Author + information
- Received December 22, 2012
- Accepted January 15, 2013
- Published online August 13, 2013.
- Christian Jackowski, MD, EMBA∗∗ (, )
- Nicole Schwendener∗,
- Silke Grabherr, MD† and
- Anders Persson, MD, PhD‡
- ∗Forensic Imaging Center Bern, Institute of Forensic Medicine, University of Bern, Bern, Switzerland
- †Centre Universitaire de Médecine Légale (CHUV)–University of Lausanne, Lausanne, Switzerland
- ‡Center for Medical Image Science and Visualization (CMIV), University of Linköping, Linköping, Sweden
- ↵∗Reprint requests and correspondence:
Prof. Dr. Christian Jackowski, University of Bern, Institute of Forensic Medicine, Buehlstrasse 20, CH-3012 Bern, Switzerland.
Objectives This study aimed to investigate post-mortem magnetic resonance imaging (pmMRI) for the assessment of myocardial infarction and hypointensities on post-mortem T2-weighted images as a possible method for visualizing the myocardial origin of arrhythmic sudden cardiac death.
Background Sudden cardiac death has challenged clinical and forensic pathologists for decades because verification on post-mortem autopsy is not possible. pmMRI as an autopsy-supporting examination technique has been shown to visualize different stages of myocardial infarction.
Methods In 136 human forensic corpses, a post-mortem cardiac MR examination was carried out prior to forensic autopsy. Short-axis and horizontal long-axis images were acquired in situ on a 3-T system.
Results In 76 cases, myocardial findings could be documented and correlated to the autopsy findings. Within these 76 study cases, a total of 124 myocardial lesions were detected on pmMRI (chronic: 25; subacute: 16; acute: 30; and peracute: 53). Chronic, subacute, and acute infarction cases correlated excellently to the myocardial findings on autopsy. Peracute infarctions (age range: minutes to approximately 1 h) were not visible on macroscopic autopsy or histological examination. Peracute infarction areas detected on pmMRI could be verified in targeted histological investigations in 62.3% of cases and could be related to a matching coronary finding in 84.9%. A total of 15.1% of peracute lesions on pmMRI lacked a matching coronary finding but presented with severe myocardial hypertrophy or cocaine intoxication facilitating a cardiac death without verifiable coronary stenosis.
Conclusions 3-T pmMRI visualizes chronic, subacute, and acute myocardial infarction in situ. In peracute infarction as a possible cause of sudden cardiac death, it demonstrates affected myocardial areas not visible on autopsy. pmMRI should be considered as a feasible post-mortem investigation technique for the deceased patient if no consent for a clinical autopsy is obtained.
Post-mortem magnetic resonance imaging (pmMRI) has become a valuable tool to noninvasively document forensic and pathological findings (1–5). pmMRI may be especially useful for visualizing pathological findings in the soft tissue, using thin cross sections that are not possible with macroscopic routine autopsy (4). However, MRI in deceased persons is far different from clinical MRI, and interpretation of unenhanced pmMRI requires special expertise (6).
Cross-sectional pmMRI is being evaluated and validated for the assessment of different causes of death. These research efforts are mainly those of forensic institutes and are thought to serve as a supplement to conventional autopsy in forensic cases to substantiate post-mortem diagnostics. pmMRI may also have the potential to provide an alternative post-mortem examination technique for clinical pathology to counterbalance the reduction in post-mortem diagnostics due to decreasing clinical autopsy rates (7,8). This trend broadly affects the quality of the healthcare systems and thereby the health of the general population.
