Pathological analysis of cadavers for educational dissection by using postmortem imaging

This study was performed primarily to clarify whether pathological analysis of cadavers for anatomical dissection is possible using postmortem imaging (PMI), and whether this is worthwhile. A total of 33 cadavers that underwent systematic anatomical dissection at our medical school also underwent PMI. Fixative solution was injected into the corpus 3–4 days after death. PMI was then performed using an 8‐slice multi‐detector CT scanner 3 months before dissection. Before dissection, a conference was held to discuss the findings of the PMI. First, two radiologists read the postmortem images without any medical information and deduced the immediate cause of death. Then, the anatomy instructor revealed the medical information available. Based on this information, the radiologist, anatomy instructor, and pathologists suggested candidate sampling sites for pathological examination. On the last day of the dissection period, the pathologists resected the sample tissues and processed them for pathological examination. In 12 of 33 cases, the presumed causes of death could be determined based on PMI alone, and revision of the cause of death described in the death certificate was considered in five (15.2%) cases, based on PMI and pathological analysis. This article presents a novel method of pathological analysis of cadavers for anatomical dissection using PMI without disturbing the anatomy education of medical students.


INTRODUCTION
Anatomy, including dissection of human cadavers, is an indispensable basic subject in the medical education program. 1,2 In the recent medical education reform, there is a tendency for weight to be placed on clinical medicine and training and various alternative methods are being tried. Integration of newer teaching modalities and modern technology encourage interest and retention of anatomical knowledge and its clinical relevance. 3 In anatomy in particular, it is intended that a medical student learns the normal physical structure, however, many cadavers used in anatomy harbor a variety of diseases making this difficult.
The cadaver can be analyzed using diagnostic imaging using computed tomography (CT) and/or magnetic resonance imaging (MRI). Postmortem imaging (PMI) seems to be the common term, however, the procedure is variously described as virtopsy in Switzerland, 4 virtual autopsy in France, 5 radio-autopsy in Germany 6 and autopsy imaging (Ai) in Japan. 7 The imaging results of the cadaver are naturally included in the anatomy education for medical students. For example, in the Mount Sinai School of Medicine, the PMI of the cadaver is regarded as a powerful teaching tool and is standard in most anatomy courses, 8 across various countries, including Ireland and Poland. [9][10][11][12][13][14] In Japan, many universities also perform PMI before anatomical dissection starts and combine imaging education with anatomy education. 15 However, it is rare that any further medical examinations of the cadaver, other than imaging, are carried out.
While dissection of the cadaver is indispensable for the anatomical education of the normal physiological structure, it cannot be overemphasized that pathological autopsy is indispensable for the understanding of disease. The rate of academic autopsies has, however, been decreasing worldwide in recent years. Despite many benefits, 16 autopsies are performed after less than 10% of all US deaths, 17 and there has been a decrease in the hospital autopsy rate from 21.6% in 1990 to 7.9% in 1999. 18 The mean hospital autopsy rate in 2013 in the UK was 0.69% of hospital deaths, 19 and the autopsy rate in Germany is now below 10%. 20 In France, the rate of autopsy in hospitals is also decreasing, 21 while in Brazil, the median autopsy rate has fallen from 19.3% in 2003 to 10.6% in 2008 (P = 0.07). 22 According to the Japan Council for Quality Health Care, the autopsy rate of 397 hospitals, which were tested by the hospital usability test in 2012, was 4.0%. 23 Pathological examination of the cadaver for anatomical dissection is done incidentally, such as making specimens of the tumors found during dissection, since the primary purpose of the anatomical dissection is anatomy education. Thus, sampling a pathological specimen systematically has been impossible, because that would interfere with the anatomy education. Therefore, we attempted pathological analysis of cadavers with PMI for systematic anatomical dissection. One purpose of this study was to clarify whether pathological analysis of cadavers for anatomical dissection using PMI is possible. Furthermore, this paper shows the rate at which a more accurate cause of death could be determined.

