3D Printing in Forensic Radiology

Introduction

Forensic radiology has been described as the use of imaging in both the antemortem and postmortem setting in order to detect and document various pathologies for medicolegal purposes. While considered small in comparison to other subspecialties within radiology, the field’s origins date back over 120 years within the United States (US) with the first reported case of X-rays being used as illustrative/demonstrative evidence in 1896. Since that time, advances in radiologic imaging, most notably the development and widespread adoption of cross-sectional imaging along with advanced post-processing techniques such as multiplanar reconstruction and surface rendering techniques, have been integrated into this subspecialty. The application of cross-sectional imaging in combination with advanced visualization techniques allows forensic radiology to provide unique and additional information, particularly in the postmortem setting in which a traditional autopsy may not be practical such as when presented with a severely decomposed body.

In 2010, Jeffery identified five major areas in which radiology is utilized by forensic radiologists including identification of bodies which are not identifiable by other means, firearm deaths in which the location (entry, exit wounds) and identification of residual fragments is required, nonaccidental injury and child abuse in which radiographic signatures are used to differentiate between recent and historical musculoskeletal trauma, barotrauma in which air embolisms are identified, and the identification of traumatic subarachnoid hemorrhage.

The application and integration of three-dimensional (3D) printing technologies to create accurate, precise, and realistic 3D anatomical models is thus consistent with the historical development and adaptation of new technologies in forensic radiology. In addition, victim-specific 3D anatomic modeling poses many advantages that are both complimentary to and create new opportunities in both the antemortem and postmortem settings including but not limited to the ability to sanitize gruesome human injuries thereby allowing previously inadmissible evidence to be presented to a jury, the ability to provide physical life size replication of injuries, the preservation and reproduction of human remains long after their disposal, and the ability to explain complex injury patterns that while clear to the expert radiologist remain confusing to the nonmedical professional.

The purpose of this chapter is to provide an overview of the relatively nascent application of 3D printed anatomic models in forensic radiology; describe, through illustrative examples, the various applications; and to provide context for both the strength and limitations of this technology as applied to forensic radiology.

Historical Overview

Since the discovery of the X-ray by Wilhelm Röntgen in 1895, the forensic sciences have long appreciated the importance of imaging. Lichtenstein described the first medicolegal use of X-rays in which a radiograph was first used to localize a bullet lodged between a male gunshot victim’s fibula and tibia on Christmas Eve of 1895 in Canada and then later presented as evidence in the assailant’s trial. The result of which led to a conviction and 14-year jail term. Due to the high contrast between foreign metal objects compared to bone and soft tissue, X-rays found their initial application in the investigation and location of gunshot wounds and associated shrapnel. Over the course of the next 125 years, conventional plane film radiography has been demonstrated to be an invaluable forensic tool as witnessed by its use in a diverse range of applications including malpractice investigations both as a method of documentation and protection against it, fatality investigation, the identification of victims through forensic dental identification, multiple victim fatalities, injury investigation such as abuse, and in nonviolent crimes. ^,

Most recently, the advent of cross-sectional imaging techniques, in particular computed tomography (CT) and magnetic resonance imaging (MRI), has allowed forensic radiology to noninvasively diagnose a range of injuries and crimes in both the premortem and postmortem settings. In fact, several institutions throughout the world, recognizing the utility of these imaging modalities, require whole-body imaging prior to traditional autopsy with the majority of cases undergoing postmortem CT.

Forensic Radiologic Imaging

As of 2020, autopsy remains the gold standard for establishing cause of death. There are multiple reasons for this including historical precedent—this is the primary method taught to and practiced by medical examiners, the relatively low technology and cost of performing an autopsy, and the ubiquity of the traditional autopsy suite. By contrast, imaging, and in particular cross-sectional imaging (MRI, CT), remain adjunct tools that rely on access to and interpretation of imaging data generated by them. Interpretation of radiologic findings is further complicated by the need to have access to radiologists with expertise in identifying injuries associated with death. ^, This is especially true in the postmortem setting in which injuries sustained are not typically encountered in the outpatient radiologic imaging setting (e.g., drowning) or tissue undergoes gross alteration due to the violent nature of the injury or delay from the time of expiration to imaging (e.g., drowning or putrification).

It is important to appreciate that performing an autopsy is not automatic in the determination of death or as part of a death investigation. Instead, its primary use is to determine unequivocally the true cause of death when the circumstances of the decedent’s death remain unclear based on physical examination of the body and death scene. In addition, autopsy, while hopefully establishing causality, does have limitations. For example, by its very nature it is unable to preserve the injury, traumatic fractures to bones and/or soft tissue damage due to lacerations and penetrating trauma are removed, and some pathologies such as early cardiac ischemia or intracranial vascular injuries are poorly visualized when compared to imaging findings. In this setting, imaging can and is playing an increasing role in not only providing unequivocal evidence on the cause of death thereby eliminating the need for autopsy but also to increase the quality of the pathologist’s advice to a coroner when determining if an autopsy should be performed.

