| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Original Research |
1 Institute of Forensic Medicine, Buehlstrasse 20, Bern 3012, Switzerland.
2 Department of Diagnostic Radiology, University Hospital Bern, Bern,
Switzerland.
Received February 9, 2005;
accepted after revision April 18, 2005.
Address correspondence to E. Aghayev
(emin.aghayev{at}irm.unibe.ch).
Abstract
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
MATERIALS AND METHODS. This article describes the measurement and evaluation of the cross-sectional areas of major blood vessels, of the diameter of the right pulmonary artery, of the volumes of thoracic aorta and spleen on MDCT, and of the volumes of heart chambers on MRI in 65 autopsy-verified cases of fatal hemorrhage or no fatal hemorrhage.
RESULTS. Most cases with a cause of death of "fatal hemorrhage" had collapsed vessels. The finding of a collapsed superior vena cava, main pulmonary artery, or right pulmonary artery was 100% specific for fatal hemorrhage. The mean volumes of the thoracic aorta and of each of the heart chambers and the mean cross-sectional areas of all vessels except the inferior vena cava and abdominal aorta were significantly smaller in fatal hemorrhage than in no fatal hemorrhage.
CONCLUSION. For the quantitative differentiation of fatal hemorrhage from other causes of death, we propose a three-step algorithm with measurements of the diameter of the right pulmonary artery, the cross-sectional area of the main pulmonary artery, and the volume of the right atrium (specificity, 100%; sensitivity, 95%). However, this algorithm must be corroborated in a prospective study, which would eliminate the limitations of this study. Quantitative postmortem cross-sectional imaging might become a reliable objective method to assess the question of fatal hemorrhage in forensic medicine.
Keywords: aorta • blood vessels • fatal hemorrhage • heart chambers • MDCT • MRI • virtopsy
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
The difficulty of the exact diagnosis of the cause of death by fatal hemorrhage is in the fact that this diagnosis is based on three premises: the decreased extent of livor mortis; a subjective assessment of the color of the visceral organs regarding their paleness; and an estimation of the volume of blood loss by measuring extravasal blood, for example, in the thoracic or abdominal cavities. Unfortunately, all these premises are weak.
The low sensitivity of the extent of livor mortis was mentioned by Knight [1] in his textbook as follows: "The phenomenon (livor) appears at a variable time after death, indeed it may not appear at all, especially in infants, old people, those with anemia. It may be so faint as almost to escape detection."
The paleness of vascular organs is a relative criterion, and paleness may have other causes than acute bleeding, such as chronic anemia; furthermore, paleness may exist in nondepending areas without bleeding.
Estimation of the volume of blood loss becomes difficult when the blood is pooled in hematomas or ecchymoses and almost impossible when the blood has disappeared on the crime scene. Indeed, and in contrast to any clinical or radiologic department, there is no verification of the conclusion of forensic autopsy because it is the last step of investigation. This also means that no statistics are available in the literature about the accuracy of the forensic assessment of fatal hemorrhage.
The radiologic development of cross-sectional imaging has allowed volumetric data sampling, which has become a basis to implement the modern radiologic examination methods in forensic medicine. The increasingly efficient role of postmortem MDCT and MRI in forensic radiology has already been reported [2-4].
At the Institute of Forensic Medicine in Bern, in collaboration with the Department of Diagnostic Radiology of the University Hospital of Bern, 90 forensic cases were radiologically examined before October 2003 using postmortem MDCT and MRI. The study was performed as part of a scientific project (Virtopsy, www.virtopsy.com) that aims to develop minimally invasive autopsy. Thali et al. [4] qualitatively reported imaging findings of the first 40 cases, among which fatal hemorrhage was the most frequent radiologically indeterminable cause of death.
In this study, we attempted to quantitatively determine the postmortem radiologic signs of death due to fatal hemorrhage. This retrospective study describes the radiologic evaluation of the cross-sectional areas of selected major blood vessels; the diameter of the right pulmonary artery; and the volume of the thoracic aorta, the four heart chambers, and the spleen with respect to the question of fatal hemorrhage.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Case Samples
Ninety cadavers, among those delivered to the Institute of Forensic
Medicine for forensic autopsy, were studied before autopsy using postmortem
MDCT and MRI. Twenty-five cases with either tissue destruction or putrefaction
were excluded from this study. Thus, 65 cases scanned using identical body
sequences were eligible for this retrospective study
(Table 1). In 16 of the 65
cases, fatal hemorrhage was diagnosed as the cause of death at autopsy. In
another three cases, "massive bleeding" and "pale and
bloodless visceral organs" were described in the autopsy report. We
combined these 19 cases in the group of "fatal hemorrhage." The
remaining 46 of 65 cases had other causes of death at autopsy. They were
examined as a control group with "no fatal hemorrhage." Both
groups had comparable demographic data. The average weight and length in the
fatal hemorrhage group were 73 kg (range, 44-100 kg) and 173 cm (153-190 cm).