As cardiac-related deaths represent the major portion of natural deaths in our community it is of particular importance that pmMRI can demonstrate pathological findings in the human heart as well (9). Tissue alterations occurring during and after myocardial ischemia are most important. Recent studies on myocardial infarction visualized on pmMRI were performed using 1.5-T systems, with rather small study populations and findings of limited significance (4,6,10). However, acute, subacute, and chronic infarction could be differentiated using the signal behavior on T1- and T2-weighted (T2w) images. Jackowski et al. (10) recently also showed that hypointensities on unenhanced T2w images without any hyperintense margin may be a sign of peracute ischemic lesions (age between minutes and approximately 1 h), findings hardly detectable at autopsy and using routine histological examination. Sudden cardiac death, without the definite myocardial finding of a fresh ischemic lesion, has challenged pathologists and forensic examiners for decades. As myocardial dissection obviously fails in reliably visualizing these very early myocardial alterations, coronary status obtained at autopsy was additionally chosen for the present study as the “gold standard” to validate the finding of myocardial hypointensities on T2w images. The aim of the present study was to validate the known 1.5-T pmMRI appearances of different infarction stages (acute, subacute, and chronic) on a 3-T system in a study population larger than those in the existing literature. Second, the finding of myocardial hypointensity on T2w imaging as a possible myocardial appearance of so-called “sudden cardiac death” was correlated to histology and, more important, to the coronary status obtained on autopsy.
Between September 2010 and December 2011, 136 forensic corpses either presenting a case history of chronic or acute cardiac anamnesis or with death under circumstances making a cardiac cause of death very likely were prospectively enrolled. In all cases, a forensic autopsy was ordered by the local authorities, and the corpses were delivered to a forensic institute equipped with an MR scanner dedicated for post-mortem examinations only. Age at death ranged from 0 to 94 years (mean: 56.2 ± 19.9 years; 31 female, 105 male). The post-mortem interval (time from death to scanning) ranged from several hours to 3 days. Relevant putrefaction gas formation was ruled out using whole-body computed tomography prior to pmMRI. In all cases, a 3-T pmMRI examination of the heart was performed in situ prior to forensic autopsy.
A total of 76 of the examined cases (mean age: 57.8 ± 16.7 years; 14 female, 62 male) presented with cardiac findings on pmMRI and were further investigated within the present study. The remaining 60 cases showed no myocardial findings on pmMRI or autopsy. Causes of death were pulmonary artery embolism, rupture of aortic aneurism, aortic dissection, intoxication, subarachnoid hemorrhage, trauma, gastrointestinal hemorrhage, pneumonia, intracerebral hemorrhage, hepatic failure, drowning, asphyxia, multiorgan failure, sepsis, and/or hemangioblastoma. Use of the image data for the present study was approved by the local ethics committee.
The corpses were wrapped either in an artifact-free body bag, a plastic foil, or a linen sheet. All subjects were scanned while in the supine position using a 3-T MRI scanner (Achiva, Philips Healthcare, Best, the Netherlands) with a 16-channel torso-XL coil. Short-axis images were acquired using conventional clinically used localizer settings (11). Scanning parameters were as follows: T1w turbo spin echo (TSE) (repetition time [TR]: 540 ms; echo time [TE]: 10 ms), T2w TSE (TR: 2,430 ms; TE: 100 ms), T2w TSE fat saturation (TR: 1,610 ms; TE: 60 ms), near proton density–weighted TSE (TR: 1,800 ms; TE: 25 ms), near proton density–weighted TSE fat saturation (TR: 2,035 ms; TE: 30 ms), and gradient echo fast field echo (TR: 450 ms; TE: 5.8 ms). Slice thickness was 3 mm and gap was 0.3 mm. Examination time was 1 h. Image interpretation was performed according to Jackowski et al. (4,10), predominantly based on the signal behavior on T2w images (Table 1). Peracute ischemic lesions, not visible on autopsy, with or without early histological alterations were diagnosed on pmMRI when cloudy, hypointense myocardial areas on T2w without any hyperintense marginal edematous reaction were present (Fig. 1). Acute ischemic lesions (autopsy aspect: yellowish discoloration; histology: fiber necrosis, granulocytous infiltration, myocardial edema) presented with a hypointense central zone surrounded by a hyperintense margin on T2w. The relation of hypointensity to hyperintense margin was variable. Earlier acute stages showed lower contrast with rather lesser developed perifocal hyperintense margins (Fig. 2) compared with the fully developed large acute infarctions (Fig. 3). The latter presented with a deep dark hypointense area surrounded by a distinctively elevated signal, resulting in a rather high-contrast alteration. The hyperintense margin was mostly more prominent in subepicardial zones. In cases of intramyocardial post-infarction hemorrhage, the typical appearance was disturbed and the alteration showed a heavily inhomogeneous appearance (Fig. 4). Subacute lesions (autopsy aspect: grayish myocardial alteration with slightly denser consistency; histology: fibroblasts with loose connective tissue formation, angiogenesis) showed mainly hyperintense myocardial zones on T2w that sometimes presented some minor zones of chronification (Fig. 5). Chronic lesions of definite myocardial scars (autopsy aspect: white, shrunken left ventricular [LV] wall; histology: nuclei-free collagenous fibers) showed loss of signal on all applied sequences and, if transmural, thinning of the affected LV wall (Figs. 6 and 7).
