Yusuke Kawasumi

Diagnosis of drowning using post-mortem computed tomography based on the volume and density of fluid accumulation in the maxillary and sphenoid sinuses.

Recent studies have reported that drowning victims frequently have fluid accumulation in the paranasal sinuses, most notably the maxillary and sphenoid sinuses. However, in our previous study, many non-drowning victims also had fluid accumulation in the sinuses. Therefore, we evaluated the qualitative difference in fluid accumulation between drowning and non-drowning cases in the present study. Thirty-eight drowning and 73 non-drowning cases were investigated retrospectively. The fluid volume and density of each case were calculated using a DICOM workstation. The drowning cases were compared with the non-drowning cases using the Mann-Whitney U-test because the data showed non-normal distribution. The median fluid volume was 1.82 (range 0.02-11.7) ml in the drowning cases and 0.49 (0.03-8.7) ml in the non-drowning cases, and the median fluid density was 22 (-14 to 66) and 39 (-65 to 77) HU, respectively. Both volume and density differed significantly between the drowning and non-drowning cases (p=0.001, p=0.0007). Regarding cut-off levels in the ROC analysis, the points on the ROC curve closest (0, 1) were 1.03ml (sensitivity 68%, specificity 68%, PPV 53%, NPV 81%) and 27.5 HU (61%, 70%, 51%, 77%). The Youden indices were 1.03ml and 37.8 HU (84%, 51%, 47%, 86%). When the cut-off level was set at 1.03ml and 27.5HU, the sensitivity was 42%, specificity 45%, PPV 29% and NPV 60%. When the cut-off level was set at 1.03ml and 37.8HU, sensitivity was 58%, specificity 32%, PPV 31% and NPV 59%.

Assessment of the relationship between drowning and fluid accumulation in the paranasal sinuses on post-mortem computed tomography.

Recent reports have detailed common computed tomography (CT) findings in drowning victims, most notably fluid accumulation in the maxillary and sphenoidal sinuses. This CT finding could help forensic doctors to diagnose drowning. This study retrospectively investigated 151 subjects: 39 drowning and 112 non-drowning cases. Pearsons chi-square tests demonstrated that fluid accumulation in the maxillary or sphenoidal sinuses was associated significantly with drowning (p=0.0001). The sensitivity of the drowning diagnosis was 97%, specificity was 35%, accuracy was 51%, positive predictive value was 34% and negative predictive value was 98%. Drowning was significantly associated with fluid accumulation in the maxillary and sphenoidal sinuses, but the specificity and positive predictive value of the drowning diagnosis were poor. Although the presence of fluid in the maxillary and sphenoidal sinuses cannot be used to diagnose drowning, the absence of the fluid can be used to virtually exclude drowning.

Hypothermic death: Possibility of diagnosis by post-mortem computed tomography

Referring to our experience with post-mortem computed tomography (CT), many hypothermic death cases presented a lack of increase in lung-field concentration, blood clotting in the heart, thoracic aorta or pulmonary artery, and urine retention in the bladder. Thus we evaluated the diagnostic performance of post-mortem CT on hypothermic death based on the above-mentioned three findings. Twenty-four hypothermic death subjects and 53 non-hypothermic death subjects were examined. Two radiologists assessed the presence or lack of an increase in lung-field concentration, blood clotting in the heart, thoracic aorta or pulmonary artery, and measured urine volume in the bladder. Pearsons chi-square test and Mann-Whitney U-test were used to assess the relationship between the three findings and hypothermic death. The sensitivity, specificity, accuracy, positive predictive value (PPV) and negative predictive value (NPV) of the diagnosis were also calculated. Lack of an increase in lung-field concentration and blood clotting in the heart, thoracic aorta or pulmonary artery were significantly associated with hypothermic death (p=0.0007, p<0.0001, respectively). The hypothermic death cases had significantly more urine in the bladder than the non-hypothermic death cases (p=0.0011). Regarding the diagnostic performance with all three findings, the sensitivity was 29.2% but the specificity was 100%. These three findings were more common in hypothermic death cases. Although the sensitivity was low, these findings will assist forensic physicians in diagnosing hypothermic death since the specificity was high.

Cardiac output obtained from test bolus injections as a factor in contrast injection rate revision of following coronary CT angiography.

Background Optimal contrast enhancement is crucial for the detection of coronary artery stenoses and atherosclerotic changes in coronary CT angiography (CTA). Purpose To demonstrate the feasibility of using the cardiac output (CO) obtained from the test bolus injection data-set (COtest) as a factor in contrast injection rate revision of the following coronary CTA. Material and Methods The test bolus injection data-sets of 52 consecutive coronary CTAs were examined. COtest was calculated from the test bolus data-set. Aortic peak enhancement (APE) was measured on the following coronary CTA. We simulated the APE at a fixed contrast injection rate of 4 mL/s (simAPE) in each patient. Results The ranges of COtest and simAPE were 2.82-7.56 L/min and 194-527 Hounsfield Units, respectively. There was a significant negative correlation (R = -0.802, P < 0.001) between simAPE and COtest. Conclusion COtest can be used for injection rate revision on coronary CTA.

