- Email: [email protected]
Non-traumatic postmortem computed tomographic (PMCT) findings of the lung
Forensic Science International 139 (2004) 39–48
Non-traumatic postmortem computed tomographic (PMCT) findings of the lung Seiji Shiotania,*, Mototsugu Kohnob, Noriyoshi Ohashib, Kentaro Yamazakic, Hidetsugu Nakayamad, Ko. Watanabee, Yuji Oyakef, Yuji Itaig,6 a
Department of Diagnostic Radiology, Tsukuba Medical Center Hospital, 1-3-1 Amakubo, Tsukuba City, Ibaraki 305-8558, Japan b Department of Emergency and Critical Care Medicine, Tsukuba Medical Center Hospital, Tsukuba, Japan c Department of Legal Medicine, Tsukuba Medical Center Hospital, Tsukuba, Japan d Department of Radiation Oncology, Tsukuba Medical Center Hospital, Tsukuba, Japan e Department of Cardiovasucular Surgery, Tsukuba Medical Center Hospital, Tsukuba, Japan f Department of Pediatrics, Tsukuba Medical Center Hospital, Tsukuba, Japan g Department of Radiology, Institute of Clinical Medicine, University of Tsukuba, Tsukuba, Japan Received 7 February 2003; accepted 15 September 2003
Abstract Purpose: We attempted to obtain postmortem computed tomographic (PMCT) images of the lung in cases of non-traumatic death and describe the results to distinguish usual postmortem findings from those of specific thoracic causes of death. Materials and methods: Our subjects were a total of 150 consecutive non-traumatic cases with cardiopulmonary arrest on arrival who were examined by CT within 2 h after certification of death between January 1993 and December 2001. PMCT images of the lung and the frequency of imaging findings (dependent density, ground glass attenuation (GGA), consolidation, pleural effusion, and endotracheal (or endobronchial) air defect) were retrospectively reviewed. Autopsy had been conducted in 16 of the cases. Results: The causes of death and frequency percentages of dependent density, GGA, consolidation, pleural effusion, and endotracheal (or endobronchial) air defect were: 91 cases of acute heart failure (AHF) (69, 66, 24, 11, 14%), 23 cases of aortic dissection (57, 39, 4, 52, 0%), 11 cases of pneumonia (18, 82, 100, 45, 27%), 23 other specified cases (52, 30, 13, 17, 9%), and two unspecified cases (0, 0, 0, 50, 0%), [total respective frequency percentages were (60, 57, 25, 21, 12%)]. Autopsy confirmed that GGA on PMCT in AHF cases corresponded to pulmonary edema. Conclusion: When PMCT of the lung shows no other shadows than dependent density, further analysis is necessary to detect the cause of death. # 2003 Elsevier Ireland Ltd. All rights reserved. Keywords: CT; Lung; Postmortem changes; Forensic radiology
1. Introduction When a patient arrives at a hospital in cardiopulmonary arrest and death is confirmed, surface inspection based only on external appearance of the corpse is utterly insufficient to detect the cause of death [1]. Since autopsy is impractical or impossible in many cases, we initiated postmortem computed tomography (PMCT) examination for detection of the * Corresponding author. Tel.: þ81-29-851-3511; fax: þ81-29-858-2773. E-mail address: [email protected] (S. Shiotani). 6 Deceased.
cause of death and as a screening method to evaluate the necessity of autopsy [2]. We previously reported that hyperattenuating aortic wall, intravascular high density fluid level (hypostasis), and dilatation of the heart were characteristic PMCT findings as postmortem changes [3–5]. Without sufficient data on the postmortem changes of the lung, it is impossible to interpret whether or not lung shadows on PMCT images are related to the cause of death. The discrimination between them is of central interest in forensic radiology [6]. However, there have been only a few published reports on PMCT [6–11] and there has been no published report describing PMCT findings of the lung in a large number of subjects.
0379-0738/$ – see front matter # 2003 Elsevier Ireland Ltd. All rights reserved. doi:10.1016/j.forsciint.2003.09.016
40
S. Shiotani et al. / Forensic Science International 139 (2004) 39–48
carcinoma (one case). Causes of death were diagnosed based on comprehensive understanding of the patient’s present illness, clinical history, PMCT findings, and autopsy findings in 16 cases (ten cases of AHF, four cases of pneumonia, one case of retained placenta, one case of pulmonary artery thromboembolism) for which we were able to obtain family permission. Among the 91 AHF cases, ten cases were pathologically confirmed as ischemic heart disease (coronary sclerosis, coronary artery thrombosis, myocardial infarction). Heart failure is a term expressing mode of death, but not the cause of death. However, we have used the term heart failure, since autopsy was performed only in a limited number of cases. After ruling out cases with obvious diagnosis of aortic dissection, SAH, and other known causes of death, the remaining 81 cases were diagnosed as AHF by referencing the clinical history such as acute onset of chest pain, dyspnea, and loss of consciousness when at least one of the following categories was satisfied: (1) abnormal electrocardiogram findings recorded in the ambulance, such as ST elevation or ventricular fibrillation [12–14] (10 cases), (2) a foamy, reddish fluid exuded from the oral cavity and the lumen of the trachea (pulmonary edema) [15,16] (12 cases), (3) a history of heart disease such as angina pectoris, myocardial infarction, or arrhythmia [12–14] (49 cases) (among these, four cases were diagnosed as AHF by the history alone), (4) coronary artery calcification on PMCT [17–19] (55 cases) (among these, 12 cases were diagnosed as AHF by coronary artery calcification alone), (5) marked left ventricular hypertrophy on PMCT (ventricular septal thickness of over 2 cm) [13,14,20,21] (12 cases). Two experienced diagnostic radiologists retrospectively reviewed the PMCT scans of the chest and judged by consensus whether each finding of dependent density, ground glass attenuation (GGA), consolidation, pleural effusion, and endotracheal (or endobronchial) air defect was positive or negative. Dependent density was diagnosed positive when subpleural increased attenuation in the dependent lung was observed [22]. Normal dependent density in the supine position disappears in the prone position; however, CT scans were obtained only in the supine position in the present study. This was because of the lack of time to
The purpose of this study was to describe the non-traumatic PMCT findings of the lung and to distinguish usual postmortem findings from those of specific thoracic causes of death.
