Traumatic brain injury (TBI) is a clinical condition characterized by lesions caused by blunt or penetrating head trauma and could result in acute and chronic manifestations such as concussion, chronic traumatic encephalopathy and diffuse axonal injury. It is the main cause of death and overall disability in pediatric population worldwide, with an annual incidence of about 475,000 cases in United States in the age group of 0–14 years; its high incidence is primarily due to the anatomical structure of the children head and their partial (or absent) capability to avoid head injuries from falls [1]. The mortality rate is particularly high in the group 0–4 years, partially due to the higher incidence of abuse in that age range [2]. In the newborn, delivery head injury is the most common cause of TBI [3]; obstetric devices such vacuum extractor or forceps represent a risk factor by compressing and tractioning fetal head through vaginal canal [1].

Newborn’s head presents unique biomechanical features aimed to minimize physique resistance during natural delivery and, consequently, brain damage. The infant skull is less rigid than adult’s one and open sutures allow a higher plasticity and deformity in response to a mechanical stress [4]. Consequently, conditions like hematomas of the scalp are very common, even in uncomplicated deliveries; they can also be clinically silent or determine slightly relevant complications. Major clinical events (i.e., intraventricular hemorrhage) are more frequently seen in presence of risk factors such low birth weight or hypoxemia [5].

Among the possible classification systems of TBI, one of the most frequently used evaluates the trauma extent, identifying:

• focal damage as lacerations, hematomas, hemorrhage, focal secondary lesions due to an increase in intracranial pressure;

• diffuse/multifocal damage as diffuse axonal injury (DAI), global ischaemia, diffuse cerebral oedema.

Moreover, the peculiar action of vacuum extractors mainly leads to compressive lesions; other mechanisms of lesion, such penetration, high or low energy impact or acceleration, should be distinguished in order to hypothesize different causes to delivery.

It is also necessary to distinguish between primary mechanical injuries and secondary changes, including anoxia, oedema, hypoxia, and ischaemia; all the above-mentioned conditions play a statistically important role in the determinism of death and neurological disabilities [6, 7]. However, both macroscopic and microscopic features may be not specific since they are shared between several pathologies. Consequently, an accurate distinction between primary and secondary damage may be complex.

As previously mentioned, both forceps and vacuum extractors are performing instruments of operative vaginal delivery [8]. They are mostly indicated for cases of complicated vaginal deliveries, including prolonged second-stage labour, cord prolapse, non-reassuring fetal heart rate, intrapartum hemorrhage and exhaustion [9]. Vacuum-assisted vaginal delivery (VAVD) is performed for 1–23% of deliveries [10] and leads to an overall complication rate of about 10% [8] (mortality hasn’t never been estimated properly, but some authors report rates 1%) [11].

These events can cause neurologic alterations with a clinical presentation characterized by low Apgar score, convulsions, distress, encephalopathy, jaundice and several aftereffects of different entity; peripheric limb palsy (brachial plexus injury is the most common) can be the direct effect of a fetal malposition or disproportion.

Neonatal cranial complications which can be described after vacuum extraction can be classified as extracranial and intracranial injuries [12, 13].

Although improvements in obstetrical management and more appropriate indications for caesarean section have led to a consistent decrease in the incidence of perinatal mechanical injuries, several issues may still arise. The evaluation of vacuum-related lesions has always represented a challenge for the medical examiner in the determination of causes, means of production, timing, and in general in the assessment of medical liability.

According to that, the aim of the following analysis is the establishment of a multidisciplinary forensic methodology focused on the assessment of timing and pathophysiology of injuries, which constitutes an essential question regarding the adequacy of procedures performed by all forensic pathologists [41]. The described procedures will be illustrated through the presentation of the results emerging from inedited multicentric case series and the preliminary evidence deriving from the use of innovative immunohistochemical markers will be discussed.

Thanks to the scientific achievements of our era, the initial radiological approach to an autopsy is spreading regardless of the single case features (age, sex, race, type of injury, diagnostic or judicial intent). Moreover, such a diagnostic approach allows an objective, standardized, shared and non-perishable datum [42]. Currently, the two main methods used in the forensic field for the purpose of determining cranial injuries are post-mortem computed tomography (PMTC) and post-mortem magnetic resonance imaging (PMMRI) [43].