Mass detection using deep convolutional neural network for mammographic computer-aided diagnosis

In recent years, a deep convolutional neural network (DCNN) has attracted great attention due to its outstanding performance in recognition of natural images. However, the DCNN performance for medical image recognition is still uncertain because collecting a large amount of training data is difficult. To solve the problem of the DCNN, we adopt a transfer learning strategy, and demonstrate feasibilities of the DCNN and of the transfer learning strategy for mass detection in mammographic images. We adopt a DCNN architecture that consists of 8 layers with weight, including 5 convolutional layers, and 3 fully-connected layers in this study. We first train the DCNN using about 1.2 million natural images for classification of 1,000 classes. Then, we modify the last fully-connected layer of the DCNN and subsequently train the DCNN using 1,656 regions of interest in mammographic image for two classes classification: mass and normal. The detection test is conducted on 198 mammographic images including 99 mass images and 99 normal images. The experimental results showed that the sensitivity of the mass detection was 89.9 % and the false positive was 19.2 %. These results demonstrated that the DCNN trained by transfer learning strategy has a potential to be a key system for mammographic mass detection computer-aided diagnosis (CAD). In addition, to the best of our knowledge, our study is the first demonstration of the DCNN for mammographic CAD application.

Postmortem computed tomography images of a broken piece of a weapon in the skull

The use of post-mortem radiological imaging is becoming increasing widespread in forensic medicine. These images can help improve the quality and safety of autopsy. We report two cases of homicide victims that had broken pieces of a weapon in their skull. Postmortem multislice computed tomography (MSCT) demonstrated metal artifacts in the skull of two homicide cases. Autopsy showed that both artifacts were pieces broken off of a sharp instrument. Some possible benefits of postmortem MSCT include the ability to reconstruct visual images of weapons and the possible prevention of injury to autopsy personnel.

High-resolution monochrome liquid crystal display versus efficient household colour liquid crystal display: comparison of their diagnostic performance with unenhanced CT images in focal liver lesions

This study compared the ability of an efficient household liquid crystal display (LCD) and a medical high-resolution LCD to detect small hepatic lesions on unenhanced computed tomography (CT) images. We obtained the images from 100 subjects who had undergone abdominal CT. They consisted of 41 patients with a single space-occupying lesion (SOL) in the liver and 59 control subjects with no SOL. Independently, five radiologists rated their confidence concerning the presence of hepatic SOLs on a continuous scale from 0 to 1. Receiver-operating characteristic (ROC) analysis was performed using the jackknife method using the program LABMRMC. We evaluated the differences in Az based on the 95% confidence intervals. The mean Az of the five observers was 0.9594 with the efficient household LCD vs. 0.9335 with the medical high-resolution LCD. The difference was –0.0422, and the 95% confidence interval was –0.1101 to 0.0257 (p=0.2203). There was no significant difference in the Az value between the two types of LCDs. The diagnostic performance with the household LCD was comparable to that with the high-resolution LCD, implying that the former type of LCD can be used to diagnose CT images.

Radiological analysis of a naturally mummified body

A body in an advanced stage of mummification was found in a concrete apartment in Japan. Natural complete mummification is very rare in Japan’s humid and temperate climate. We performed multi-slice computed tomography (MSCT) to obtain information on the body prior to autopsy. MSCT clearly illustrated old fractures in the right lower leg. Bone resists destruction during body decomposition. If antemortem medical records of the deceased are available, MSCT scanning can provide information for positive identification. At autopsy, the cervical tissues presented as a dried mass, and it was difficult to separately remove the hyoid bone and thyroid cartilage. Fractures of the large horn of the hyoid bone and superior horn of the thyroid cartilage, which are not observed in all strangulation cases, strongly suggest criminal activity. The diagnosis of these fractures is of great value. In our case, MSCT revealed that there was no fracture in the hyoid bone or ossified area of the thyroid cartilage. Hard tissues are usually well preserved in mummies. Although MSCT images have limits in mummies because of the severe dryness of soft tissues and organs, they could become a useful tool not only for personal identification, but also for the identification of neck compression.

Usefulness of postmortem computed tomography before forensic autopsy for alerting forensic personnel to tuberculosis infection

Since May 2009, we have performed multislice computed tomography (MSCT) prior to forensic autopsy for cases of suspicious death. In the present case, innumerable widely scattered nodules in both pulmonary fields on MSCT were indicative of miliary tuberculosis (TB). At autopsy, both lungs were submerged in formalin fluid immediately after removal from the body. Miliary TB was finally diagnosed based on microscopic findings. TB is a disease that autopsy room workers need to be aware of to protect themselves. Unfortunately, because little medical information about deceased individuals is usually available before forensic autopsy, the diagnosis of TB is frequently not made until autopsy. This leads to a much higher incidence of TB in autopsy room staff members even if they wear protective clothing. Therefore, MSCT before forensic autopsy may identify suspected cases of miliary TB in advance and thus help to prevent TB infection in forensic autopsy personnel.

Heat haematoma: Post-mortem computed tomography findings

A heat haematoma is an epidural haematoma often found in severely burned bodies as a heat-induced artefact.1e4 This haematoma is observed with intense burns of the head, which cause the dura mater to contract and exfoliate from the skull, resulting in the exudation of blood from the venous sinuses. A heat haematoma is formed when heat causes this exuded blood to coagulate. Macroscopically, a heat haematoma is brick red. It is not a vital reaction. In forensics, one must carefully distinguish a heat haematoma from an epidural haematoma caused by blunt trauma. Given the current wide use of post-mortem computed tomography (CT), the examiner should be able to distinguish these two haematomas on CT. The CT findings of heat haematomas are seldom reported.5,6 We report a case of heat haematoma in a post-mortem CT examination.