2. Materials and methods The subject of our study was 150 consecutive non-traumatically deceased cases with cardiopulmonary arrest on arrival who were examined by CT within 2 h after certification of death between January 1993 and December 2001. Transportation time by ambulance or helicopter from the point of patient pick-up to the hospital was 30 min or less. The cases included 99 males and 51 females ranging in age from 0 to 91 years old (mean age, 63 years old). Cardiopulmonary resuscitation (CPR) was done in 133 of 150 cases including tracheal intubation, artificial respiration, cardiac massage, and infusion. CPR was not done in 17 of 150 cases because they had already exhibited early signs of death (postmortem hypostasis and rigor mortis). Whole body PMCT (CT; Accel Proceed, GE Medical Systems, Milwaukee, WI) was performed in the supine position in conventional scan mode without using helical scan mode. The chest was scanned at 1.5 cm intervals with 1.0 cm collimation and was observed at appropriate mediastinal window setting (window level (WL)/window width ðWWÞ ¼ 3040/300– 400) and lung window setting (WL/WW ¼ 600 to 700/ 1600–1700). The causes of death of the patients were acute heart failure (AHF) (91 cases), aortic dissection (23 cases), pneumonia (11 cases), other specified causes (23 cases), and unspecified causes (two cases). The other specified causes of death were choking on food (seven cases), terminal-stage cancer (one case each of esophageal, lung, and adrenal cancer), cerebral hemorrhage (three cases), subarachnoidal hemorrhage (SAH) (two cases), gastrointestinal perforation (two cases), postoperative bleeding (one case of coarctation of the aorta), retained placenta (one case), acute asthmatic attack (one case), sepsis (one case), pulmonary artery thromboembolism (one case) and rupture of hepatocellular Table 1 Postmortem CT findings of the lung and frequency percentage Postmortem CT findings
Dependent density Ground glass attenuation Consolidation Pleural effusion Endotracheal air defect
Diagnosed cause of death
Total frequency (n ¼ 150)
Acute heart failure (n ¼ 91)
Aortic dissection (n ¼ 23)
Pneumonia (n ¼ 11)
63 60 22 10 13
13 9 1 12 0
2 9 11 5 3
(69) (66) (24) (11)* (14)
(57) (39) (4)* (52)
(18)* (82) (100)* (45) (27)
Other specified (n ¼ 23)
Unspecified (n ¼ 2)
12 7 3 4 2
0 0 0 1 (50) 0
(52) (30)* (13) (17) (9)
Values in parentheses are in percent. * P < 0:05 (significant difference between frequency percentage in individual cause of death and total frequency).
90 85 37 32 18
(60) (57) (25) (21) (12)
S. Shiotani et al. / Forensic Science International 139 (2004) 39–48
perform a second PMCT on the same subject in the prone position, and the PMCT scans were quickly done between regularly scheduled CT examinations of live patients in our institution. For these reasons, we use the word ‘‘dependent density’’ comprehensively to represent a band of increased attenuation in the dependent lung. GGA was diagnosed positive when we observed a hazy increased attenuation of the lung not associated with obscuration of underlying vessels [22]. Consolidation was diagnosed positive when we observed homogeneous increase in lung parenchymal
41
attenuation attributed to obscuration of underlying vessels [22]. Endotracheal (or endobronchial) air defect was diagnosed positive when something such as food or water excluding the air was present in the trachea or main bronchi. Other findings were also described if present. For statistical evaluation, the test for the proportion of binominal or normal distribution was used. A P-value of less than 0.05 was considered to be indicative of a statistically significant difference between the frequency of each PMCT finding for the individual cause of death and total frequency.
Fig. 1. (a, b) A 36-year-old man with a history of arrhythmia who had been orally administered with antiarrhythmic drug. He suddenly became unconscious while talking with his family. Administrative autopsy finding was mild coronary sclerosis only. The cause of death was defined as acute heart failure probably due to arrhythmia. (a) PMCT at the level of lung base showed dependent density only. (b) Dependent density on PMCT corresponded to pulmonary congestion microscopically. The capillaries were tightly packed with erythrocytes. There was no exudation into the alveolar lumen.
42
S. Shiotani et al. / Forensic Science International 139 (2004) 39–48
3. Results Table 1 shows the PMCT findings of the lung and their frequency percentages. Dependent density was the most frequently observed finding. In each cause of death, there was no statistically significant difference between frequency except pneumonia and the total frequency. Dependent density in autopsy cases corresponded to pulmonary congestion pathologically (Fig. 1). Autopsies in AHF cases showed that GGA corresponded to pulmonary edema without dilated perivascular lymphatic spaces pathologically (Figs. 2 and 3). Fig. 4
shows GGA in an aortic dissection case. Seven cases of other specified causes of death showing GGA were terminal-stage cancer (two cases), SAH (two cases), choking on food (one case), cerebral hemorrhage (one case), and rupture of hepatocellular carcinoma (one case). All cases that presented pulmonary edema on physical examination showed GGA. Three cases of other specified causes of death showing consolidation were sepsis (one case), SAH (one case), and choking on food (one case). Fig. 5 shows consolidation in a SAH case. Pleural effusion in aortic dissection cases were visually higher in density than water density (hemothorax).
Fig. 2. (a, b) A 60-year-old man who suddenly felt chest discomfort while mountaineering and was taking a rest at a cottage. He became unconscious and died there. Administrative autopsy found that he had acute myocardial infarction due to thrombus of the coronary artery. (a) PMCT at the level of carina showed bilateral diffuse GGA. (b) GGA of the right upper lobe at the level of carina on PMCT corresponded to pulmonary edema microscopically. The capillaries are congested and the alveoli are filled with homogenous, eosinophilic fluid with erythrocytes.
S. Shiotani et al. / Forensic Science International 139 (2004) 39–48
43
Fig. 3. (a, b) A 47-year-old man who had extreme shortage of sleep owing to hard work and mental stress. He moaned while taking a nap and died suddenly. Administrative autopsy finding was thrombus in the coronary artery. (a) PMCT at the level of carina showed bilateral diffuse GGA. (b) GGA of the right upper lobe at the level of carina on PMCT corresponded to pulmonary edema microscopically. The alveoli are filled with an eosinophilic coagulum of protein with erythrocytes.
Four cases of other specified causes of death showing pleural effusion were terminal-stage cancer (two cases), gastrointestinal perforation (one case), and rupture of hepatocellular carcinoma (one case). Two cases of other specified cases of death showing endotracheal (or endobronchial) air defect were choking on food. In all positive endotracheal (or endobronchial) air defect cases, the defect consisted of water or soft tissue density with air bubbles. Fig. 6 shows endobronchial air defect in a pneumonia case. Other PMCT findings of the lung were one case of multiple lung metastases, two cases of chronic pulmonary inflammation (history of Mycobacterium avium–intracellu-
lare complex infection), two cases of pneumothorax (due to central venous catheterization through subcalvian vein), one case of interstitial emphysema, and five cases of rib fracture.
4. Discussion Dependent density was the most frequently observed PMCT finding of the lung. Live CT (CT obtained with the patient alive) usually shows a vertical density gradient [23,24] and normal dependent density can be interpreted as
44
S. Shiotani et al. / Forensic Science International 139 (2004) 39–48
Fig. 4. (a, b) A 75-year-old man who had history of hypertension. He became unconscious after sudden episode of chest pain. (a) PMCT of the mediastinal window setting indicated aortic dissection. The cause of massive air retention in the SVC was unknown. (b) PMCT of the lung window setting at the level of carina showed bilateral diffuse GGA.
