant. AFM could add objective surface and mechanical descriptors to complement macroscopic evaluation, radiology, and conventional microscopy, helping pathologists distinguish resuscitation artifacts from thoracic fractures unrelated to resuscitative maneuvers. In an era when CPR-related rib injuries are common and medico-legal disputes over their origin are frequent, such a tool could prove decisive.

Bone is also highly informative in cases of thermal exposure, including fires and explosions. Experimental work has shown that heating produces progressive physicochemical and structural modifications of bone mineral: crystallite parameters begin to change at temperatures of 500 degrees Celsius and above, while marked crystallite growth and recrystallization become prominent at higher temperatures, commonly reported around 700 degrees Celsius and beyond, with downstream effects on the stability and fragility of calcined remains. Post-mortem computed tomography, or PMCT, has already proven valuable in assessing burned victims, including the recognition of typical heat-related changes and the differentiation of heat-related fractures from traumatic ones. The review proposes a correlative framework in which PMCT provides non-destructive, case-level documentation of fracture patterns, ATR-FTIR spectroscopy and X-ray diffraction characterize temperature-related mineral and matrix transformations, and AFM targets selected regions for nanoscale topographic and nanomechanical mapping. The open question is whether thermal overprinting modifies or preserves the nanoscale features relevant to forensic interpretation, and AFM is positioned to answer it.

Soft tissue tells a different story, one in which water plays an unexpectedly large role. Studies on human iliotibial band samples investigated how variations in water content influence the mechanical properties of post-mortem soft tissues, and the findings were clear: lower water content increases the elastic modulus of the tissue. This has direct implications for anyone designing experiments on post-mortem soft tissues, since hydration status must be controlled for or at least reported. With specific regard to estimating the post-mortem interval, the elapsed time since death and one of the central questions in forensic practice, De-Giorgio and colleagues performed a preliminary investigation on skeletal muscle samples collected at different post-mortem intervals, applying AFM-based measurements of the Young’s modulus and hysteresis, the latter reflecting energy dissipation during loading cycles. Their results indicated that hysteresis is less influenced by inter-individual variability than the elastic modulus, making it a parameter worth prioritizing in future PMI-focused studies. This matters because traditional PMI methods, which rely on qualitative assessment of livor mortis, rigor mortis, and algor mortis, are fundamentally constrained by the examiner’s subjective evaluation and by intrinsic factors including age, sex, muscle mass, and ambient temperature.

Hair, remarkably durable and easy to sample, offers a third window into the post-mortem nanoworld. Mansilla and colleagues compared hair samples from contemporary autopsies with those from archaeological mummies to examine nanostructural cellular features. Compared with modern hair, the mummy samples exhibited a depletion of fatty acids and a reduction in the regularity of the cuticular scale pattern, although the scales themselves remained remarkably preserved even after centuries. The review’s authors suggest extending this analytical approach by comparing mummy hair across different historical periods and geographic regions, potentially revealing correlations between long-term environmental exposure and nanostructural hair alterations. Such a comparative framework could eventually help investigators assess whether remains are consistent with a claimed historical context or with a more recent death.

Perhaps the most forward-looking sections of the review concern substrates that have barely been touched in forensic settings. Skin, which has been extensively investigated with AFM in dermatological research on the epidermis and dermis, remains underexplored in forensic pathology, despite the fact that establishing wound vitality and age is often critical for medico-legal reconstruction and remains methodologically challenging. AFM-derived parameters such as surface topography, local stiffness mapping, and viscoelastic signatures could be investigated for their association with the post-trauma interval, provided that major confounders including hydration status, sampling site, and fixation protocol are explicitly controlled. The authors also point to dried bloodstains, where experimental studies have used AFM to estimate stain age by tracking changes in the mechanical properties of red blood cells over time, with subsequent work showing that substrate type, whether glass, metal, or ceramic, significantly affects the mechanical signature. Gunshot residue analysis represents another avenue, since AFM has already been applied to high-resolution characterization of GSR particle morphology and size distribution, with evidence that particle size distributions may vary as a function of shooting distance. Extending such analyses to peri-wound skin at entry and exit sites could reveal reproducible patterns related to muzzle-to-target distance and ammunition type. Even the subtle electrical current mark seen in electrocution victims, currently assessed mainly through histopathology and scanning electron microscopy with energy-dispersive X-ray analysis, could yield additional microfeatures to AFM’s nanomechanical mapping.