Objective
From October, 2010, to September, 2015, 100 cadavers underwent systematic anatomical dissection at Fukui University, School of Medicine. During their lifetime, permission for PMI was obtained from 33 of the donors.
Systematic anatomical dissection is conducted according to the Japanese law 'A Body Donation Law'. The consent for PMI of the donated corpus was obtained from the bereaved individual. This study obtained approval for the research program from the Ethical Review Board of the University of Fukui, School of Medicine.

Treatment of the corpus
Approximately 10 L of fixative solution (mixture of formalin, alcohol and glycerin) was injected into the corpus over 10-30 min through a femoral artery 3-4 days after death. Only the fixative solution was injected, and around 1 L was discharged (egested). The fixative seems to collect in the abdominal cavity and the thoracic cavity, thus, whether pathologically significant ascites and pleural effusion were present could not be determined. The brain was resected from the corpus 1 week after death because the gross neuroanatomy training of the resected brain is performed by a different part of the curriculum from the gross anatomy training of the human body. It is not possible to evaluate the brain histopathologically, since the brain is used for the gross neuroanatomy training of medical students.

Taking PMI
The PMI was performed at the Ai center at the University of Fukui, 3 months before dissection, using an 8-slice multi-detector CT scanner (Hitachi Medico, Tokyo, Japan) used exclusively for autopsies, as described previously. [24][25][26] The corpus was placed in the supine position, and a full-body scan from the vertex to the toes was performed. The scanning conditions were 120 kV, 250 mA, 8 × 2.5 collimation, 1.125 pitch, 0.8 s rotation time, 5 mm slice thickness and 5 mm increments. No contrast agent was used in these cases. The corpus was appropriately treated with dignity at all times. In the pyloric region of stomach, stent found as per imaging. Over all walls of stomach layers, signet ring cell adenocarcinoma is detected.
In the right ilium, osteoblastic tumor.
In oral cavity, implant and torus palatinus.
Cancer cells present from the muscular layer to the serosa. Cancerous peritonitis.

III. Gastric cancer
Bone marrow occupied by small tumor cells.
Epithelial binding is found and regarded as metastasis of the cancer, but seems to be different from the signet ring cells found in stomach.
10 79/M Postoperative state of chest wall.

Not determined
Stent in the coronary artery.
Bypass operation of the right coronary artery, anastomosis to a circumflex artery.
(No neutrophil infiltration in myocardium.) The valve is also postreplacement.
Gall stones. Artificial hip prosthesis. Osteolytic lesion in the sacrum.
Artificial prostheses of bilateral knee joints.
Right adrenal gland swelling Accessory spleen.

(S/O aspiration pneumonia)
Pulmonary edema and pneumonia.
Atypical cells in the spleen, C/W leukemic cells.
Hemosiderosis of the spleen.
Sacral lesion not detectable.

PMI conference
Before anatomical dissection began, a PMI conference including two radiologists, an anatomy instructor, two pathologists and a radiological technologist was held. First, two radiologists (KK and TS) read the postmortem images of the corpus without any medical information and deduced the immediate cause of death. The algorithm for diagnosis is described in a previous article. [24][25][26] Then, the anatomy instructor (SI) revealed the medical information. The medical information concerning the corpus included only the age, sex, cause of death and the place of death described in the death certificate. Based on this information, the radiologist, anatomy instructor and pathologists suggested candidate sampling sites.

Conventional histopathology
Based on the PMI conference, the pathologists (SN and KI) resected the sample tissues from the lesions of the corpus on the last day of the dissection period. The sampled tissue from the lesion was sliced into 5 mm thick sections. The average number of samples was around four, one at the Marked emphysema and pneumonia.

No abdominal findings.
Diverticulum in the right side of the urinary bladder, 5 cm in diameter. minimum and seven at the maximum, per case ( Table 1). The tissues were embedded in paraffin and sectioned with a thickness of 6 μm. For routine histological examination, each section was stained with hematoxylin and eosin (H-E). In a case with suspected amyloidosis or tuberculosis, Congo-red staining or Ziehl-Neelsen staining was performed, respectively.