effects of vascular distention and poor ventilation of the dependent lung [25]. In this study, CT scans were obtained in the supine position alone, making it impossible to distinguish between normal and pathological dependent density. However, the frequency of dependent density for live CT is reported to be nearly 30% [25], while that for our PMCT was 60%. Such a discrepancy has two likely causes. PMCT was obtained at the resting expiratory level, and live CT during inspiration reduces dependent density [26]. Further, hypostasis in the lung [15,27] and transudation, which is known to be more severe in the dependent lung [28], exaggerates dependent density. We performed PMCT scans within 2 h of death and the ratio of dependent density was deemed to increase with the passage of time. The frequency of depen-
dent density in pneumonia cases was lower than total frequency. With pneumonia cases, there is the possibility that consolidation and GGA obscures dependent density; therefore, dependent density was considered a common but nonspecific CT finding after death, similar to hyperattenuating aortic wall, intravascular high density fluid level (hypostasis), and dilatation of the heart on PMCT [3–5]. GGA was the second most frequently observed PMCT finding. Because that autopsies in AHF cases confirmed GGA corresponded to pulmonary edema, and all cases noted as having pulmonary edema on physical examination had GGA on PMCT, we can extrapolate that GGA on PMCT in AHF cases indicates pulmonary edema. When left-sided heart failure without concomitant right-sided heart failure
S. Shiotani et al. / Forensic Science International 139 (2004) 39–48
45
Fig. 5. A 50-year-old man who was found lying on the floor while at work. PMCT of the brain indicated subarachnoidal hemorrhage (not shown). PMCT at the level of carina showed bilateral consolidation and GGA.
occurs, the mean pulmonary filling pressure rises because of the shift of large volumes of blood from the systemic circulation into the pulmonary circulation [29]. If this pressure rises above a value approximately equal to the colloid osmotic pressure of the plasma (about 28 mmHg), fluid begins to filter out of the capillaries into the interstitial spaces and alveoli, resulting in pulmonary edema [29] and increased lung density on CT [30,31]. Chronic heart failure patients with marked elevation of left arterial pressure for long periods of time develop capacious lymphatic channels which are capable of removing large quantities of fluid from the pulmonary interstitial spaces [32]. However, in patients whose pulmonary circulation is not accustomed to the high pressure levels, pulmonary edema occurs so rapidly that it occasionally causes death in less than one-half hour [33] which is not enough time to dilate perivascular lymphatic spaces draining fluid from the lung. Pulmonary edema does not appear in all cases of death, because the mean circulatory pressure (¼pressure measured everywhere in the systemic circulation when all blood flow is stopped) is 7 mmHg after death [34], which does not exceed the colloid osmotic pressure of the plasma. One of the important questions following cardiac arrest in AHF cases is whether it was primarily due to pump failure or to an arrhythmia and the distinction cannot be certainly made [13]. Since GGA on PMCT in AHF cases is regarded as pulmonary edema, as described above, about half of AHF cases in our study could be classified as death due to pump failure. Some cases of aortic dissection and SAH also showed GGA. The frequency of aortic regurgitation following ascending aortic dissection was reported in about two-thirds of such cases [35], and sudden aortic regurgitation resulted in left ventricular heart failure. Pulmonary edema following
intracranial disease such as SAH has been known as neurogenic pulmonary edema. Weir et al. [36] reported that pulmonary edema following SAH was diagnosed in 71% of autopsy cases. We must be careful when we interpret cases of AHF and pneumonia, as it is difficult to clearly differentiate them using only PMCT images. In clinical practice, a death certificate is usually prepared based on a comprehensive understanding of the patient’s present illness, clinical history, and PMCT findings. For instance, a history saying ‘‘This apparently healthy man complained of sudden chest pain, fell to the ground and died’’ without showing aortic dissection on PMCT images would suggest AHF. However, a history of respiratory infection with signs and symptoms such as dyspnea and fever would suggest the cause of death to be pneumonia. While the pathology is the same, mild exudation into the alveoli shows GGA on CT but severe exudation shows consolidation. Therefore, the difference between GGA and consolidation on PMCT in one case is the difference in the degree of exudation. Effusions into serous cavities are said to be common postmortem findings [15]. However, in our study, the frequency of pleural effusion was lower than that of dependent density. Pleural effusion accompanied with aortic dissection was the evidence of hemothorax and was specific to the disease. Also, pleural effusion in pneumonia cases was consistent with parapneunonic effusion. Therefore, we consider pleural effusion to be one of the specific PMCT findings that suggests the cause of death. Endotracheal (or endobronchial) air defect in AHF and pneumonia cases means secretion from lungs flooded with fluid, and for that in cases of choking on food to mean food itself. When interpreting the PMCT finding of endotracheal
46
S. Shiotani et al. / Forensic Science International 139 (2004) 39–48
Fig. 6. (a, b) A 2-year-old boy who had been suffering from a severe cold for 1 week. Although he drank milk in the morning, he was found to have stopped breathing in the afternoon. (a) PMCT of the lung window setting showed bilateral diffuse GGA. (b) PMCT of the mediastinal window setting at the level of carina showed fluid filling of bilateral main bronchi (endobronchial air defect). (c) Bilateral diffuse GGA of the right upper lobe at the level of carina on PMCT corresponded to pneumonia microscopically. The alveolai are filled with polymorphonuclear leukocytes.
S. Shiotani et al. / Forensic Science International 139 (2004) 39–48
(or endobronchial) air defect, we should first rule out choking on food by examination of the history. Rib fracture, pneumothorax, and interstitial emphysema are complications accompanying CPR. Nagel et al. [37] reported that pulmonary edema associated with CPR, was seen in 46% of their autopsy cases. GGA was seen in 57% of our cases, most of which suggested pulmonary edema, similar to the results of Nagel et al. However, the ratio of GGA varied widely in each cause of death, ranging from zero to 82%. As we have described, there are various possible reasons for findings of pulmonary edema in every case, and therefore we could not clearly determine that pulmonary edema was due to CPR. Pulmonary edema in the study of Nagel et al. might have been due to the cause of death itself or to secondary changes of the causes of death, rather than from the effect of CPR. One drawback to our study was that we were able to identify a radiologic–pathologic correlation in only a limited number of cases; also, we were unable to perform high resolution CT (HRCT) although it is the more preferred method to analyze minute parenchymal changes of the lung. Since percentage of autopsies is decreasing not only in Japan but in other countries, a multi-institutional study is desirable for verification of CT findings correlating with autopsy findings. However, we believe that the large extent of our PMCT subject data provides clues for discerning the tendency of postmortem changes in the lungs and for further future analysis. In conclusion, when PMCT of the lung shows no shadows other than dependent density, further examination is necessary to determine the specific cause of death.
Acknowledgements We thank Ms. Yumiko Moriyama for her assistance in manuscript preparation.
References [1] K. Taki, H. Kato, K. Hirahara, Y. Ito, Difficulty to diagnose the cause of death in emergency room, Jpn J. Legal Med. 52 (1998) 223–226 (in Japanese). [2] S. Shiotani, M. Kono, N. Ohashi, K. Yamazaki, M. Kajitani, M. Wada, Postmortem CT for detection of the cause of death in cardiopulmonary arrest on arrival (CPAOA) patient, Radiology 217 (P) (2000) 117. [3] S. Shiotani, M. Kohno, N. Ohashi et al., Hyperattenuating aortic wall on postmortem computed tomography (PMCT), Radiat. Med. 20 (2002) 201–206. [4] S. Shiotani, M. Kohno, N. Ohashi, K. Yamazaki, Y. Itai, Postmortem intravascular high density fluid level (hypostasis): CT findings, J. Comput. Assist. Tomogr. 26 (2002) 892–893. [5] S. Shiotani, M. Kohno, N. Ohashi et al., Dilatation of the heart on postmortem computed tomography (PMCT): comparison with live CT, Radiat. Med. 21 (2003) 29–35.