The review is candid about the obstacles standing between these visions and courtroom reality. The evidence base is constrained by small sample sizes, high instrumentation costs, and methodological heterogeneity spanning differences in sampling, preparation, AFM operational modes, and reporting standards. To support AFM as a reliable forensic tool, and ultimately strengthen its potential admissibility in court, the authors call for robust statistical reporting including appropriate summary measures, dispersion indices, and explicit assessment of reproducibility and measurement uncertainty. They also advocate multicentric research designs to increase sample representativeness and distribute financial and logistical burdens, along with standardized post-mortem imaging biobanks built on harmonized acquisition protocols to facilitate cross-study comparison and accelerate the integration of AFM-based metrics into forensic science. Bone, skeletal muscle, and hair remain the most practical starting points, being generally accessible at autopsy and collectable with limited manipulation. If those foundations are laid, the nanoscale may soon become a routine crime scene, one measured not in meters but in nanometers.

Subject of Research: Applications of atomic force microscopy to human post-mortem tissues in forensic medicine

Article Title: Atomic force microscopy in human post-mortem samples: current applications and future directions

Article References: De-Giorgio, F., & Benedetti, B. (2026). Atomic force microscopy in human post-mortem samples: current applications and future directions. International Journal of Legal Medicine. https://doi.org/10.1007/s00414-026-03976-x

Image Credits: AI Generated

DOI: 10.1007/s00414-026-03976-x

Keywords: atomic force microscopy, forensic science, post-mortem interval, forensic pathology, bone nanomechanics, nanoindentation, skeletal muscle, mummified hair, thermal bone alteration, wound age estimation, gunshot residue, legal medicine

The research, led by Alice Matheux and colleagues at the Dijon University Hospital in France, set out to define postmortem reference concentrations for amlodipine, distinguishing levels compatible with harmless therapeutic exposure from those that may have contributed to death. The team reviewed 33 autopsy cases in which amlodipine had been quantified in postmortem jugular vein blood. Crucially, the classification of each case was made on the basis of autopsy findings and circumstantial evidence, deliberately independent of the measured amlodipine concentration, to avoid circular reasoning. Cases were assigned to one of two groups: deaths that were amlodipine-related or possibly related, designated G1, and deaths clearly unrelated to the drug, designated G2. This design allowed the researchers to compare the concentration distributions of the two groups statistically and to derive tentative interpretive thresholds from the boundary between them.

The analytical backbone of the study was a validated liquid chromatography coupled with high-resolution mass spectrometry method, abbreviated LC-HRMS. This technique separates the drug from the complex matrix of decomposing blood and then identifies and quantifies it by its exact mass and fragmentation pattern, offering specificity that older immunoassay or low-resolution methods cannot match. High-resolution mass spectrometry has become a cornerstone of modern forensic toxicology precisely because it can confirm molecular identity with near-certainty, reducing the risk of false positives that could unfairly implicate a medication in someone’s death. Once concentrations were obtained, the two groups were compared using a two-tailed Student’s t-test, a standard parametric test for assessing whether the means of two datasets differ beyond what random variation would predict.

The results were striking. Of the 33 cases, 16 fell into the amlodipine-related or possibly related group, with a median postmortem blood concentration of 115 micrograms per liter, and an interquartile range running from 85.5 to 255.5 micrograms per liter. The 17 unrelated deaths in the comparison group showed a median of just 41 micrograms per liter, with a first quartile of 31 and a third quartile of 87 micrograms per liter. The difference between the groups was statistically significant, with a p-value of 0.01, meaning there is roughly a one percent probability that a gap of this size would arise by chance if the two groups were truly drawn from the same distribution. From these figures, the authors proposed practical interpretive boundaries for casework.

