Determining the anatomical and physiological impact of penetrating wounds involves relating knowledge of anatomy with possible projectile penetration paths. Our penetration path module is being designed both for use in MediSim, to identify injuries that correlate with wounds and for use in TraumAID, to evaluate penetrating injuries to the chest and abdomen (Appendix B). Delp [10] addresses injuries to the legs.
Penetrating injuries caused by gunshot or stab wounds are modeled using a rotatable 3D graphical model of the torso with the appropriate anatomical structures. External wounds corresponding to either gunshot or stab wounds can be entered onto any location on the surface of the 3D body model. In the case of ballistic injuries, the presence of multiple entry and exit wounds complicates the process of determining which organs have been injured.
A combinatorial analysis of the surface wounds in these cases leads to various penetration path hypotheses. For any given pair of wounds, we define a wound path space as the space of possible paths a bullet may have taken from one wound to the other. Similarly, for a single external bullet wound, a wound path space describes the space of possible trajectories from the wound to a bullet lodged in the body. A penetration path hypothesis may consist of one or more wound path spaces, depending on the number of wounds.
Gunshot and stab wounds have wound path spaces that correspond to their respective regions of uncertainty. The wound path space for a gunshot wound is represented by two cones joined at their bases, with the cone apices positioned at the external wounds or at known bullet locations. The greatest uncertainty in any continuous line trajectory corresponds to the area at the adjoining bases of the cones. For a stab wound, a truncated cone is used to model the wound path space, corresponding to uncertainty in the direction of the knife blade.
Once all the possible pairings of wounds for a given set of penetrating injuries have been determined, the system establishes which anatomical structures are affected for each wound path space. This is done by checking for intersection the polygonal surfaces that make up the geometric representations of the wound path spaces and the anatomical structures. At this point, identification of (potentially) affected anatomical structures is used differently in MediSim and TraumAID. For TraumAID, the system currently highlights and lists those anatomical structures determined to lie in the wound path space. By exploring and displaying all the penetration path and injured organ possibilities for a given set of injuries, the system provides a means of visualizing the anatomy involved. This can help in bridging the gap between knowledge of the anatomy involved in a particular injury and the physiological manifestations associated with that injury. For MediSim, the a posteriori probabilities of anatomical structure involvement can be used to update the a posteriori probabilities of particular injuries. One or more high-likelihood injuries can then be attributed to the casualty.