As a result of the advantages and application of imaging in the forensic sciences, the use of postmortem CT and MR (PMCT, PMMR, sometimes referred to as the virtual autopsy or virtopsy ) has grown significantly. The concept of PMCT was first proposed in Israel in 1994 in response to certain religious communities as an alternative to invasive postmortem procedures. In 1997, in response to similar requests, a PMMR service was established in the United Kingdom. Since that time, further evolution of the application of PMCT has occurred resulting in the establishment of a mobile CT scanner for on scene assessment of mass casualties and more recently the development of an entire mass casualty imaging system including acquisition, reporting, secure data transfer, and storage known as Fimag. Weustink et al. also reported on the use of both PMCT and PMMR in combination with ultrasound guided biopsy as a method for minimally invasive autopsy to either substitute or augment a conventional autopsy. Hoey et al. described the use of PMCT which they referred to as CATopsy for the prediction of death in trauma patients, while Rutty et al. have described the utility of PMCT as an investigative tool in the case of an intentional neonatal upper airway obstruction. Jackowski et al. have reported on the use of PMCT for the identification of venous air embolism. Similarly, the use of both PMCT and PMMR has been described in the postmortem assessment of massive gas embolism following severe decompression sickness resulting from a diving accident. PMCT has also been widely used in injuries related to foreign metal objects including handgun and rifle injuries as well as knife wounds to multiple organs such as the aorta and brainstem. Within the medicolegal context, PMCT has also been described as a method for establishing the cause of fatal outcomes following medical intervention in the hospital setting.

An obvious question is that if PM imaging—most notably PMCT and PMMR—is to replace conventional autopsy, what is the diagnostic accuracy of this approach? Several studies have indicated that rather than replace conventional autopsy, PM imaging serves to augment information provided by either approach thereby increasing the yield of information provided by either. That is, information obtained from either is complimentary. For example, Donchin et al. reported that in a blinded study involving 25 trauma victims in which autopsy findings were compared to PMCT radiologic findings, PMCT identified 70.5% pathologic states compared to 74.8% detected by autopsy. Farkash et al. used PMCT in fatal military penetrating trauma; and although they found it useful, the authors noted that PMCT is not without limitation including limits in detecting superficial injuries of the extremities and the exact route of fragments. Levy et al. noted that in the case of gunshot-related fatalities, PMCT underestimated the number of gunshot wounds (78 detected by autopsy vs. 68 by PMCT). Similarly, Cirielli et al. described the use of PMCT in 23 postmortem investigations in which virtual autopsy matched the findings of traditional autopsy in 15 (65%) of cases, whereas traditional autopsy was needed in the remaining 8(35%) but that validity of virtual autopsy was highest for traumatic deaths.

From the studies described above it is apparent is that PM imaging diagnostic accuracy can be improved by the use of multiple PM imaging modalities. For example, Rutty et al. have reported that PMCT when combined with PM coronary angiography (PMCTA) was able to identify the cause of death, as established by autopsy, in 92% of the 241 cases reviewed. The use of multiple PM imaging modalities provides the ability to more completely asses the spectrum of injuries that may occur such as the identification of subarachnoid hemorrhage in which PMMR has proven to be more sensitive than PMCT. To further illustrate the synergistic nature of both modalities, Bollinger et al. reported on the use of PMCT and PMMR in the determination of death by hanging in a car. In this report the authors noted that data obtained from both imaging techniques identified the cause of death being cerebral hypoxia as opposed to a brainstem lesion induced by a hang-man fracture. In addition, PMMR identified soft tissue damage including hemorrhage and bleeding. Similarly, Oesterhelweg et al. described how both PMCT and PMMR provide complimentary information that in combination improve the accuracy of the diagnosis of the cause of death in the presence of laryngeal foreign bodies. Finally, Durnhofer et al. provided an excellent review of the use of both PMCT and PMMR as a virtual autopsy tool across a range of causes of death including natural causes, accident, suicide, homicide, or iatrogenic causes.