According to the study, postmortem whole blood concentrations below approximately 90 micrograms per liter, and more precisely below 87 micrograms per liter based on their data, appear compatible with non-toxic exposure to amlodipine. Conversely, concentrations above roughly 250 to 300 micrograms per liter, with 255.5 micrograms per liter marking the threshold in their dataset, may indicate toxicity that contributed to death. These numbers may seem abstract, but they carry enormous weight in courtrooms and inquest rooms, where a toxicologist’s interpretation of a single blood concentration can determine whether a death is ruled accidental, suicidal, natural or suspicious. Between the two thresholds lies a gray zone where interpretation must lean heavily on autopsy findings, scene investigation and medical history.

Yet the study’s most conceptually important finding concerns not the lethal threshold but the apparently benign one. In the group of amlodipine-unrelated deaths, the median concentration of 41 micrograms per liter was 2.73 times higher than the therapeutic plasma concentrations observed in living patients, which typically range from 3 to 15 micrograms per liter. In other words, people who died of entirely unrelated causes, and who were presumably taking amlodipine at prescribed doses, showed postmortem blood levels that would look alarming if judged by clinical standards. A naive reading of such a result could wrongly suggest drug toxicity in a death that had nothing to do with the medication, or conversely could erode confidence in the entire interpretive framework.

The explanation lies in a phenomenon known as postmortem redistribution. After death, the body’s circulatory and cellular systems gradually break down, and drugs that had been concentrated in tissues begin to leak back into the blood. Organs such as the lungs, liver and myocardium act as reservoirs for certain compounds, particularly basic, lipophilic drugs that bind to acidic phospholipids in cell membranes. Amlodipine, with its long half-life and large volume of distribution, is a textbook candidate for this process. As tissue stores equilibrate with blood after death, concentrations in sampled blood can rise substantially above anything the living body ever experienced. The published blood-to-plasma ratio for amlodipine, described as 1.48 or 1.7 depending on the reference, does not fully account for the observed elevation, which points toward redistribution as the dominant mechanism.

Postmortem redistribution is one of the most treacherous pitfalls in forensic toxicology, and it is not unique to amlodipine. Drugs as diverse as digoxin, tricyclic antidepressants and many antipsychotics show similar postmortem drift, which is why forensic guidelines generally recommend sampling blood from peripheral sites such as the femoral vein, where redistribution tends to be less pronounced. The Dijon team sampled jugular vein blood, a central site more susceptible to redistribution effects, which the authors acknowledge as a consideration. Autopsy findings associated with elevated amlodipine concentrations have been described in the literature, and the new dataset adds weight to the argument that direct extrapolation from living therapeutic ranges to postmortem measurements is fundamentally unreliable for this drug.

The practical implications ripple outward from the mortuary. Emergency physicians know amlodipine overdose as a feared presentation, capable of producing refractory vasodilatory shock that resists standard resuscitation and has historically required unconventional measures such as high-dose insulin-glucose therapy. Fatalities from intentional and accidental overdose, including cases involving infants and elderly patients, have been documented for decades. But proving that amlodipine caused a particular death requires reliable concentration benchmarks, and until now, forensic practitioners have had to rely on scattered case reports and small series. By providing group-based medians and quartiles from systematically classified cases, the French team has given toxicologists a defensible starting point for interpretation, one anchored in the reality of postmortem chemistry rather than the pharmacology of the living.

The authors are careful to frame their conclusions as provisional. Thirty-three cases, while a meaningful addition to the literature, remain a modest sample, and the overlap between the two groups means individual concentrations near the thresholds cannot be interpreted in isolation. They call for further large-scale, multicenter studies to establish validated postmortem reference ranges for amlodipine, ideally incorporating peripheral blood sampling, tissue measurements and standardized case classification across institutions. Until such datasets exist, the Dijon results stand as both a warning and a tool: a warning that postmortem blood levels of this common medication can be dramatically inflated by the chemistry of death itself, and a tool that, used alongside autopsy findings and clinical context, helps distinguish the patient who died despite their blood pressure pills from the one who died because of them.