Regarding the first technique (PMCT), it must be stated that following the brilliant results achieved in the forensic field since its application [44, 45, 46], the use of radiographic methods has become a practical standard in many countries, including Switzerland.

In the field of childhood brain trauma, it is estimated that CT is used in up to 70% of cases related to in vivo clinical practice [47]. Despite exposure to a radiant source, this data demonstrates its undisputed usefulness for the purpose of detecting such injuries; since in the forensic field the exposure of the body to radiation is not a critical issue, the use of PMCT can be considered the diagnostic gold standard. Direct imaging allows to obtain an accurate visualization in axial or coronal planes, allowing further software reconstructions (e.g., sagittal plane).

Specifically, the sensitivity of this method for the isolated skull fractures detection is higher than MRI [48, 49]. Furthermore, PMCT of the brain allows identification of the structure involved in the hemorrhagic event (parenchyma, meningeal layers, bone structures, periosteum, soft tissues) (Fig. 1) and estimation of the pathologic process that spread within tissues in order to assess its causal role in the death event; from the point of view of expansive lesions, this technique is therefore the most suitable.

In fact, in cases of cerebral blood collection a CT scan can identify direct signs such as hyperdensity and indirect signs such as midline shift, hydrocephalus, and fractures [50]. In some specific situations, such as pneumocephalus or the presence of arterial or venous gas emboli, radiographic detection appears to be the only possibility of diagnosis, unless the autopsy investigation is conducted through investigative underwater procedures that are not confirmed in normal forensic practice [51].

In any case, phenomena of a gaseous nature always require an accurate differential diagnosis with respect to the post-mortal genesis of a putrefactive nature which, although slowed down in newborns due to poor visceral bacterial colonization, actually poses the risk of a misdiagnosis. To this end, specific methods such as the one proposed by Eggers et al. [52], based on the estimate of the gas content in the right ventricle and hepatic parenchyma to establish the extent of postmortem phenomena.

Obviously, in order to obtain the most accurate anatomical view possible of certain structures (e.g., skull base), the use of high-resolution algorithms for soft or bone tissues is useful.

Finally, another possibility offered by the radiographic technique is to proceed with contrasted angiography of the lesional areas (PMCTA). The infusion of oily media such as Angiofil [53, 54] through the modified use of heart lung machines to remedy the absence of post-mortem circulatory diffusion allows to obtain the contrast of the vascular lumen in the absence of tissue diffusion, except where permitted by easily highlighted solutions of continuity. For these purposes, PMCT is a valuable aid to correctly address autopsy procedures.

As for PMMRI, it is undeniable that the use of magnetic fields constitutes in vivo a widely used method, due to the advantages in terms of absence of emitted radiation and the high sensitivity in distinguishing between areas of different densities. In the forensic field, this technique represents a recent innovation [55]. In fact, among the factors limiting its diffusion there is the prolonged time of diagnostic acquisition, the high cost associated with the procedure, the absence of solid scientific elements in the literature and the poor knowledge of the impact of post-mortal alterations on the final report.

Regarding the last point, the inadequacy of the quantitative reconstructions currently available (T1, T2, diffusion-weighted) was observed especially in the case of hypoxic-ischemic damage and, therefore, the need to develop specific filters for the post-mortem examination [56]; various solutions have been proposed, including a linear correction model based on the post-fatal temperature of the investigated tissues [57]. However, the analysis of post-mortal changes carried out by means of magnetic resonance elastography (MRE) has allowed to estimate a latency of about 24 hours between fatal event and the onset of transformative radiological changes, offering useful insights in terms of timing window of execution [58]. Regarding the time required for the three-dimensional reconstruction of the image, which is why this technique is often used in individual body districts (rather than “panoramic” as in the case of PMTC), it is certainly true that the small size of the bodies of newborns allows a total-body acquisition not possible in adults [59].