Imaging findings were reported to the team of forensic pathologists who were responsible for the forensic autopsy, enabling histological examination of lesions detected on pmMRI but not visible on macroscopic examination during dissection.
Autopsy was carried out either directly after scanning or the following morning by board-certified forensic pathologists. The cardiac dissection was adapted to match short-axis images on pmMRI by slicing the myocardium in base-parallel slices. The coronary orifices and the apex were used as anatomic landmarks that allowed comparison of similar slice sections on imaging and autopsy. Documentation of the coronary status was performed either by slicing the coronaries or by longitudinal dissection. Distended photographic documentation was carried out. Histological examination was performed on the regular forensic basis in the anterior, septal, posterior, and lateral LV myocardium; His bundle; anterior and posterior papillary muscle; and right ventricular myocardium. Additionally, all lesions detected on pmMRI or macroscopically visible on dissection were included in histological examinations (targeted histology). Histological stains included hematoxylin and eosin, elastic van Gieson, and chromotrop-aniline-blue.
Imaging findings were morphologically correlated to autopsy findings in terms of location and stage assessment. In cases of pmMRI lesions assessed as peracute infarction and showing no visible alteration on autopsy, correlation to autopsy-obtained coronary status was additionally performed.
Myocardial staging was performed on the basis of pmMRI findings, autopsy aspect, and histological findings according to Jackowski et al. (4).
In 76 of 136 cases (55.9 %), at least one ischemic lesion was detected on pmMRI, and cardiac-related cause of death was present as the final forensic case assessment (Table 1). Within the 76 cases, a total of 124 myocardial lesions were detected on pmMRI. These lesions consisted of 25 chronic, 16 subacute, 30 acute, and 53 peracute myocardial alterations as assessed on pmMRI. Chronic, subacute, and acute lesions were confirmed macroscopically and histologically. There was 100% agreement between pmMRI and autopsy. Lesions described as peracute on pmMRI could not be observed macroscopically on autopsy. These 53 peracute lesions were distributed among 42 cases (e.g., one LAD event could result in hypointense lesions within the anterior wall, within the anterior septum, and/or within the anterior papillary muscle).
In 33 of the peracute lesions (62.3%), very early histological alterations, such as loss of striation, contraction band necrosis, wavy fibers, or hypereosinophilia, could be observed in targeted histological investigations. In 20 of the peracute lesions (37.7%), no such early histological alterations could be detected.
In 45 of the peracute lesions (84.9%), a matching coronary finding could be observed on autopsy. In these cases, either an acute coronary event (fresh thrombotic occlusion, soft plaque rupture) or a severe chronic stenosis within the main supplying coronary artery was present. Eight lesions assessed as peracute on pmMRI presented without a matching coronary event on autopsy. Within these 8 lesions, targeted histological examination was positive in 3 (Cases #4 and #7), severe concentric hypertrophy was present in 2 (Case #3; heart weight: 690 g), death occurred after cocaine intoxication in 1 (Case #12), and death occurred after sexual activity in 2 (Case #62). All cases presenting with peracute lesions on pmMRI showed indirect signs of cardiac failure on autopsy, such as internal congestion and pulmonary edema, and were forensically assessed as sudden cardiac death.