Subject of Research: Postmortem amlodipine concentrations and postmortem redistribution in forensic toxicology

Article Title: Postmortem amlodipine concentrations and postmortem redistribution: a retrospective study of 33 cases

Article References: Matheux, A., Amadieu, C., Pasquet, A., François-Purssell, I., Loiseau, M., & Guerard, P. (2026). Postmortem amlodipine concentrations and postmortem redistribution: a retrospective study of 33 cases. International Journal of Legal Medicine. https://doi.org/10.1007/s00414-026-04014-6

Image Credits: AI Generated

DOI: 10.1007/s00414-026-04014-6

Keywords: amlodipine, forensic toxicology, postmortem redistribution, LC-HRMS, calcium channel blocker, autopsy, blood concentration thresholds, cardiogenic shock, pharmacokinetics, legal medicine, drug overdose, postmortem blood analysis

The core argument of the correspondence is that fitness-to-drive assessment has become fixated on the biomarker at the expense of context. Biomarkers such as carbohydrate-deficient transferrin in serum, ethyl glucuronide in hair, and phosphatidylethanol in dried blood spots are powerful indicators of alcohol consumption, and their technical performance has been extensively validated. Recent comparative work, cited in the correspondence, has examined whether serum carbohydrate-deficient transferrin or hair ethyl glucuronide offers the more reliable tool for diagnosing alcohol abuse, and protocols now exist for quantifying phosphatidylethanol in dried blood spots to monitor abstinence. Studies in driver’s license regranting cases have combined hair ethyl glucuronide, urinary ethyl glucuronide and ethyl sulfate, and capillary dried blood spot phosphatidylethanol to reconstruct a person’s drinking history. These methods are genuinely impressive pieces of analytical chemistry.

The problem, Palmiere argues, is not the chemistry but the interpretation. Biomarker concentrations do not exist in a social vacuum. Patterns of alcohol consumption vary enormously between cultures, and reference intervals and decision thresholds are typically derived from populations whose drinking habits, body compositions, and genetic backgrounds may differ from those of the person being tested. A threshold calibrated in one population may flag as abnormal a pattern of consumption that is unremarkable in another, or conversely may miss hazardous drinking that presents atypically. Without awareness of these population-level differences, a laboratory value that looks like a clean, objective answer can quietly encode the demographics of the population in which it was validated.

The same concern applies with even greater force to the screening questionnaires that sit alongside laboratory testing. The Alcohol Use Disorders Identification Test, developed under the auspices of the World Health Organization, is one of the most widely used instruments for detecting excessive drinking in primary care and in traffic medicine. Its diagnostic performance in detecting alcohol use disorders in the general population has been studied extensively, and validated translations exist, including a French version whose reliability and validity in primary care settings have been documented. But translation is not equivalence. Idioms about drinking, culturally shaped patterns of what counts as a drink, and differing willingness to disclose personal behavior to an authority figure can all shift how respondents answer. A questionnaire that performs well in its original cultural setting may systematically under- or overestimate alcohol problems in respondents from other backgrounds, and the correspondence suggests that this possibility is too rarely acknowledged in the medico-legal context.

Linguistic barriers compound the psychometric ones. Fitness-to-drive evaluations are high-stakes encounters in which the person being assessed has strong incentives to present themselves favorably, and the assessor must judge not only what is said but how it is said. When the assessed person and the assessor do not share a fluent common language, or when interpretation is improvised rather than professional, subtle cues of evasiveness or misunderstanding may reflect communication difficulty rather than deception or impairment. Screening instruments administered in a language the respondent does not command lose their validated properties entirely, yet the correspondence implies that such administrations still occur and that their results still carry weight in licensing decisions.