However, PMMRI remains an extremely useful and sensitive technique for the study of soft tissues, as in cases characterized by the presence of intra- and extra-cranial hemorrhages, bruises, abrasions and lacerations of the scalp [60], axonal damage and alterations in the relationship between gray matter and white matter [61]. On the basis of a famous pilot study [62], in fact, subsequent studies have shown how this technique allows to appreciate with high sensitivity (100%) and specificity (98–100%) signs attributable to bleeding lesions [63, 64, 65] as well as suffering neuronal ischemic. Likewise, it must be underlined the PMMRI suitability to distinguish a hemorrhagic spread into phases.

As stated for PMCT, the post-mortal absence of blood circulation prevents the execution of “clinical” (arterial and venous) sequences; however, special alternative sequences (even without the use of contrast medium) have been proven to allow identification of vascular macro and micro-lesions useful for directing the autopsy procedure. Moreover, we note the existence in the literature of experiences aimed at introducing contrast media for the vascular study of various districts, including the cerebral one (PMMRI angiography); numerous compounds, including mixtures of iodinated compounds and PEG [66], would allow to obtain brilliant results in T1-weighted images through infusion procedures very similar to those foreseen for PMCTA. Ultimately, PMMRI is extremely promising in the forensic field, especially for the neonatal study; the progressive spread of this technique will allow in the immediate future to have more data and protocols for the implementation of post-fatal investigations, in order to equate (or replace) PMCT (Fig. 2, Ref. [43]).

The peculiarities related to the two techniques are shown in Table 1.

Following the execution of post-mortem radiological investigations, it is highly probable that the diagnostic suspicion is already highly oriented on the basis of the lesions found. However, given the large number of factors to be taken into consideration in a neonatal autopsy (gestational age, degree of development, malformations, perinatal ischemic suffering, state of the adnexa, intercurrent infections, maternal intake of teratogenic substances, etc.) it is necessary to use a methodical approach, aimed at investigating all possible factors determining death. Not surprisingly, this phase is the most delicate as operator-dependent: the technique identified, the correct documentation of the findings and the execution of the histological samples require the presence of a pathologist expert in the field. Furthermore, numerous scientific societies in the sector have drawn up specific protocols, such as Guidelines on Autopsy Practice edited by the Royal College of Pathologists [67].

For this purpose, head and brain macroscopic examinations should always be preceded by a scrupulous external examination of the fetus or newborn. Height and weight, crown-rump length and head circumference must be recorded, as well as the state of nutrition [50].

In particular, the following steps should be taken [68]:

(1) Measurement of head circumference and diameter to evaluate fetal growth;

(2) In a prone position, posterior foramen magnum inspection and cerebro-spinal fluid preservation [69, 70, 71];

(3) Recording of the color of the sclera, iris and conjunctiva (post-bleeding jaundice) followed by extraction of vitreous fluid. A further microscopic examination of the fundus oculi should be achieved, considering that up to 70% of children with subdural haemorrhage also have retinal haemorrhage [72].

Furthermore, the external inspection phase is of great importance to diagnose stillbirth: a live birth, for example, must be free from degenerative changes from post-mortal stay in the amniotic fluid, such as skin maceration, articular laxity or overlapping cranial bones. On the other hand, the recognition of elements such as head swelling constitute inspective birth-related vital signs. Then, a complete search for eventual malformations should be made and documented.

In this phase, any external lesion should be carefully measured, photographed, described, and sampled. Tissue sampling is particularly important to establish the nature of the lesion and its age. There could be some cases, like abdominal wall defects, in which a true differential diagnosis between malformation and birth trauma is required.

Regarding autoptic techniques, as proceeded in adults, pericranial soft tissues can be dissected by an intermastoid incision. Following this procedure, the scalp must be retracted and an accurate examination of the inner galea and temporal muscles must be assessed. In this phase, the measurement of fontanels is required because of the possible distention caused by internal foreign masses, bleedings, or hydrocephalus.