47
[6] M.J. Thali, K. Yen, W. Schweitzer et al., Virtopsy, a new imaging horizon in forensic pathology: virtual autopsy by postmortem multislice computed tomograpy (MSCT) and magnetic resonance imaging (MRI)—a feasibility study, J. Forensic Sci. 48 (2003) 386–403. [7] P. Krantz, S. Holtas, Postmortem computed tomography in a diving fatality, J. Comput. Assist. Tomogr. 7 (1983) 132–134. [8] Y. Donchin, A.I. Rivkaind, J. Bar-Ziv, J. Hiss, J. Almog, M. Drescher, Utility of postmortem computed tomography in trauma victims, J. Trauma 37 (1994) 552–558. [9] S.K. Wallace, W.A. Cohen, E.J. Stern, D.T. Reay, Judicial hanging: postmortem radiographic, CT, and MR imaging features with autopsy confirmation, Radiology 193 (1994) 263–267. [10] Y. Shirakawa, K. Osaka, M. Yamashita et al., Benefit of postmortem whole body computed tomography in traffic accident victims, JJAAM 7 (1996) 273–280 (in Japanese). [11] U. Farkash, A. Scope, M. Lynn et al., Preliminary experience with postmortem computed tomography in military penetrating trauma, J. Trauma 48 (2000) 303–308. [12] I. Gordon, H.A. Schapiro, Deaths from acute neurogenic cardiovascular failure, in: I. Gordon, H.A. Schapiro (Eds.), Forensic Medicine, Churchill Livingstone, Edinburgh, 1982, pp. 150–159. [13] D.S. Pinto, M.E. Josephson, Sudden cardiac death, in: V. Fuster, R.W. Alexander, R.A. O’rourke (Eds.), The Heart, McGraw-Hill, New York, 2001, pp. 1015–1048. [14] R.J. Myerbrg, A. Castellanos, Cardiac arrest and sudden cardiac death, in: E. Braunwald, D.P. Zipes, P. Libby (Eds.), Heart Disease, vol. 1, W.B. Saunders Company, Philadelphia, 2001, pp. 890–931. [15] I. Gordon, H.A. Schapiro, Rapid deaths from interference with oxygenation of the tissues, in: I. Gordon, H.A. Schapiro (Eds.), Forensic Medicine, Churchill Livingstone, Edinburgh, 1982, pp. 80–94. [16] M.M. Givertz, W.S. Colucci, E. Braunwald, Clinical aspects of heart failure: high-output heart failure; pulmonary edema, in: E. Braunwald (Eds.), Heart Disease, vol. 1, W.B. Saunders Company, Philadelphia, 2001, pp. 534–561. [17] I. Gordon, H.A. Schapiro, The medical investigation of the cause of death: sudden, rapid and unexpected deaths in adults, children and infants, in: I. Gordon, H.A. Schapiro (Eds.), Forensic Medicine, Churchill Livingstone, Edinburgh, 1982, pp. 160–179. [18] J.J. Mahmarian, Computed tomography of the heart, in: V. Fuster, R.W. Alexander, R.A. O’rourke (Eds.), Hurst’s the heart, McGraw-Hill, New York, 2001, pp. 567–588. [19] R.M. Steiner, D.C. Levin, Radiology of the heart, in: E. Braunwald (Eds.), Heart Disease, vol. 1, W.B. Saunders Company, Philadelphia, 2001, pp. 237–272. [20] C.G. Skioldebrand, C.O. Overfors, C. Mavroudis, M.J. Lipton, Assessment of ventricular wall thickness in vivo by computed transmission tomography, Circulation 61 (1980) 960–965. [21] K. Lackner, P. Thurn, Computed tomography of the heart: ECG-gated and continuous scans, Radiology 140 (1981) 413– 420. [22] R.S. Fraser, N.L. Muller, N. Colman, P.D. Pare, Terms for CT of the lungs, in: R.S. Fraser, N.L. Muller, N. Colman, P.D. Pare (Eds.), Diagnosis of Diseases of the Chest, vol. 1, W.B. Saunders Company, Philadelphia, 1999, pp. xxxiii–xxxvi.
48
S. Shiotani et al. / Forensic Science International 139 (2004) 39–48
[23] A.B. Millar, D.M. Denison, Vertical gradients of lung density in healthy supine men, Thorax 44 (1989) 485–490. [24] L.J. Rosenblum, R.A. Mauceri, D.E. Wallenstein et al., Density patterns in the normal lung as determined by computed tomography, Radiology 137 (1980) 409–416. [25] D.R. Aberle, G. Gamsu, C.S. Ray, I.M. Feuerstein, Asbestosrelated pleural and parenchymal fibrosis: detection with highresolution CT, Radiology 166 (1988) 729–734. [26] P.J. Robinson, L. Kreel, Pulmonary tissue attenuation with computed tomography: comparison of inspiration and expiration scans, J. Comput. Assist. Tomogr. 3 (1979) 740–748. [27] C.J. Polson, The sign of death, in: C.J. Polson, D.J. Gee (Eds.), The Essentials of Forensic Medicine, Pergamon Press, Oxford, 1973, pp. 3–43. [28] M. Tagliabue, T.C. Casella, G.E. Zincone, R. Fumagalli, E. Salvini, CT and chest radiography in the evaluation of adult respiratory distress syndrome, Acta Radiologica 35 (1994) 230–234. [29] A.C. Guyton, J.E. Hall, Cardiac failure, in: A.C. Guyton, J.E. Hall (Eds.), Medical physiology, W.B. Saunders, Philadelphia, 1996, pp. 265–273. [30] L.W. Hedlund, P. Vock, E.L. Effmann, M.M. Lischiko, C.E. Putman, Hydrostatic pulmonary edema an analysis of lung
[31]
[32]
[33]
[34]
[35]
[36] [37]
density changes by computed tomography, Invest. Radiol. 19 (1984) 254–262. P. Scillia, M. Delcoix, P. Lejeune et al., Hydrostatic pulmonary edema: evaluation with thin-section CT in dogs, Radiology 211 (1999) 161–168. R.C. Schalant, E.H. Sonnenblick, Pathophysiology of heart failure, in: R.C. Schlant, R.W. Alexander (Eds.), The Heart, McGraw-Hill, New York, 1994, pp. 515–556. A.C. Guyton, J.E. Hall, Pulmonary circulation; pulmonary edema; pleural fluid, in: A.C. Guyton, J.E. Hall (Eds.), Medical physiology, W.B. Saunders Company, Philadelphia, 1996, pp. 414–421. A.C. Guyton, J.E. Hall, Cardiac output, venous return, and their regulation, in: A.C. Guyton, J.E. Hall (Eds.), Medical Physiology, W.B. Saunders Company, Philadelphia, 1996, pp. 239–252. P.T. O’Gara, R.W. DeSanctis, Aortic dissection, in: J. Loscalzo, M.A. Creager, V.J. Dzau (Eds.), Vascular Medicine, Little, Brown and Company, Boston, 1996, pp. 927–950. B.K. Weir, Pulmonary edema following fatal aneurysm rupture, J. Neurosurg. 49 (1978) 502–507. E.L. Nagel, E.G. Fine, J.P. Krischer, J.H. Davis, Complications of CPR Crit. Care. Med. 9 (1981) 424.