First histological signs of myocardial infarction appear after approximately 2 to 4 h of survival time. Early acidosis due to the consumption of adenosine triphosphate causes the myocytes to become increasingly susceptible to eosinophilic stains as a first visible sign on routine histological examination (12). Contraction band necrosis may also be observed (13). Loss of striation in single fibers or smaller groups of fibers, swollen (cellular edema) fibers, or thinned and wavy fibers may be observed as well (14,15). On histological examination, there is usually a thin (0.3 to 0.5 mm) subendocardial layer of surviving tissue due to direct oxygen diffusion through the endocardium (exception: intraventricular thrombus and thickened endocardium) (14,16,17). As a frequent finding, intramyocardial hemorrhage occurs focally within the infarcted tissue. The inflammatory cellular reaction is commenced by the infiltration of polymorphonuclear leukocytes. The amount of infiltration progressively increases until the 4th day. After the 5th day, the polymorphonuclear leukocytes disappear. Mononuclear cells remove necrotic fibers by phagocytosis. Newly formed blood capillaries (angiogenesis) growing into the infarcted regions can be observed. Along these vessels, fibroblasts reach the necrotic areas. The myofibroblasts express alpha-smooth muscle actin and start the formation of collagen (types I and III) (16,18–20). The formation of collagen is moderately prominent at 3 weeks and reaches its maximum at about 3 months. The fibrous tissue reaction can also spread into noninfarcted myocardial regions, but to a distinctively lesser extent (21–23). Revascularization seems to influence these ongoing alterations only during the first 4 to 6 h after a coronary occlusion, resulting in a reduction of the infarct size but not in a complete reversibility (24,25). Later reperfusion does not stop the necrosis but may accelerate subsequent inflammatory and remodeling reactions.
These chronologically occurring and histologically detectable alterations cause specific MR signal appearances that also allow for an infarct stage estimation based on the combination of the signal alterations on different weightings. The appearances of the acute, subacute, and chronic stages on pmMRI have been presented in earlier publications (4,10). The 71 myocardial lesions (acute, subacute, and chronic) in the present study support the results of recent studies in very limited populations and using 1.5-T (4,10,26). However, it is the description of the peracute myocardial ischemic lesion in a large number that needs to be discussed thoroughly as this finding has an impact on future post-mortem cardiac diagnostics.
In a recent work, the finding of a hypointensity on T2w was first described in six cases using 1.5-T (10). In five of the six cases, there was an explanatory coronary finding reported, and the sixth case presented with a severe hypertrophic alteration. It was concluded that hypointensities on T2w may represent the area of myocardial affection in peracute infarction.
A single in the present study (Case #25) has been previously been reported as having a comparable myocardial lesion in combination with the nonenhanced visualization of the coronary thrombus considered to explain the ischemic hypointensity (26). However, the present study could, for the first time, collect data on a large number of cases presenting with hypointense T2w lesions correlating well to coronary events. In 42 cases, 53 hypointense T2w lesions were present that were not visible on macroscopic dissection. Knowledge about the pmMRI finding allowed for a targeted histological examination that showed early ischemic alterations in 62.3% of lesions. On the basis of the present study, it is to be expected that these ischemic alterations would have not been detected on routine histological examination without advanced knowledge of the MR finding. In 37.7%, no histological alteration could be found within the affected (MR diagnosis) myocardium. However, a comparison of hypointensity on MR to coronary status on autopsy yielded more satisfactory results. In 84.9%, a coronary finding could be observed that was able to explain an ischemic situation within the affected myocardium. Relating these results to routine autopsy, it is to be expected that this group of study cases well explains the majority of those unsatisfactory autopsy cases ultimately to be assessed as sudden cardiac death but without verifiable myocardial alterations. In these cases, pmMRI may provide a valuable benefit to post-mortem investigations because it demonstrates a possible myocardial origin of a fatal ventricular arrhythmia, whereas the ventricular arrhythmia itself cannot be verified post-mortem. On the other hand, it helps to distinctively improve the hit rate of histological investigations.