For newborns, the “butterfly” technique (full reflection of the four bony flaps) should be preferred to perform head tissue dissection [73]. Next, the head is tipped forwards and laterally, and the occipital pole is gently lifted with a finger or scalpel handle to inspect the falx and tentorium for hemorrhage and tears, so that they can be examined in this way; the state of hemispheres can be observed as well, paying attention to slightly evident subdural blood extravasations. Following the midline bone removal, the access to the longitudinal fissure allows an in situ lateral ventricles opening to assess possible bleedings. The extraction procedure shouldn’t differ from the adult standard one; in case of very low parenchymal consistence, it could be necessary the help from an assistant. Unlike other circumstances, water suspension of the cerebral parenchyma is highly unrecommended, because hemorrhagic stratifications could be washed away.

Once that skull base has been exposed, dura mater should be adequately removed, and eventual observed fractures should also be documented. It is mandatory to confirm preliminary diagnosis of fractures made during imaging investigation, as radiology reports may often denote fractures of the skull or metaphysis of long bones or ribs that, at autopsy, prove to be anatomic variants of bone formation, aberrant vascular channels, or variant sutures in cases of skull fracture. For this purpose, bone sampling should be executed routinely whenever microscopical confirmation is needed.

With the encephalic mass disposed on the table, a preliminary macroscopic evaluation allows appreciation of the volume, mass and eventual alterations of the brain, whose locations and extensions must be evaluated. Due to his typically poor consistence and the necessity of histological processing, the whole brain should be submerged for at least 2 days in a 20% formalin solution.

Next, the medical examiner will proceed to organ sectioning (frontal, transverse and/or sagittal). This phase is crucial for an appropriate routine sampling, which should include several areas of the cortex and underlying white matter (frontal, parietal, occipital, temporal lobes). It may be appropriate to sample the corpus callosum. A section of the hippocampus should be obtained, as well as a section of each level of the brain stem and cerebellum. For the spinal cord, the dura should be opened, and a crosscut should be made at 1 cm levels throughout its extent. Each major division should be sampled for histological study along with the dura. If lesions are found, more extensive sampling should be performed.

In conclusion, all the areas of histopathological interest should be documented and sampled for further microscopical examination (Fig. 3).

An adequate histopathological examination requests at least a basic knowledge of the physical and chemical mechanisms which produce the main alterations seen in such cases. The most evident alteration that a traumatic injury produces on infant brain corresponds to blood extravasation in all layers of the intra and extracranial tissues, which can be adequately assessed by traditional hematoxylin and eosin (H&E) histology or by histochemical techniques further analysed across the present section. However, the primary biomechanical force applied to brain matter contributes partly to parenchymal damage [74]: a consequence of the trauma is cerebral edema, which involves an alteration of the blood circulation and, consequently, a new pathological entity: the hypoxic-ischemic lesion [75].

Although traditional methods don’t allow a sufficient estimation of the lesional timing, they must necessarily be performed. In fact, they make it possible to establish an incontrovertible judgment of the vitality of the lesions, as well as to establish their overall extent (judgment based strictly on the number of samples taken) and to identify the underlying areas of suffering worthy of immunohistochemical study.

In order to correctly fix the samples performed, the material must be fixed in 20% formalin for at least 48 hours and then processed and embedded in paraffin. Microtome sections of about 4 $\mu$m thickness should be executed and then stained with H&E.

The microscopic observation of the samples strictly depends on the district examined. At the level of the extra-cranial soft tissues, for example, the evidence of extravasal red blood cells contained between the layers is in itself an evident sign of trauma, determined in this case by the suction produced by the vacuum device used for the extraction of the fetus. Observing the characteristics of this type of injury allows to arrive at a rough timing from the onset according to criteria already consolidated within the literature, such as those proposed by Janssen [76]. Of equal interest, due to the complex diagnosis required of the forensic pathologist, is the question relating to the viability of the lesion; even in this case, the literature in the sector proposes consolidated methods of differential diagnosis based on the integrity of the red blood cells, the presence of fibrin reaction, the degree of leukocyte infiltration and, finally, the presence of granulation tissue [77].