The average weight and length in the no-fatal-hemorrhage group were 71 kg
(range, 37-127 kg) and 171 cm (110-193 cm). The age range of the 19 cases in
fatal hemorrhage was 25-83 years (average, 50 years). The age range of the 46
cases in the no-fatal-hemorrhage group was 13-95 years (average, 49 years).
The fatal hemorrhage group consisted of 14 men and five women (ratio, 2.8:1),
and the no-fatal-hemorrhage group, of 34 men and 12 women (ratio, 2.8:1).
|
Imaging and Autopsy
For imaging studies, the bodies were wrapped in two radiologically
artifact-free body bags (Rudolf Egli AG). Combined MDCT and MRI examinations
were performed at an average of 28 hours after death, with 12 cases examined
within the first 12 hours, 20 cases within 1 day, nine cases within 2 days,
four cases within 3 days, one case within 4 days, and one within 5 days after
death.
MDCT was performed on a LightSpeed QX/i unit (GE Healthcare). Slice thickness was 1.25 mm. Measurements were calculated on between 780 and 1,800 axial cross-sections per body (average, 1,000 images). The duration of MDCT was 8-15 minutes. For MRI, we used a standardized protocol and additional sequences to address specific questions on a 1.5-T Signa EchoSpeed Horizon scanner (version 5.8, GE Healthcare). Axial planes of T1-weighted (TR/TE, 400/15) fast spin-echo or T2-weighted (4,000/100) fast spin-echo sequences of the body with a slice thickness of 4 or 5 mm and a gap of 1 mm were systematically used. MRI studies usually required 2-3 hours. Between 300 and 600 images per body (average, 400 images) were acquired. Image interpretation was performed by board-certified radiologists.
Autopsy, with examination of all three body cavities, was performed at an average of 52 hours (range, 24-142 hours) after death and at an average of 12 hours after imaging by board-certified forensic pathologists.
Measurements
All measurements were performed on a workstation (Advantage Windows 4.1, GE
Healthcare).
On axial MDCT images, the cross-sectional area of vessels was measured using the Spline-Function tool from Display Tools (Fig. 1) and the diameter, using the Direct-line Distance tool from Display Tools. The walls of the vessels were not included in the measurement. Measurements of the cross-sectional areas in square millimeters on 2D axial MDCT images of the following vessels were performed: first, the ascending aorta, just above the coronary orifices (Fig. 1); second, the descending aorta (DA1) on the same level as the ascending aorta (Fig. 1); third, the descending aorta (DA2) just below the orifice of the inferior vena cava; fourth, the abdominal aorta on the level of the kidney hilum (when a discrepancy existed in the position of the kidneys, the lower level was the one used for evaluation); fifth, the main pulmonary artery (MPA) on the same level as the ascending aorta (Fig. 1); sixth, the superior vena cava (SVC) on the same level as the ascending aorta (Fig. 1); seventh, the inferior vena cava (IVC1) just below its orifice; eighth, the inferior vena cava (IVC2) just below the liver; and ninth, the inferior vena cava (IVC3) just above the confluence of the iliac veins.
|
|
Measurements of the volume of the heart chambers were performed in MRI using Mass Analysis software version 5.1 (MR analytic system, Medis), including the right atrium (RA), right ventricle (RV), left atrium (LA), and left ventricle (LV). In the Mass Analysis software, manual outlining of contours of the heart chambers in the axial plane of each 2D MR image was performed. Furthermore, a 3D reconstruction of the chamber and a generated report of the measured volume were printed.
For identification of the quantitative differences between fatal hemorrhage
and no-fatal-hemorrhage, cutoff values separating fatal hemorrhage and
no-fatal-hemorrhage cases were defined for all locations, and the sensitivity
and specificity of these were calculated. Mean values of fatal hemorrhage were
compared with the mean values of no-fatal-hemorrhage using the Student's
t test (two-tailed, 
= 0.05).
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
The volumes of the aorta, the heart chambers, and the spleen were also considerably smaller in fatal hemorrhage. The average volume of the thoracic aorta in fatal hemorrhage was 18 cm3, and in no-fatal-hemorrhage it was 43 cm3. Thus, in no-fatal-hemorrhage the thoracic aorta contained on average more than two times the blood volume of the fatal hemorrhage cases. Volumes of the right atrium, left atrium, right ventricle, and left ventricle in fatal hemorrhage were accordingly only 18%, 23%, 24%, and 50%, respectively, of those in no-fatal-hemorrhage (Table 1).