Non-traumatic postmortem computed tomographic (PMCT) findings of the lung Seiji Shiotania,*, Mototsugu Kohnob, Noriyoshi Ohashib, Kentaro Yamazakic, Hidetsugu Nakayamad, Ko. Watanabee, Yuji Oyakef, Yuji Itaig,6 a
Department of Diagnostic Radiology, Tsukuba Medical Center Hospital, 1-3-1 Amakubo, Tsukuba City, Ibaraki 305-8558, Japan b Department of Emergency and Critical Care Medicine, Tsukuba Medical Center Hospital, Tsukuba, Japan c Department of Legal Medicine, Tsukuba Medical Center Hospital, Tsukuba, Japan d Department of Radiation Oncology, Tsukuba Medical Center Hospital, Tsukuba, Japan e Department of Cardiovasucular Surgery, Tsukuba Medical Center Hospital, Tsukuba, Japan f Department of Pediatrics, Tsukuba Medical Center Hospital, Tsukuba, Japan g Department of Radiology, Institute of Clinical Medicine, University of Tsukuba, Tsukuba, Japan Received 7 February 2003; accepted 15 September 2003
Abstract Purpose: We attempted to obtain postmortem computed tomographic (PMCT) images of the lung in cases of non-traumatic death and describe the results to distinguish usual postmortem findings from those of specific thoracic causes of death. Materials and methods: Our subjects were a total of 150 consecutive non-traumatic cases with cardiopulmonary arrest on arrival who were examined by CT within 2 h after certification of death between January 1993 and December 2001. PMCT images of the lung and the frequency of imaging findings (dependent density, ground glass attenuation (GGA), consolidation, pleural effusion, and endotracheal (or endobronchial) air defect) were retrospectively reviewed. Autopsy had been conducted in 16 of the cases. Results: The causes of death and frequency percentages of dependent density, GGA, consolidation, pleural effusion, and endotracheal (or endobronchial) air defect were: 91 cases of acute heart failure (AHF) (69, 66, 24, 11, 14%), 23 cases of aortic dissection (57, 39, 4, 52, 0%), 11 cases of pneumonia (18, 82, 100, 45, 27%), 23 other specified cases (52, 30, 13, 17, 9%), and two unspecified cases (0, 0, 0, 50, 0%), [total respective frequency percentages were (60, 57, 25, 21, 12%)]. Autopsy confirmed that GGA on PMCT in AHF cases corresponded to pulmonary edema. Conclusion: When PMCT of the lung shows no other shadows than dependent density, further analysis is necessary to detect the cause of death. # 2003 Elsevier Ireland Ltd. All rights reserved. Keywords: CT; Lung; Postmortem changes; Forensic radiology
1. Introduction When a patient arrives at a hospital in cardiopulmonary arrest and death is confirmed, surface inspection based only on external appearance of the corpse is utterly insufficient to detect the cause of death [1]. Since autopsy is impractical or impossible in many cases, we initiated postmortem computed tomography (PMCT) examination for detection of the * Corresponding author. Tel.: þ81-29-851-3511; fax: þ81-29-858-2773. E-mail address: [email protected] (S. Shiotani). 6 Deceased.
cause of death and as a screening method to evaluate the necessity of autopsy [2]. We previously reported that hyperattenuating aortic wall, intravascular high density fluid level (hypostasis), and dilatation of the heart were characteristic PMCT findings as postmortem changes [3–5]. Without sufficient data on the postmortem changes of the lung, it is impossible to interpret whether or not lung shadows on PMCT images are related to the cause of death. The discrimination between them is of central interest in forensic radiology [6]. However, there have been only a few published reports on PMCT [6–11] and there has been no published report describing PMCT findings of the lung in a large number of subjects.
0379-0738/$ – see front matter # 2003 Elsevier Ireland Ltd. All rights reserved. doi:10.1016/j.forsciint.2003.09.016
40
S. Shiotani et al. / Forensic Science International 139 (2004) 39–48
carcinoma (one case). Causes of death were diagnosed based on comprehensive understanding of the patient’s present illness, clinical history, PMCT findings, and autopsy findings in 16 cases (ten cases of AHF, four cases of pneumonia, one case of retained placenta, one case of pulmonary artery thromboembolism) for which we were able to obtain family permission. Among the 91 AHF cases, ten cases were pathologically confirmed as ischemic heart disease (coronary sclerosis, coronary artery thrombosis, myocardial infarction). Heart failure is a term expressing mode of death, but not the cause of death. However, we have used the term heart failure, since autopsy was performed only in a limited number of cases. After ruling out cases with obvious diagnosis of aortic dissection, SAH, and other known causes of death, the remaining 81 cases were diagnosed as AHF by referencing the clinical history such as acute onset of chest pain, dyspnea, and loss of consciousness when at least one of the following categories was satisfied: (1) abnormal electrocardiogram findings recorded in the ambulance, such as ST elevation or ventricular fibrillation [12–14] (10 cases), (2) a foamy, reddish fluid exuded from the oral cavity and the lumen of the trachea (pulmonary edema) [15,16] (12 cases), (3) a history of heart disease such as angina pectoris, myocardial infarction, or arrhythmia [12–14] (49 cases) (among these, four cases were diagnosed as AHF by the history alone), (4) coronary artery calcification on PMCT [17–19] (55 cases) (among these, 12 cases were diagnosed as AHF by coronary artery calcification alone), (5) marked left ventricular hypertrophy on PMCT (ventricular septal thickness of over 2 cm) [13,14,20,21] (12 cases). Two experienced diagnostic radiologists retrospectively reviewed the PMCT scans of the chest and judged by consensus whether each finding of dependent density, ground glass attenuation (GGA), consolidation, pleural effusion, and endotracheal (or endobronchial) air defect was positive or negative. Dependent density was diagnosed positive when subpleural increased attenuation in the dependent lung was observed [22]. Normal dependent density in the supine position disappears in the prone position; however, CT scans were obtained only in the supine position in the present study. This was because of the lack of time to
The purpose of this study was to describe the non-traumatic PMCT findings of the lung and to distinguish usual postmortem findings from those of specific thoracic causes of death.
