Objectives To validate virtual autopsy against conventional autopsy, which is the current reference standard. Materials and methods We directly compared the findings of virtual autopsy with conventional autopsy in five forensic cases including a stab wound and four traffic accidents. Results Except for non-displaced linear fractures, virtual autopsy identified bone fractures better than conventional autopsy, including fractures of the C-spine, facial bones and pelvic bones. It better detected air inside the body, i.e. pneumothorax (with or without mediastinal shift), pneumomediastinum, air embolism in the heart and vessels, and pneumocephalus. However, soft tissue ruptures or tears were not detected by virtual autopsy. In addition, virtual autopsy was falsely positive for intraventricular hemorrhage and hemoperitoneum. By using statistical calculation of the findings those could be confirmed on conventional autopsy, the sensitivity of virtual autopsy was 75% and the positive predictive value was 95%. Conclusion A number of inherent and correctable pitfalls exist in virtual autopsy, potentially limiting its use as the sole investigation in the deceased. However, it could be very useful as a screening tool prior to conventional autopsy.