Structural factors add a third layer of inequity. Access to driving is not a luxury in many communities; it is a precondition of employment, caregiving, and basic participation in social life, particularly in rural and peri-urban areas poorly served by public transport. A person who loses a license after a positive screening result faces consequences that ripple through an entire household, and the ability to navigate the regranting process, including repeated testing, documentation requirements, and sometimes costly follow-up assessments, is unevenly distributed across social groups. People with precarious employment, limited health literacy, or insecure immigration status may struggle to comply with procedural demands that more advantaged drivers meet without difficulty, so identical medical findings can lead to very different practical outcomes depending on the driver’s circumstances.

The correspondence situates these concerns within a specific legal framework, drawing on the Swiss Federal Act on Road Traffic and the Swiss Ordinance on the Admission of Persons and Vehicles to Road Traffic. Swiss law, like the law of many European jurisdictions, ties license retention and regranting to medical assessments of fitness to drive, including in cases involving alcohol abuse and psychotropic substance use. Scholarly work on medicinal drugs and fitness to drive, and on psychotropic substance abuse and driving license eligibility in Italy, illustrates how widespread these assessment regimes have become. Where the law mandates a medical judgment, the fairness of the underlying instruments becomes a matter of legal as well as clinical significance, because an inequitable test translates directly into an inequitable sanction.

What would a more equitable assessment look like? The correspondence does not propose abandoning biomarkers or questionnaires, which would be neither feasible nor scientifically justified. Instead, it points toward a more contextualized practice: awareness of the population-specific limits of reference intervals and cut-offs, use of professionally translated and locally validated instruments wherever possible, professional interpretation rather than ad hoc language support, and a structural understanding of how compliance burdens fall unevenly. It also implies that assessors should treat convergent evidence, combining laboratory markers, validated questionnaires, clinical examination, and driving history, rather than elevating any single value into a verdict. The literature on combining carbohydrate-deficient transferrin with ethyl glucuronide in hair, for example, already reflects the logic that no single marker is definitive; the equity argument extends that logic from analytical uncertainty to social context.

The wider significance of the argument reaches beyond traffic medicine. Fitness-to-drive assessment is one of many medico-legal domains in which standardized instruments, developed and validated in particular populations, are applied to increasingly diverse societies. Forensic toxicology, psychiatric evaluation, and occupational medicine all face versions of the same problem: tools that are objective within their validation samples can become instruments of disparity when transplanted without scrutiny. The correspondence’s contribution is to name this problem explicitly in a domain where the stakes, both for road safety and for individual liberty and livelihood, are unusually concrete.

Published as a correspondence in the International Journal of Legal Medicine, the piece is deliberately concise, and it raises questions more than it settles them. How large are the cultural and linguistic effects on biomarker interpretation and questionnaire performance in real licensing populations? Which jurisdictions have already adapted their assessment protocols, and which continue to apply one-size-fits-all thresholds? Those empirical questions now await the field’s attention. What the correspondence makes clear is that the pursuit of road safety through medical assessment cannot be separated from the pursuit of fairness in how that assessment is designed, administered, and interpreted. A system that protects the public from genuinely impaired drivers while treating every driver, whatever their language or background, as equally credible and equally comprehensible, is the standard toward which the field is being asked to move.

Subject of Research: Equity and validity in medical fitness-to-drive assessment

Article Title: Beyond the biomarker: cultural, linguistic, and structural equity in fitness-to-drive medical assessment

Article References: Palmiere, C. (2026). Beyond the biomarker: cultural, linguistic, and structural equity in fitness-to-drive medical assessment. International Journal of Legal Medicine. https://doi.org/10.1007/s00414-026-03937-4

Image Credits: AI Generated

DOI: 10.1007/s00414-026-03937-4

Keywords: fitness to drive, legal medicine, biomarkers, alcohol abuse screening, AUDIT, carbohydrate-deficient transferrin, ethyl glucuronide, phosphatidylethanol, health equity, psychometrics, road traffic medicine, Swiss road traffic law