2. Materials and methods The subject of our study was 150 consecutive non-traumatically deceased cases with cardiopulmonary arrest on arrival who were examined by CT within 2 h after certification of death between January 1993 and December 2001. Transportation time by ambulance or helicopter from the point of patient pick-up to the hospital was 30 min or less. The cases included 99 males and 51 females ranging in age from 0 to 91 years old (mean age, 63 years old). Cardiopulmonary resuscitation (CPR) was done in 133 of 150 cases including tracheal intubation, artificial respiration, cardiac massage, and infusion. CPR was not done in 17 of 150 cases because they had already exhibited early signs of death (postmortem hypostasis and rigor mortis). Whole body PMCT (CT; Accel Proceed, GE Medical Systems, Milwaukee, WI) was performed in the supine position in conventional scan mode without using helical scan mode. The chest was scanned at 1.5 cm intervals with 1.0 cm collimation and was observed at appropriate mediastinal window setting (window level (WL)/window width ðWWÞ ¼ 3040/300– 400) and lung window setting (WL/WW ¼ 600 to 700/ 1600–1700). The causes of death of the patients were acute heart failure (AHF) (91 cases), aortic dissection (23 cases), pneumonia (11 cases), other specified causes (23 cases), and unspecified causes (two cases). The other specified causes of death were choking on food (seven cases), terminal-stage cancer (one case each of esophageal, lung, and adrenal cancer), cerebral hemorrhage (three cases), subarachnoidal hemorrhage (SAH) (two cases), gastrointestinal perforation (two cases), postoperative bleeding (one case of coarctation of the aorta), retained placenta (one case), acute asthmatic attack (one case), sepsis (one case), pulmonary artery thromboembolism (one case) and rupture of hepatocellular Table 1 Postmortem CT findings of the lung and frequency percentage Postmortem CT findings
Dependent density Ground glass attenuation Consolidation Pleural effusion Endotracheal air defect
Diagnosed cause of death
Total frequency (n ¼ 150)
Acute heart failure (n ¼ 91)
Aortic dissection (n ¼ 23)
Pneumonia (n ¼ 11)
63 60 22 10 13
13 9 1 12 0
2 9 11 5 3
(69) (66) (24) (11)* (14)
(57) (39) (4)* (52)
(18)* (82) (100)* (45) (27)
Other specified (n ¼ 23)
Unspecified (n ¼ 2)
12 7 3 4 2
0 0 0 1 (50) 0
(52) (30)* (13) (17) (9)
Values in parentheses are in percent. * P < 0:05 (significant difference between frequency percentage in individual cause of death and total frequency).
90 85 37 32 18
(60) (57) (25) (21) (12)
S. Shiotani et al. / Forensic Science International 139 (2004) 39–48
perform a second PMCT on the same subject in the prone position, and the PMCT scans were quickly done between regularly scheduled CT examinations of live patients in our institution. For these reasons, we use the word ‘‘dependent density’’ comprehensively to represent a band of increased attenuation in the dependent lung. GGA was diagnosed positive when we observed a hazy increased attenuation of the lung not associated with obscuration of underlying vessels [22]. Consolidation was diagnosed positive when we observed homogeneous increase in lung parenchymal
41
attenuation attributed to obscuration of underlying vessels [22]. Endotracheal (or endobronchial) air defect was diagnosed positive when something such as food or water excluding the air was present in the trachea or main bronchi. Other findings were also described if present. For statistical evaluation, the test for the proportion of binominal or normal distribution was used. A P-value of less than 0.05 was considered to be indicative of a statistically significant difference between the frequency of each PMCT finding for the individual cause of death and total frequency.
Fig. 1. (a, b) A 36-year-old man with a history of arrhythmia who had been orally administered with antiarrhythmic drug. He suddenly became unconscious while talking with his family. Administrative autopsy finding was mild coronary sclerosis only. The cause of death was defined as acute heart failure probably due to arrhythmia. (a) PMCT at the level of lung base showed dependent density only. (b) Dependent density on PMCT corresponded to pulmonary congestion microscopically. The capillaries were tightly packed with erythrocytes. There was no exudation into the alveolar lumen.
42
S. Shiotani et al. / Forensic Science International 139 (2004) 39–48
3. Results Table 1 shows the PMCT findings of the lung and their frequency percentages. Dependent density was the most frequently observed finding. In each cause of death, there was no statistically significant difference between frequency except pneumonia and the total frequency. Dependent density in autopsy cases corresponded to pulmonary congestion pathologically (Fig. 1). Autopsies in AHF cases showed that GGA corresponded to pulmonary edema without dilated perivascular lymphatic spaces pathologically (Figs. 2 and 3). Fig. 4
shows GGA in an aortic dissection case. Seven cases of other specified causes of death showing GGA were terminal-stage cancer (two cases), SAH (two cases), choking on food (one case), cerebral hemorrhage (one case), and rupture of hepatocellular carcinoma (one case). All cases that presented pulmonary edema on physical examination showed GGA. Three cases of other specified causes of death showing consolidation were sepsis (one case), SAH (one case), and choking on food (one case). Fig. 5 shows consolidation in a SAH case. Pleural effusion in aortic dissection cases were visually higher in density than water density (hemothorax).
Fig. 2. (a, b) A 60-year-old man who suddenly felt chest discomfort while mountaineering and was taking a rest at a cottage. He became unconscious and died there. Administrative autopsy found that he had acute myocardial infarction due to thrombus of the coronary artery. (a) PMCT at the level of carina showed bilateral diffuse GGA. (b) GGA of the right upper lobe at the level of carina on PMCT corresponded to pulmonary edema microscopically. The capillaries are congested and the alveoli are filled with homogenous, eosinophilic fluid with erythrocytes.
S. Shiotani et al. / Forensic Science International 139 (2004) 39–48
43
Fig. 3. (a, b) A 47-year-old man who had extreme shortage of sleep owing to hard work and mental stress. He moaned while taking a nap and died suddenly. Administrative autopsy finding was thrombus in the coronary artery. (a) PMCT at the level of carina showed bilateral diffuse GGA. (b) GGA of the right upper lobe at the level of carina on PMCT corresponded to pulmonary edema microscopically. The alveoli are filled with an eosinophilic coagulum of protein with erythrocytes.
Four cases of other specified causes of death showing pleural effusion were terminal-stage cancer (two cases), gastrointestinal perforation (one case), and rupture of hepatocellular carcinoma (one case). Two cases of other specified cases of death showing endotracheal (or endobronchial) air defect were choking on food. In all positive endotracheal (or endobronchial) air defect cases, the defect consisted of water or soft tissue density with air bubbles. Fig. 6 shows endobronchial air defect in a pneumonia case. Other PMCT findings of the lung were one case of multiple lung metastases, two cases of chronic pulmonary inflammation (history of Mycobacterium avium–intracellu-
lare complex infection), two cases of pneumothorax (due to central venous catheterization through subcalvian vein), one case of interstitial emphysema, and five cases of rib fracture.
4. Discussion Dependent density was the most frequently observed PMCT finding of the lung. Live CT (CT obtained with the patient alive) usually shows a vertical density gradient [23,24] and normal dependent density can be interpreted as
44
S. Shiotani et al. / Forensic Science International 139 (2004) 39–48
Fig. 4. (a, b) A 75-year-old man who had history of hypertension. He became unconscious after sudden episode of chest pain. (a) PMCT of the mediastinal window setting indicated aortic dissection. The cause of massive air retention in the SVC was unknown. (b) PMCT of the lung window setting at the level of carina showed bilateral diffuse GGA.