Immunohistochemical staining
Immunostaining and special staining were performed as needed. The method of immunostaining in our laboratory has been described previously. 27,28 That is, 4 μm thick sections of paraffin-embedded tissue were deparaffinized with xylene, which was then replaced with ethanol. After washing with water, the intrinsic peroxidase activity was blocked with a 0.03% H 2 O 2 solution dissolved in absolute methanol at room temperature for 15 min, and the section was then rinsed with phosphate buffered saline (PBS). The sections in PBS were heated in a microwave oven at 750 W for 15 min for heatmediated antigen retrieval. The sections were then reacted at 4°C overnight with the antibody, which was diluted according to the directions for use. All antibodies were commercially purchased: CK5/6, TTF-1, p63, AE1/AE3 (Nichirei Biosciences, Tokyo, Japan), CK7, CK20, CD15, CD30, CEA, S-100, PSA, AFP, EMA, p40 (Agilent, Glostrup, Denmark), Chromogranin A, Synaptophysin CD56, GCDFP-15 (Leica Biosystems, Wetzlar, Germany). Detection of antibody reactivity was performed using a commercially available kit (Dako EnVision Plus-horseradish peroxidase (HRP); Agilent, Glostrup, Denmark), as follows. The sections were reacted with a goat-anti-mouse EnVision-HRP-enzyme conjugate for 30 min at room temperature, followed by rinsing with PBS. The peroxidase color reaction was visualized by incubating the sections with the chromogen, 0.02% 3-3-diaminobenzidine tetrahydrochloride (Sigma Chemical Co., St Louis, MO, USA), at room temperature for 10 min. The sections were then washed with water. After the immunostaining procedures were completed, the sections were lightly counterstained with hematoxylin. As a negative control, the exact same procedure was performed without the primary antibody. The pathological diagnosis was done based on these procedures, and the specimens were analyzed in the pathology laboratory.

RESULTS
The summary of all cases that underwent pathological analysis is shown in

Case 13. A 76-year-old man
Osteolytic lesions in the bilateral ilium (Fig. 1a), right ischium and left rib were found on PMI. These findings suggested metastasis of a malignant tumor. In addition, a soft tissue tumor was also noted in the right femoral region (Fig. 1b). Whether this soft tissue tumor was primary or secondary was an issue for discussion at the PMI conference. The cause of death given on the death certificate was multiple myeloma. Iliac bone (Fig. 1c) and a soft tissue tumor (Fig. 1d) were obtained, and pathological specimens were prepared. These pathological specimens showed atypical plasma cells in the bone marrow of the ilium (Fig. 2a,b), and the soft tissue tumor consisted of eosinophilic amorphous material (Fig. 2c,d). The material showed apple-green birefringence with Congo-red staining. The material was thus amyloid deposits, and the soft tissue tumor was diagnosed as amyloidoma.

Case 14. An 86-year-old woman
On PMI, a cardiac effusion was found. The effusion was suspected to be bloody (Fig. 3a), so cardiac tamponade was the suspected cause of death. However, there seemed to be no rupture of a cardiac wall or aortic dissection (Fig. 3b-e).
The aortic root was obtained for pathological examination. The pathological specimens showed cancer metastasis in the adventitia of the aortic root and pericardium (Fig. 4a,b). In order to determine the origin of the cancer, immunostaining was performed. Cytokeratin 7 was positive, but cytokeratin 20 was negative (Fig. 4c,d). Though lung adenocarcinoma was suspected, immunostaining for TTF-1 was negative (data not shown). The cause of death on the death certificate was acute pulmonary vascular failure. However, the examination showed cardiac tamponade as the immediate cause of death, and cancer metastasis as the cause of the cardiac tamponade.
Case 20. An 88-year-old woman with Hodgkin's lymphoma Bilateral pleural effusions (right > left) were noted on PMI. Mediastinal lymph nodes were markedly swollen (Fig. 5a). However, mediastinal lymph nodes could not be obtained, but pulmonary hilar lymph nodes were taken instead (Fig.  5b,c). On examination of the specimens of the pulmonary hilar lymph nodes, the lymph follicular structure was preserved, but nodular fibrosis was seen (Fig. 6a). Lymphocytes were infiltrating into the surrounding adipose tissue. Some  multinucleated giant cells were found in the background of small lymphocytes and fibroblasts (Fig. 6b,c). Some nuclei had prominent nucleoli, and binucleated Reed-Sternberg cells were also found. Hodgkin's lymphoma was suspected, and immunostaining was performed. The immunostaining showed that the binucleated Reed-Sternberg cells were positive for CD15 and CD30 (Fig. 6d,e), and Hodgkin's lymphoma was diagnosed.