Also, the average diameter of the right pulmonary artery was 69% smaller in fatal hemorrhage than in no-fatal-hemorrhage (Table 1). One hundred four (50%) of a total of 209 vessels were collapsed in fatal hemorrhage (Fig. 2), and only 101 (20%) of 506 vessels in no-fatal-hemorrhage (Table 2). In contrast to fatal hemorrhage, in all 46 cases of no-fatal-hemorrhage the SVC, IVC1, MPA, RPA and DA1 were never collapsed.
|
Fatal hemorrhage and no-fatal-hemorrhage cases were best separated when a cutoff value of 6 mm for the RPA, 130 mm2 for the MPA, or 13 cm3 for the right atrium was chosen. Specificity of these cutoff values was 100%, and their sensitivities were 73.5% (RPA), 63% (MPA), and 79% (right atrium).
The Student's t test showed significant differences between fatal hemorrhage and no-fatal-hemorrhage for all parameters except the IVC3, abdominal aorta, and volume of the spleen (Table 1).
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
The idea of measuring postmortem the cross-sectional area of major blood vessels on MDCT was born after the recent development of postmortem imaging as a subdiscipline of forensic medicine. The experienced eyes of radiologists observed an outstanding difference in the cross-sectional areas of major blood vessels between living patients and cadavers. Our recently collected material allowed the statistical quantification of this difference.
The clinical association of hypovolemia and hypotension with flattened (i.e., collapsed) vessels, especially the inferior vena cava, on abdominal CT scans has been known since the late 1980s and early 1990s [5-8]. Jeffrey and Federle [5] reported a strong correlation of hypovolemia with flattening of the inferior vena cava at multiple levels on CT scans in a patient who had suffered a substantial blunt abdominal trauma. In their retrospective review of 100 trauma patients, they revealed flattening in seven cases. Six of the seven patients had major hemorrhage. A control group of 100 abdominal CT scans showed no flattening in 98 cases. Taylor et al. [6] concluded that a flat vena cava may signal impending cardiovascular collapse. More recently, Sivit et al. [7] described the hypoperfusion complex in 27 of 1,018 children. Caval flattening was present in all 27 cases. Mirvis et al. [8] noted the presence of a flattened inferior vena cava on abdominal CT in 10 of 13 shock patients after blunt abdominal trauma. However, Eisenstat et al. [9] reported abdominal CT scans of 500 nontrauma patients; the flat vena cava was observed in 14% of all patients, and only a minority (30%) had hypotension or evidence of hypovolemia.
On postmortem CT examinations, Thali et al. [4] first described a reduction of the diameter of the abdominal aorta, the so-called aortic collapse sign. In the preliminary impression of those authors, a collapse of the descending aorta may be a reproducible radiologic sign of massive or even fatal hemorrhage.
In contrast to MRI, MDCT is a less expensive cross-sectional radiology method with faster scanning times—approximately 8-15 minutes in our experience. On the other hand, the limited soft-tissue-density differentiation of MDCT [3] does not allow an accurate measurement of the volume of the heart chambers. For this reason, in this study the volume of the four heart chambers was defined using MRI data. For the remaining measurements, MDCT appeared to be the proper method.
Our measurements of the cross-sectional areas of major blood vessels, presented in Table 1, show the clear difference between fatal hemorrhage and no-fatal-hemorrhage, with persons in fatal hemorrhage having noticeably smaller cross-sectional areas. Also, the volumes of the aorta, heart chambers, and spleen were smaller in fatal hemorrhage. Using the Student's t test, the difference between the mean values in fatal hemorrhage and no-fatal-hemorrhage was confirmed for nearly all parameters. The IVC3, abdominal aorta, and volume of the spleen do not seem to vary with fatal hemorrhage; the ranges of values for these three parameters were almost similar in fatal hemorrhage and no-fatal-hemorrhage (Table 1).
|
No-fatal-hemorrhage cases could be excluded when a cutoff value of 6 mm for the RPA, 130 mm2 for the MPA, or 13 cm3 for the right atrium was chosen (specificity, 100%). The sensitivity of these parameters was 73.5% (RPA), 63% (MPA), and 79% (right atrium).