It is not fully understood why 3-T MR seems to be more sensitive for peracute myocardial lesions than conventional examination methods. Hypotheses include the early drop in pH value that shortens T2 relaxation. A possible contribution of decreased interstitial water has also been discussed due to insufficient arterial blood supply while the venous drainage is not affected. However, the T2 effect seems to be dominating.
Recently, direct thrombus imaging techniques have been clinically implemented on the basis of non–contrast-enhanced T1w MR and shown to be effective in detecting coronary thrombus as well (27). Post-mortem, direct thrombus imaging using nonenhanced MR has been shown to visualize coronary thrombosis as well. In contrast to clinical conditions, T2w images have proved to be more effective than T1w MR when it comes to post-mortem application of direct thrombus imaging techniques (26). The myocardium as well as the coronary status may be investigated in a single, non–contrast-enhanced pmMRI examination.
First of all, there was no adequate control group available. There were 60 cases examined without myocardial findings and presenting a noncardiac cause of death. However, this group was not taken as control group because these cases have not been chosen in advance as such. The results of the investigations including pmMRI have led to the assessment as cases without myocardial findings. If a case presented a finding, it was treated as a study case and further investigated as such. However, there would be no false-positives or false-negatives to be mentioned within the 60 excluded cases.
The investigators were not blinded. We had to decide whether it was more important to blind the investigator or to obtain targeted histological samples from the peracute lesions detected on pmMRI. Only the informed forensic pathologist was able to obtain such samples during autopsy. With respect to the main study question, we assessed the value of histological confirmation as greater than that of blinding of the forensic pathologist. However, that remains arguable.
Furthermore, the gold standard chosen, namely, autopsy (myocardium and coronary arteries), and histological examination have limitations. Especially with peracute infarction, cases lacking a coronary event could have also been interpreted as false-positive. However, 3 of the 8 lesions without a coronary finding were verified as peracute by early histological signs. Two more lesions occurred in a heart weighting about 690 g. As a general autopsy rule for pathologists, it is accepted that heart weights over 500 g can explain a sudden cardiac death due to ischemic events without a coronary finding. In these cases, even a healthy coronary system may not remain able to sufficiently supply the hypertrophic myocardium. One lesion occurred in a case related to cocaine intoxication. Coronary spasms from cocaine intoxications are well known and also clinically described (28–32). In these cases, ischemic myocardial alterations may be present without verifiable coronary findings. Thereby 6 of the 8 lesions without a matching coronary event provide further reasonable explanations for an ischemic lesion.
The only arguable case seems to be a young female patient who died suddenly right after sexual activity and showed indirect signs of cardiac failure but no cardiac or coronary findings on autopsy. The risk for myocardial infarction with the coronary arteries being normal is remarkably higher in younger individuals (33) and in women (34), and an imbalance between oxygen demand and supply as well as intense sympathetic stimulation are considered to be risk factors for myocardial infarction without coronary findings (28). Therefore, this case presented with a high-risk profile for myocardial infarction without coronary findings. However, a post-mortem validation of the pmMRI finding via the chosen gold standard was not possible.
The failure of the gold standard (autopsy) to detect peracute infarction was likely the result of the short survival times of these cases. A survival time of a few minutes does not allow for sufficient vital reactions within the myocardium, such as edema or necrosis, to develop. In peracute infarctions, initial ischemia and death of the person follow each other in such close succession that no vital reactions can be expected within the affected myocardium. In very early cases, macroscopic autopsy and even histological examination may fail to differentiate between initial ischemic myocardium and global ischemic myocardium at time of death. In these cases, autopsy still depends on coronary diagnostics showing severe stenosis or occlusions. According to the literature on pathology, only one-half of peracute cases present with an acute coronary lesion (35). In our forensically selected case material, the percentage of coronary events was distinctively higher. However, there were also peracute cardiac deaths without coronary findings. These cases presented with extracardiac signs of cardiac failure but no myocardial findings.
