Downloaded from http://journals.lww.com/amjforensicmedicine by BhDMf5ePHKbH4TTImqenVICwrFV2GrwdDzk7Vsyg/bD9OoaHVaqn+NapTdbupuoL on 09/07/2020 Head and Brain Postmortem Computed Tomography—Autopsy Correlation in Hospital Deaths Serenella Serinelli, MD, PhD,* Timothy E. Richardson, DO, PhD,* Sylvie Destian, MD,† Kanish Mirchia, MD,* Michael Williams, MD,* Mariangeles Medina-Perez, MD,† and Lorenzo Gitto, MD* Abstract: The use of postmortem computed tomography (PMCT) to sup- port autopsy pathology has increased in recent decades. To some extent, PMCT has also been contemplated as a potential alternative to conven- tional postmortem examination. The purpose of this study was to investi- gate the ability of PMCT to detect specific pathologic findings in the head and brain in natural hospital deaths. We examined postmortem CT images and autopsy data from 31 sub- jects who died at SUNY (State University of New York) Upstate University Hospital between 2013 and 2018. Each subject underwent a noncontrast PMCT and a traditional autopsy. A neuroradiologist analyzed PMCT im- ages for head and brain abnormalities. The autopsies were performed by pathologists who were aware of the radiology results. In our series, PMCT was able to detect the majority of the significant space-occupying lesions, although it was not always reliable in ascertaining their nature. Postmortem computed tomography revealed findings usually challenging to detect at autopsy. Unfortunately, there were also situations in which PMCT was misleading, showing changes that were difficult to in- terpret, or that could be related to postmortem events. Therefore, we conclude PMCT should be used as an adjunct rather than a substitute to autopsy. Key Words: postmortem computed tomography, autopsy, head, brain, hospital deaths (Am J Forensic Med Pathol 2020;41: 163–175) P ostmortem computed tomography (PMCT) is considered an important tool in the investigation of death, and its comple- mentary function to autopsy has increased in the past decades. Be- cause a gradual decline in the rates of both forensic and clinical autopsies has been observed, PMCT has been contemplated not only as an aide to conventional autopsy but also as a possible al- ternative. 1,2 This could be particularly important in situations where an autopsy cannot be performed for personal, logistic, or cultural reasons. 3 Even though PMCT has been studied and applied predomi- nantly in forensic practice, 4–7 it can also be useful to detect path- ological findings that allow us to understand the causes of death in the nonforensic hospital setting. Some authors have reported that PMCT can be more accurate than clinical diagnosis for iden- tifying the cause of death in hospitalized patients. 8,9 However, in accordance with Graziani et al, 10 we believe that testing the ability of the PMCT to detect specific pathological findings is a more im- portant step than understanding the usefulness of this tool in deter- mining the cause of death. Determination of the cause of death cannot always be based solely on the images, in the same way as it cannot be based only on the gross autopsy results: there are situations in which the cause of death is highly related to the circumstances surrounding the death, to the relationship between external and internal findings, or the histological or genetic assessment of specimens collected during autopsy. Therefore, we believe that the first essential step is to understand how well the PMCT correlates with the autopsy in the identification of specific findings that can be encoun- tered in a hospital setting. To our knowledge, no study has in- vestigated the ability of PMCT to detect specific pathologic findings in the head and brain in comparison with the autopsy in natural hospital deaths. MATERIALS AND METHODS Data from subjects who died at SUNY (State University of New York) Upstate University Hospital in Syracuse (NY) between 2013 and 2018 and who underwent a PMCT followed by autopsy were collected. The cases analyzed were all natural nonforensic deaths in- volving hospitalized patients. All bodies were in supine position and were scanned in sealed body bags. Noncontrast PMCT im- ages were obtained on a Toshiba Aquilion One 320 slice CT scan- ner. Axial 1 1 mm images were acquired in bone and soft tissue for the head, and archived to the hospital Fuji PACS. Axial, sagit- tal, and coronal 3 3 mm reformations obtained on the workstation from the axial source data were sent to PACS as well. The head and brain findings were reported by a board-certified neuroradiol- ogist. The neuroradiologist was provided with information regard- ing the age of the deceased and the medical history, but was unaware of the autopsy findings. After PMCT, a full autopsy was performed by board- certified pathologists in accordance with the techniques described elsewhere. 11–13 In particular, the scalp was incised from a point behind the ear, across the vertex to the corresponding point on the other side. The soft tissues were reflected, and the skull was opened. The brain and dura mater were removed and placed in 10% formalin solution for fixation for 2 weeks or longer, to be an- alyzed by a board-certified neuropathologist for gross and histo- logical findings. The pathologists were informed of the patient's medical history and conclusions of the PMCT. A data extraction sheet was developed to record age, sex, race, cause of death, PMCT, and autopsy macroscopic/microscopic findings. Gross and histological findings from the autopsy were compared with the results obtained with the PMCT. RESULTS The 31 subjects ranged in age from 5 to 85 years (mean, 63 years). Thirteen were females, and 18 were males; 20 were white, 10 African American, and 1 Asian. The causes of death involving cerebral conditions are summa- rized in Table 1. The first column lists cases in which brain condi- tions were the only cause of death. For the cases of the second Manuscript received September 28, 2019; accepted December 10, 2019. From the *Department of Pathology, and †Department of Radiology, State University of New York-Upstate Medical University, Syracuse, NY. The authors report no conflict of interest. Reprints: Serenella Serinelli, MD, PhD, Department of Pathology, State University of New York-Upstate Medical University, 750 East Adams St., Syracuse, NY 13210. E-mail: serinels@upstate.edu. Copyright © 2020 Wolters Kluwer Health, Inc. All rights reserved. ISSN: 0195-7910/20/4103–0163 DOI: 10.1097/PAF.0000000000000538 ORIGINAL ARTICLE Am J Forensic Med Pathol • Volume 41, Number 3, September 2020 www.amjforensicmedicine.com 163 Copyright © 2020 Wolters Kluwer Health, Inc. All rights reserved.