The combined application of these cutoff values resulted in the algorithm illustrated in Figure 3. Whenever the RPA was less than 6 mm or the MPA was less 130 mm2 or the right atrium was less than 13 cm3, fatal hemorrhage was specifically detected with a specificity of 95%. When only the CT- or MRI-derived parameters were used, the specificity was 79%. The prevalence of the cutoff values in fatal hemorrhage is presented in Table 3.
|
Our study has the following limitations: First, the selection of the cases of fatal hemorrhage and no-fatal-hemorrhage was based on autopsy protocols and thus on a subjective assessment of blood volume in the corpse. Therefore, an error in case selection cannot be excluded. Second, postmortem changes of the vessels during the time between death and the radiologic scan were not taken into account. Recently, Stein et al. (presented at the 2004 annual meeting of German Society of Forensic Medicine) reported changes in the shape of the ascending aorta in all 137 cases examined using postmortem CT after the fifth postmortem day. Because we performed only one postmortem scan per body at an average of 28 hours, we were not able to correct our measurements for the time interval between death and the radiologic scan. A study with more cases and several consecutive scans per body would be needed to specify and quantify postmortem changes of the vessels. Third, we used absolute values and we did not correct for the body habitus; thus, we cannot exclude that weight, size, body surface area, or body mass index might influence the measured parameters among adults. Fourth, we used MDCT and MRI, which of course would be expensive in routine forensic work. MRI is more expensive, more time-consuming, and less available, although it might provide all the information needed. On the other hand, MDCT is much faster and less expensive but it also cannot differentiate the lumen of the right atrium under the condition of hypovolemia with possible collapse. It is therefore doubtful whether MDCT alone can increase the sensitivity of 79% to 95% after the second step of the algorithm, as observed by the addition of the third MRI-based step, measuring the volume of the right atrium.
At present only few institutes of forensic medicine in the world work on specific forensic CT or MRI scanners. Approximately 10 institutions have access to a scanner. The easiest way for a forensic institute to perform postmortem imaging is to collaborate with a radiology partner. Such a model exists in Bern within the Virtopsy project [2-4]. Previously published reports have shown that postmortem cross-sectional imaging has great potential in forensic medicine both as a noninvasive, objective documentation and as a diagnostic method [2-4]. Until scientific statistical proof of the value of imaging in specific forensic questions is obtained, imaging is not reimbursed.
The actual costs for examining one body using our reported method [2, 4], including transportation of the body; whole-body MDCT; and MRI of the head, thorax, and abdomen, add up to about twice those of a traditional autopsy.
In certain cultural circles in which conventional autopsy is stigmatized or even forbidden, such an approved method would allow sound medicolegal practice and support for the juridical system without violating religious prohibitions or personal reservations. At the same time, the rapid development of radiologic techniques enables decreased fees for CT and MRI examinations and improves their accuracy, favoring the development of postmortem imaging similar to the implementation of DNA technology in the field of forensic medicine.
In summary, we offer the following conclusions: First, most patients who died as a result of fatal hemorrhage showed collapsed vessels. Mean cross-sectional areas of all vessels except the inferior vena cava at the lowest level and the abdominal aorta were significantly smaller in fatal hemorrhage than in no-fatal-hemorrhage. Second, the finding of a collapsed superior vena cava, main pulmonary artery, or right pulmonary artery, together or alone, is a clear clue to fatal hemorrhage. Furthermore, a collapsed inferior vena cava or ascending or descending aorta also points to a fatal hemorrhage as a combined or a single cause of death. Third, to differentiate cases of fatal hemorrhage from other causes of death, we hypothetically propose a three-step algorithm with measurements of two vessels and one heart chamber. The retrospective specificity and sensitivity of this algorithm after the third step were 100% and 95%, respectively. However, both MDCT and MRI are needed, and this hypothesis must be corroborated or corrected in prospective studies, eliminating the limitations of this study. Fourth, provided that our hypothesis can be confirmed, quantitative postmortem cross-sectional imaging might become a reliable objective method to assess the question of fatal hemorrhage in forensic medicine.
Acknowledgments
We thank Elke Spielvogel, Carolina Dobrowolska, Christoph Laeser, Verena
Beutler, and Karin Zwygart (Department of Radiology, University Hospital of
Bern) and Urs Koenigsdorfer and Roland Dorn (Institute of Forensic Medicine,
University of Bern) for their excellent help and data acquisition during the
radiologic examinations and the forensic autopsy.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
This article has been cited by other articles:
|
|
 |
|
  R. Dirnhofer, C. Jackowski, P. Vock, K. Potter, and M. J. Thali VIRTOPSY: Minimally Invasive, Imaging-guided Virtual Autopsy RadioGraphics, September 1, 2006; 26(5): 1305 - 1333. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