Histology was limited to routinely used stains (hematoxylin and eosin, elastic van Gieson, and chromotrop-aniline-blue). There are also immunohistochemical investigation techniques that may gain further insight (36). However, in very early infarction-age stages, these techniques also have limitations; for this reason, and because coronary artery findings are expected to be stable over time and to be positive at the earliest stages of myocardial ischemia, coronary artery findings were considered the best choice for validation.
Image contrast was different between the study cases, depending on the actual core temperatures of the bodies. Forensic corpses have core temperatures mainly ranging from 4°C to about 30°C (with several exceptions, e.g., burned or frozen corpses). Especially epicardial and body fat present with low signal on T2w and T1w imaging. Therefore, there is a need to define echo, inversion, and repetition times for individual core temperature ranges intending to increase the contrast-to-noise ratio on pmMRI. This is going to be addressed in an upcoming study by our group.
Furthermore, post-mortem alterations related to the time since death, such as rigor mortis, which also occurs on myocardium, affected the comparability between different study cases. Post-mortem contracted myocardium seems to have a lower posterior descending branch than prior to the development of rigor or after rigor has relaxed. On post-mortem imaging, blood is not a homogeneous fluid but has developed a two-layer appearance due to the sedimentation of cellular components. Therefore, blood presents (e.g., on T2w pmMRI) as a signal-intense upper serum layer and a lower dark erythrocytes layer (37). Beginning putrefaction gas accumulations might cause susceptibility artefacts even when putrefaction has not yet developed to a macroscopically detectable stage.
Although one of the study cases nicely showed a coronary finding (fresh left anterior descending coronary artery thrombus) on short-axis pmMRI, we were not able to investigate the entire coronary system on pmMRI. In most of the cases, the proximal parts of the left and right coronary arteries were not covered by the most basal short-axis images that have been optimized to cover the LV only.
Because pmMRI cannot be used to assess late enhancements, is limited to pure morphological imaging. However, this disadvantage seems to be counterbalanced by the excellent image quality due to the absence of any cardiac motion– or breathing-related artefacts as well as fewer scan-time restrictions.
The results of the presented study demonstrate that unenhanced pmMRI is able to visualize and discriminate the different stages of myocardial infarction, including peracute infarction stages not visible on autopsy. Thereby, it can support forensic autopsy by allowing for targeted histological examination and may serve as a post-mortem examination technique alternative to conventional autopsy in clinical pathology. Clinicians and especially cardiologists should become aware of this fast and bloodless possible option in obtaining post-mortem diagnoses in diseased cardiac patients who will not undergo clinical autopsy. Consent for a 1-h pmMRI examination may be more easily obtained as it is expected to be more acceptable to the next of kin compared with clinical autopsy. Thereby, the negative effects of declining clinical autopsy numbers on clinical quality control and the health of the population in general may be counterbalanced.
The authors thank Dr. Morten Keller-Sutter and his team of forensic examiners and autopsy technicians for the experienced support during post-mortem autopsy documentation. The authors also thank Katrin Renfer for her assistance with manuscript preparation.
This work was supported in part by a grant from Philips AG Healthcare, Zürich, Switzerland (to Prof. Dr. Jackowski). The authors have reported that they have no other relationships relevant to the contents of this paper to disclose.
- Abbreviations and Acronyms
- left ventricular
- magnetic resonance imaging
- post-mortem magnetic resonance imaging
- Received December 22, 2012.
- Accepted January 15, 2013.
- American College of Cardiology Foundation
- Jackowski C.,
- Christe A.,
- Sonnenschein M.,
- Aghayev E.,
- Thali M.J.
- Polak M.
- Sun Y.,
- Weber K.T.
- Litwin S.E.,
- Litwin C.M.,
- Raya T.E.,
- Warner A.L.,
- Goldman S.
- Jansen C.H.,
- Perera D.,
- Makowski M.R.,
- et al.
- Tun A.,
- Khan I.A.
- Reynolds H.R.,
- Srichai M.B.,
- Iqbal S.N.,
- et al.
- Farb A.,
- Tang A.L.,
- Burke A.P.,
- Sessums L.,
- Liang Y.,
- Virmani R.