DISCUSSION
Although the cadavers for anatomical dissection were in the fixed state, interpretation of PMI was more valuable than expected. It was possible to diagnose cardiac tamponade (Case No. 14) and tumor or nodular lesions (Case Nos. 1, 3, 8, 11, 13, 15, 20, 23, 29, 31 and 32) with PMI, similar to clinical imaging. With respect to ascites and pleural effusions, since the possibility of effusion of the fixing solution could not be ruled out, the evaluation was difficult.
Furthermore, with respect to the specimens made from the cadavers, not only hematoxylin and eosin staining, but also some immunostaining was possible. On immunostaining, reactivity of CK-7, CK-20, CD30, CD15, CD56, CEA, Chromogranin A, Synaptophysin, GCDFP-15, and CK5/6 was obtained, but TTF-1, Ki-67, androgen receptor, alpha-amylase, S-100, and p63 did not react. There might be a difference between cell surface antigens and nuclear antigens. A further examination of the reactivity of immunostaining of pathology specimens from cadavers for anatomical dissection may be required.
The total number of tissues and organs for histopathological evaluation might seem to be very small (from 1 to 7 only). The pathologists (SN and KI) sampled only the site which had abnormal findings on PMI in the PMI conference, and there was a time (temporal) limitation to have to obtain sample tissues and organs from many of the cadavers on the final day of the anatomy practice for the medical student. The pathologists knew the findings of the death certificate and PMI in PMI conference, and they therefore performed a minimum sampling based on these findings.
The death certificate of Japan is significant for two main reasons: (i) it proves human death medically and legally, and (ii) it provides mortality statistics by cause of death . However, the cause of death information obtained from death certificates is often inaccurate and incomplete. The accuracy of the underlying cause of death recorded on death certificates when multiple diseases are present in Japan has been discussed. 29 The concordance rate of major underlying cause of death (cancer, heart disease and pneumonia) reported on death certificates compared with a reference standard of pathologist assessment based on autopsy data and clinical records was evaluated. The results of this article were 'The concordance rate was relatively high for cancer (81%) but low for heart disease (55%) and pneumonia (9%).The overall concordance rate was 48%'. On the other hand, of the 33 cases in our study, the cause of death of 11 cases was not able to be determined by PMI findings and pathological analysis. Five cases (Case Nos.14, 15, 16, 17 and 30) had discrepancies among cause of death (COD) in the death certificate and COD by PMI findings and pathological analysis. The concordance rate was 77% (17 cases / 22 cases).
As for as Case 14, the discrepancy between COD in the death certificate and COD by total findings was explained above. Although heart failure should be avoided as the immediate COD on the death certificate, the immediate COD in  14. PMI-CT shows a pericardial effusion (a, arrow heads). The pericardial fluid forms a niveau (horizontal surface) that suggests bloody fluid. The macroscopic view of the heart shows no cardiac wall rupture on the posterior side (b) or the anterior side (c). There is no dissection of the aortic root (d). The pathology specimen was obtained from the aortic root (e). PMI-CT, postmortem imaging-computed tomography. the death certificate of Case 15 was chronic heart failure. We considered bronchopneumonia as the COD from the total findings. The underlying COD in the death certificate of Case 16 was pneumonia, but the liver pathologically showed liver cirrhosis. The immediate COD in the death certificate of Case 17 was pneumonia, but micro abscesses in myocardium were found. We suspect that sepsis was more accurate as the COD than pneumonia. As for as Case 30, we doubted that the 69 year-old-woman had 'died of old age' as recorded in the death certificate, however, pathological analysis revealed infectious renal cyst, pyonephrosis and perinephric abscess. Sepsis was suspected as COD from these findings.
Of the 33 cases, it was possible to presume (or guess) the causes of death from PMI alone in 12 (36.4%). This is almost equivalent to the rate of determining the cause of death from PMI alone in forensic police-related cases.