Conversely, the intracranial side can be affected by hemorrhagic phenomena of a different nature. Epidural hemorrhage, despite being a rare complication in the newborn for the anatomical reasons already mentioned [36], has a high specificity in relation to the use of mechanical devices for operative delivery, as the deformation of the cranial bones, made extremely malleable and scarcely unified with each other by the amplitude and immaturity of the fontanelles, it allows a notable deformation of the arterial and venous meningeal vessels.

Subdural hemorrhage is more likely to happen, and its origin must be properly sought in the dural vessels: scientific evidence shows that the origin of these blood extravasations from the rupture of the bridging veins is typical of adulthood [78]. Consequently, the use of the classification proposed by Di Maio and Dana [79] for dating this type of lesion should be used with caution, as it has not yet been verified in the neonatal setting. Subarachnoid hemorrhage, being typically linked to blunt traumas, is also not a frequent finding in vacuum-assisted delivery [80]; the reconstruction of the time of onset however follows the same principles previously stated for subdural hemorrhage. A finding of absolute importance, given the underlying clinical impact, is certainly that of cerebral contusion.

The discovery of free red blood cells within the brain parenchyma always requires the search for signs of neuronal distress. Therefore, in close relation to the timing of observation, it was observed that cerebral edema represents the most common immediate reaction to the traumatic insult, associated with neuronal vacuolization; apoptosis begins about 45–60 minutes after the insult, resulting incompatible with immediate death; after about 12–24 hours it would be possible to observe signs of nuclear swelling. After about 2 hours, it is possible to observe neutrophils diapedesis [81, 82, 83]; further timing characterizations are made possible by the application of immunohistochemical methods which will be further highlighted later. These histological criteria represent a guideline influenced by numerous inter-individual factors, including the persistence of pathological states [84], for which further investigation is necessary. Furthermore, it emerges that histological diagnosis in the case of trauma is not based exclusively on neuronal alterations but also on the integration of changes detected within different tissue components that can reflect the traumatic event (Fig. 4).

The immunohistochemical investigation currently represents the method with the greatest specificity for dating a traumatic and hypoxic-ischemic brain injury, also allowing to establish the neonatal survival time following the application of the mechanical insult.

The use of an anti-CD15 reaction (myeloid line marker), for example, reduces the detection latency of neutrophils within the lesion area to just 10 minutes; the CD68 antibody, marker of the macrophage line, is positive a few hours after the application of the traumatic insult, making it much more useful than traditional techniques (which allow to identify Erythrophages and Siderophages with a latency of a few days) [83]. A further highly sensitive cortical marker of traumatic injury, as well as positive just 2–4 hours after the event, appears to be apolipoprotein E (ApoE) expressed by neurons and neuropil affecting the entire injured hemisphere [85]. Furthermore, an important aspect is the detection of neuronal apoptosis using the TUNEL (TdT-mediated dUTP nick end labeling) technique, with a latency time of about 2 hours and a high specificity for TBI [86].

The peculiar traumatic action produced by the vacuum device, however, not causing alterations typical of traumatic injury (i.e., mass acceleration and concussion), requires investigations aimed at studying further pathological mechanisms and the use of specific markers of neuronal hypoxic-ischemic damage (HID).

In this way it is possible to observe a loss in the antibody reaction directed against the Glial Fibrillary Acid Protein (GFAP) molecule starting about 3 hours after the traumatic insult [75].

The positivity to aquaporin 4 (AQP4), the expression of which is directly linked to the development of edema [87], becomes clinically significant (above the basal level) starting from 24 hours after the trauma [75] (Fig. 5).

Based on current scientific knowledge, it is also possible to use additional markers, such as $\beta$ Amyloid Precursor Protein ($\beta$ APP), positive starting from the acute phase (3 hours) both at the neuronal and at the glial level [88, 89]. Other markers of hypoxic-ischemic damage, such as ORP150 (Oxygen regulated protein 150) and HSP70 (Heat Shock Protein 70), are of help in effectively distinguishing the areas involved in the ischemic process that show positivity already in the acute phase (48 hours); others, such as HSP90, show late positivization ($>$48 hours) [90]. A brief synopsis of the immunohistochemical reagents described in this paper is proposed in Table 2.