3D Anatomical Modeling in Forensic Radiology

As 3D printing and anatomical modeling technologies evolve, their application in forensic radiology is similarly growing. The most obvious of which is the use of 3D models in the courtroom as demonstrative evidence. The presentation of real evidence, particularly in situations involving a violent death that results in dismemberment or severe trauma, is often either illegal or deemed too graphic for presentation to a jury thereby biasing their ability to render an appropriate and fair verdict. In addition, actual evidence may not be permissible or practical due to the fact that it is in an advanced state of decay or putrification. Traditionally, photographic evidence is typically presented as an alternative; however, this technology is not without limitation including the fact that the subject material can undergo distortion either as a photograph or a 3D virtual model, photographic representation results in the loss of all depth information, and the presentation of an object does not necessarily provide a true representation of the physical dimensions of the object. The use of 3D printed models alleviates all of these problems by providing life size representations of the actual subject/anatomical information that is deemed admissible as evidence and sufficiently sanitized thereby rendering them acceptable to a lay jury. In addition, the use of color printing technologies allows accurate representation of multiple anatomic structures and injuries such as bone fractures, vessels, cardiac infarction, ruptured organs, and bite wounds.

Several case studies demonstrate the value of 3D printing in the courtroom and point to the emerging and increasingly important role of this technology in the judicial process. Baier et al. have described the use of micro-CT to identify the circumstances of an assault that resulted in the death of the victim. By creating a 3D model of the skull, the pathologist was able to determine the number of assault weapons and perpetrators. This evidence, along with an actual 3D model, was used as part of evidence resulting in the conviction of two defendants who were sentenced to life in prison. A second case illustrates the use of 3D printing in particularly gruesome circumstances that involved the discovery of a dismembered corpse within a suitcase submerged in a canal in the United Kingdom. In this situation, a submerged suitcase was discovered by canal workers. Suspicious of the contents, police were called followed by a CT scan at a local hospital revealing a severed head, left lower leg, and arms of an individual. Subsequent searching of the canal revealed a second suitcase with the remainder of the victim. 3D printing allowed the virtual fitting of the suitcase remains to a third set of charred remains located at the crime scene linking the two sites. While the use of 3D printing provided a more complete description of the crime, the combined evidence when presented to the suspect resulted in a full confession. The third and final example involves the presentation of 3D printed models associated with the death of a female subject due to strangulation and blunt force trauma. The perpetrator of the crime in this situation was the boyfriend and biological father of the female victim’s children. Fig. 13.1 shows the 3D models that were used as demonstrative evidence in the court case that resulted in the conviction of the suspect and a sentence of life in prison.

Outside of the courtroom, 3D printed anatomic models are being used for a variety of applications including accidental deaths and traumatic injuries. Eckhardt et al. described the use of 3D surface scanning and printing in the case of accidental death due to autoerotic asphyxia not only as a method for identifying and illustrating the method of death but also as an effective method of digitally preserving the evidence prior to alteration due to histologic sampling and autopsy. 3D anatomic models have also been used in the antemortem setting in which the initial trauma can be identified based on presurgical CT imaging of the subject. 3D printing of cranial bones in traumatic head injuries has also been reported in which 3D printed versions of the recomposed bony structures are fitted to the cadaver thereby assisting in the postmortem identification of the victim by means of facial recognition. Lastly, Barrera et al. have described the use of 3D printing in nonaccidental rib cage factures in children. When presented to members of the Child Protection Team of the author’s institution, the authors reported on the instant and spontaneous physical manipulation of the models illustrating the importance of understanding the size of the actual model.

The use of 3D models in the forensic sciences raises the question of accuracy both in terms of spatial fidelity of the data and the accurate representation of injury/pathology. Spatial fidelity is in large part determined by the resolution of the data used to create the model as well as the various processes including segmentation, smoothing, and cropping of data that are part of the 3D modeling process. While CT data are generally considered superior when compared to MR in terms of spatial fidelity, the accuracy of the final model is also determined by acquisition parameters used, for example, in the case of a postmortem CT scan. Slice thickness and in plane pixel dimensions are the ultimate limiting factors in determining the degree of spatial accuracy the final model will have. If a CT scan is performed with slice thicknesses of 5 mm or greater, it is unreasonable to envision a model in which objects less than this dimension can be accurately reproduced. Given these limitations, several reports of the spatial accuracy of 3D models in forensics have already been published indicating that their accuracy is between −0.4 mm and 1.2 mm and 2 mm.

Accurate reproduction of the object being printed is also affected by the correct identification and segmentation of the anatomic structures comprising the model. This requires additional expertise necessary to identify structures that are either distorted or corrupted due to injury or by postmortem changes. Forensic radiologists hence provide unique expertise necessary to perform this task. Post-processing operations commonly performed as part of the model creation process such as smoothing and interpolation can also distort the original imaging data further verifying the essential role of the forensic radiologist in ensuring the authenticity of the final product.