effects of vascular distention and poor ventilation of the dependent lung [25]. In this study, CT scans were obtained in the supine position alone, making it impossible to distinguish between normal and pathological dependent density. However, the frequency of dependent density for live CT is reported to be nearly 30% [25], while that for our PMCT was 60%. Such a discrepancy has two likely causes. PMCT was obtained at the resting expiratory level, and live CT during inspiration reduces dependent density [26]. Further, hypostasis in the lung [15,27] and transudation, which is known to be more severe in the dependent lung [28], exaggerates dependent density. We performed PMCT scans within 2 h of death and the ratio of dependent density was deemed to increase with the passage of time. The frequency of depen-
dent density in pneumonia cases was lower than total frequency. With pneumonia cases, there is the possibility that consolidation and GGA obscures dependent density; therefore, dependent density was considered a common but nonspecific CT finding after death, similar to hyperattenuating aortic wall, intravascular high density fluid level (hypostasis), and dilatation of the heart on PMCT [3–5]. GGA was the second most frequently observed PMCT finding. Because that autopsies in AHF cases confirmed GGA corresponded to pulmonary edema, and all cases noted as having pulmonary edema on physical examination had GGA on PMCT, we can extrapolate that GGA on PMCT in AHF cases indicates pulmonary edema. When left-sided heart failure without concomitant right-sided heart failure
S. Shiotani et al. / Forensic Science International 139 (2004) 39–48
45
Fig. 5. A 50-year-old man who was found lying on the floor while at work. PMCT of the brain indicated subarachnoidal hemorrhage (not shown). PMCT at the level of carina showed bilateral consolidation and GGA.
occurs, the mean pulmonary filling pressure rises because of the shift of large volumes of blood from the systemic circulation into the pulmonary circulation [29]. If this pressure rises above a value approximately equal to the colloid osmotic pressure of the plasma (about 28 mmHg), fluid begins to filter out of the capillaries into the interstitial spaces and alveoli, resulting in pulmonary edema [29] and increased lung density on CT [30,31]. Chronic heart failure patients with marked elevation of left arterial pressure for long periods of time develop capacious lymphatic channels which are capable of removing large quantities of fluid from the pulmonary interstitial spaces [32]. However, in patients whose pulmonary circulation is not accustomed to the high pressure levels, pulmonary edema occurs so rapidly that it occasionally causes death in less than one-half hour [33] which is not enough time to dilate perivascular lymphatic spaces draining fluid from the lung. Pulmonary edema does not appear in all cases of death, because the mean circulatory pressure (¼pressure measured everywhere in the systemic circulation when all blood flow is stopped) is 7 mmHg after death [34], which does not exceed the colloid osmotic pressure of the plasma. One of the important questions following cardiac arrest in AHF cases is whether it was primarily due to pump failure or to an arrhythmia and the distinction cannot be certainly made [13]. Since GGA on PMCT in AHF cases is regarded as pulmonary edema, as described above, about half of AHF cases in our study could be classified as death due to pump failure. Some cases of aortic dissection and SAH also showed GGA. The frequency of aortic regurgitation following ascending aortic dissection was reported in about two-thirds of such cases [35], and sudden aortic regurgitation resulted in left ventricular heart failure. Pulmonary edema following
intracranial disease such as SAH has been known as neurogenic pulmonary edema. Weir et al. [36] reported that pulmonary edema following SAH was diagnosed in 71% of autopsy cases. We must be careful when we interpret cases of AHF and pneumonia, as it is difficult to clearly differentiate them using only PMCT images. In clinical practice, a death certificate is usually prepared based on a comprehensive understanding of the patient’s present illness, clinical history, and PMCT findings. For instance, a history saying ‘‘This apparently healthy man complained of sudden chest pain, fell to the ground and died’’ without showing aortic dissection on PMCT images would suggest AHF. However, a history of respiratory infection with signs and symptoms such as dyspnea and fever would suggest the cause of death to be pneumonia. While the pathology is the same, mild exudation into the alveoli shows GGA on CT but severe exudation shows consolidation. Therefore, the difference between GGA and consolidation on PMCT in one case is the difference in the degree of exudation. Effusions into serous cavities are said to be common postmortem findings [15]. However, in our study, the frequency of pleural effusion was lower than that of dependent density. Pleural effusion accompanied with aortic dissection was the evidence of hemothorax and was specific to the disease. Also, pleural effusion in pneumonia cases was consistent with parapneunonic effusion. Therefore, we consider pleural effusion to be one of the specific PMCT findings that suggests the cause of death. Endotracheal (or endobronchial) air defect in AHF and pneumonia cases means secretion from lungs flooded with fluid, and for that in cases of choking on food to mean food itself. When interpreting the PMCT finding of endotracheal
46
S. Shiotani et al. / Forensic Science International 139 (2004) 39–48
Fig. 6. (a, b) A 2-year-old boy who had been suffering from a severe cold for 1 week. Although he drank milk in the morning, he was found to have stopped breathing in the afternoon. (a) PMCT of the lung window setting showed bilateral diffuse GGA. (b) PMCT of the mediastinal window setting at the level of carina showed fluid filling of bilateral main bronchi (endobronchial air defect). (c) Bilateral diffuse GGA of the right upper lobe at the level of carina on PMCT corresponded to pneumonia microscopically. The alveolai are filled with polymorphonuclear leukocytes.
S. Shiotani et al. / Forensic Science International 139 (2004) 39–48
(or endobronchial) air defect, we should first rule out choking on food by examination of the history. Rib fracture, pneumothorax, and interstitial emphysema are complications accompanying CPR. Nagel et al. [37] reported that pulmonary edema associated with CPR, was seen in 46% of their autopsy cases. GGA was seen in 57% of our cases, most of which suggested pulmonary edema, similar to the results of Nagel et al. However, the ratio of GGA varied widely in each cause of death, ranging from zero to 82%. As we have described, there are various possible reasons for findings of pulmonary edema in every case, and therefore we could not clearly determine that pulmonary edema was due to CPR. Pulmonary edema in the study of Nagel et al. might have been due to the cause of death itself or to secondary changes of the causes of death, rather than from the effect of CPR. One drawback to our study was that we were able to identify a radiologic–pathologic correlation in only a limited number of cases; also, we were unable to perform high resolution CT (HRCT) although it is the more preferred method to analyze minute parenchymal changes of the lung. Since percentage of autopsies is decreasing not only in Japan but in other countries, a multi-institutional study is desirable for verification of CT findings correlating with autopsy findings. However, we believe that the large extent of our PMCT subject data provides clues for discerning the tendency of postmortem changes in the lungs and for further future analysis. In conclusion, when PMCT of the lung shows no shadows other than dependent density, further examination is necessary to determine the specific cause of death.
Acknowledgements We thank Ms. Yumiko Moriyama for her assistance in manuscript preparation.
References [1] K. Taki, H. Kato, K. Hirahara, Y. Ito, Difficulty to diagnose the cause of death in emergency room, Jpn J. Legal Med. 52 (1998) 223–226 (in Japanese). [2] S. Shiotani, M. Kono, N. Ohashi, K. Yamazaki, M. Kajitani, M. Wada, Postmortem CT for detection of the cause of death in cardiopulmonary arrest on arrival (CPAOA) patient, Radiology 217 (P) (2000) 117. [3] S. Shiotani, M. Kohno, N. Ohashi et al., Hyperattenuating aortic wall on postmortem computed tomography (PMCT), Radiat. Med. 20 (2002) 201–206. [4] S. Shiotani, M. Kohno, N. Ohashi, K. Yamazaki, Y. Itai, Postmortem intravascular high density fluid level (hypostasis): CT findings, J. Comput. Assist. Tomogr. 26 (2002) 892–893. [5] S. Shiotani, M. Kohno, N. Ohashi et al., Dilatation of the heart on postmortem computed tomography (PMCT): comparison with live CT, Radiat. Med. 21 (2003) 29–35.