Throughout this series of pathological analyses, there were some interesting cases. Amyloidoma of the right femoral region was diagnosed in Case No. 13 by combined PMI and pathological examination. Soft tissue amyloidoma is a rare condition that presents primarily in the abdomen and/or mediastinum and more uncommonly on the extremities. [30][31][32] There have been only a handful of case reports since 1998. We were able to see an interesting case by doing a pathological analysis using PMI of cadavers for anatomical dissection.
In Case 14, although pericardial tamponade as the cause of death was diagnosed by PMI, the cause of the pericardial tamponade was not found. On pathological analysis using immunostaining, it was possible to deduce the cause of death in greater detail. Even in Case 20, it was possible to diagnose the Hodgkin's lymphoma correctly by doing the pathological analysis with immunostaining.
By performing PMI on cadavers for anatomical dissection and analyzing the pathological findings, it was possible to perform a pathological anatomy examination similar to what is called 'minimally invasive autopsy'. 33 In some cases, it was possible to identify interesting findings and understand the cause of death in more detail.
Moreover, it is thought to be important that students understand therapeutic signs and the pathological lesions of the cadaver when doing anatomical dissection to learn normal anatomy.
It is needless to say that pathological autopsy is very important to investigate the cause of death and the pathophysiological analysis of disease. However, a falling autopsy rate has been noted for quite a long time. The tendency for a decrease in the autopsy rate began in about the 1990s. In our hospital, there were 88 autopsies a year in 1989, which decreased to 34   Histopathological findings of the aortic root.Metastatic cancer is found in the adventitia of the aortic root (a). The cancer shows solid growth, and some cancer nests have central necrosis (b). Both specimens were stained with hematoxylin and eosin. The immunostaining for CK7 is positive (c), but CK20 is negative (d). Both specimens were counterstained with hematoxylin.
year corresponds to approximately 74% of the pathologic autopsies at our hospital in 2010.
As part of anatomy education, autopsy training is performed to study normal anatomy, but the cadavers dissected are never 'normal' humans. If a 'normal' human is necessary, only a simulator should be used. We also believe that it becomes one of the purposes of anatomy training to realize that the corpus being dissected was a human who suffered from illness when alive. In this way, the pathological perspective is required also in anatomy training.
A novel method of pathological analysis of cadavers for anatomical dissection using PMI without disturbing the anatomy education of medical students was presented in this paper. The pathological specimen obtained using PMI was able to be used not only for hematoxylin and eosin (HE) staining, but also some immunostaining. With this new method, it was possible to infer the cause of death in approximately 36.4% of cases. This method seems to be useful for understanding the pathophysiology and cause of death for not only the medical student, but also the pathologist.

DISCLOSURE STATEMENT
None declared.   20. Chest PMI-CT shows that mediastinal lymph nodes and right hilar lymph nodes are markedly swollen (dotted line area) (a). A pleural effusion with right-sided predominance is also seen. The macroscopic view of the right lung (b) and the cut surface of the right lung at the hilar level (c). A swollen hilar lymph node is seen. PMI-CT, postmortem imaging-computed tomography. Figure 6 Histopathological findings of the right hilar lymph node. In the specimens of the pulmonary hilar lymph nodes, lymph follicular structure is not preserved, but nodular fibrosis is seen (Fig. 6a). Some multinucleated giant cells are found in the background of small lymphocytes and fibroblasts (Fig. 6b,c). Some nuclei have prominent nucleoli, and binucleated Reed-Sternberg (R-S) cells are also found. The immunostaining shows that the binucleated R-S cells are positive for CD15 and CD30 (Fig. 6d,e). With these findings, the diagnosis is Hodgkin's lymphoma.