47
[6] M.J. Thali, K. Yen, W. Schweitzer et al., Virtopsy, a new imaging horizon in forensic pathology: virtual autopsy by postmortem multislice computed tomograpy (MSCT) and magnetic resonance imaging (MRI)—a feasibility study, J. Forensic Sci. 48 (2003) 386–403. [7] P. Krantz, S. Holtas, Postmortem computed tomography in a diving fatality, J. Comput. Assist. Tomogr. 7 (1983) 132–134. [8] Y. Donchin, A.I. Rivkaind, J. Bar-Ziv, J. Hiss, J. Almog, M. Drescher, Utility of postmortem computed tomography in trauma victims, J. Trauma 37 (1994) 552–558. [9] S.K. Wallace, W.A. Cohen, E.J. Stern, D.T. Reay, Judicial hanging: postmortem radiographic, CT, and MR imaging features with autopsy confirmation, Radiology 193 (1994) 263–267. [10] Y. Shirakawa, K. Osaka, M. Yamashita et al., Benefit of postmortem whole body computed tomography in traffic accident victims, JJAAM 7 (1996) 273–280 (in Japanese). [11] U. Farkash, A. Scope, M. Lynn et al., Preliminary experience with postmortem computed tomography in military penetrating trauma, J. Trauma 48 (2000) 303–308. [12] I. Gordon, H.A. Schapiro, Deaths from acute neurogenic cardiovascular failure, in: I. Gordon, H.A. Schapiro (Eds.), Forensic Medicine, Churchill Livingstone, Edinburgh, 1982, pp. 150–159. [13] D.S. Pinto, M.E. Josephson, Sudden cardiac death, in: V. Fuster, R.W. Alexander, R.A. O’rourke (Eds.), The Heart, McGraw-Hill, New York, 2001, pp. 1015–1048. [14] R.J. Myerbrg, A. Castellanos, Cardiac arrest and sudden cardiac death, in: E. Braunwald, D.P. Zipes, P. Libby (Eds.), Heart Disease, vol. 1, W.B. Saunders Company, Philadelphia, 2001, pp. 890–931. [15] I. Gordon, H.A. Schapiro, Rapid deaths from interference with oxygenation of the tissues, in: I. Gordon, H.A. Schapiro (Eds.), Forensic Medicine, Churchill Livingstone, Edinburgh, 1982, pp. 80–94. [16] M.M. Givertz, W.S. Colucci, E. Braunwald, Clinical aspects of heart failure: high-output heart failure; pulmonary edema, in: E. Braunwald (Eds.), Heart Disease, vol. 1, W.B. Saunders Company, Philadelphia, 2001, pp. 534–561. [17] I. Gordon, H.A. Schapiro, The medical investigation of the cause of death: sudden, rapid and unexpected deaths in adults, children and infants, in: I. Gordon, H.A. Schapiro (Eds.), Forensic Medicine, Churchill Livingstone, Edinburgh, 1982, pp. 160–179. [18] J.J. Mahmarian, Computed tomography of the heart, in: V. Fuster, R.W. Alexander, R.A. O’rourke (Eds.), Hurst’s the heart, McGraw-Hill, New York, 2001, pp. 567–588. [19] R.M. Steiner, D.C. Levin, Radiology of the heart, in: E. Braunwald (Eds.), Heart Disease, vol. 1, W.B. Saunders Company, Philadelphia, 2001, pp. 237–272. [20] C.G. Skioldebrand, C.O. Overfors, C. Mavroudis, M.J. Lipton, Assessment of ventricular wall thickness in vivo by computed transmission tomography, Circulation 61 (1980) 960–965. [21] K. Lackner, P. Thurn, Computed tomography of the heart: ECG-gated and continuous scans, Radiology 140 (1981) 413– 420. [22] R.S. Fraser, N.L. Muller, N. Colman, P.D. Pare, Terms for CT of the lungs, in: R.S. Fraser, N.L. Muller, N. Colman, P.D. Pare (Eds.), Diagnosis of Diseases of the Chest, vol. 1, W.B. Saunders Company, Philadelphia, 1999, pp. xxxiii–xxxvi.
48
S. Shiotani et al. / Forensic Science International 139 (2004) 39–48
[23] A.B. Millar, D.M. Denison, Vertical gradients of lung density in healthy supine men, Thorax 44 (1989) 485–490. [24] L.J. Rosenblum, R.A. Mauceri, D.E. Wallenstein et al., Density patterns in the normal lung as determined by computed tomography, Radiology 137 (1980) 409–416. [25] D.R. Aberle, G. Gamsu, C.S. Ray, I.M. Feuerstein, Asbestosrelated pleural and parenchymal fibrosis: detection with highresolution CT, Radiology 166 (1988) 729–734. [26] P.J. Robinson, L. Kreel, Pulmonary tissue attenuation with computed tomography: comparison of inspiration and expiration scans, J. Comput. Assist. Tomogr. 3 (1979) 740–748. [27] C.J. Polson, The sign of death, in: C.J. Polson, D.J. Gee (Eds.), The Essentials of Forensic Medicine, Pergamon Press, Oxford, 1973, pp. 3–43. [28] M. Tagliabue, T.C. Casella, G.E. Zincone, R. Fumagalli, E. Salvini, CT and chest radiography in the evaluation of adult respiratory distress syndrome, Acta Radiologica 35 (1994) 230–234. [29] A.C. Guyton, J.E. Hall, Cardiac failure, in: A.C. Guyton, J.E. Hall (Eds.), Medical physiology, W.B. Saunders, Philadelphia, 1996, pp. 265–273. [30] L.W. Hedlund, P. Vock, E.L. Effmann, M.M. Lischiko, C.E. Putman, Hydrostatic pulmonary edema an analysis of lung
[31]
[32]
[33]
[34]
[35]
[36] [37]
density changes by computed tomography, Invest. Radiol. 19 (1984) 254–262. P. Scillia, M. Delcoix, P. Lejeune et al., Hydrostatic pulmonary edema: evaluation with thin-section CT in dogs, Radiology 211 (1999) 161–168. R.C. Schalant, E.H. Sonnenblick, Pathophysiology of heart failure, in: R.C. Schlant, R.W. Alexander (Eds.), The Heart, McGraw-Hill, New York, 1994, pp. 515–556. A.C. Guyton, J.E. Hall, Pulmonary circulation; pulmonary edema; pleural fluid, in: A.C. Guyton, J.E. Hall (Eds.), Medical physiology, W.B. Saunders Company, Philadelphia, 1996, pp. 414–421. A.C. Guyton, J.E. Hall, Cardiac output, venous return, and their regulation, in: A.C. Guyton, J.E. Hall (Eds.), Medical Physiology, W.B. Saunders Company, Philadelphia, 1996, pp. 239–252. P.T. O’Gara, R.W. DeSanctis, Aortic dissection, in: J. Loscalzo, M.A. Creager, V.J. Dzau (Eds.), Vascular Medicine, Little, Brown and Company, Boston, 1996, pp. 927–950. B.K. Weir, Pulmonary edema following fatal aneurysm rupture, J. Neurosurg. 49 (1978) 502–507. E.L. Nagel, E.G. Fine, J.P. Krischer, J.H. Davis, Complications of CPR Crit. Care. Med. 9 (1981) 424.