Diagnostic Accuracy of Liver Damage Based on Postmortem Computed Tomography Findings in High-Energy Trauma

Abstract

The liver is an organ that is frequently injured by blunt trauma. In clinical medicine, contrast-enhanced computed tomography (CT) is useful for diagnosing liver damage. However, detection of liver injury is difficult with postmortem CT (PMCT) that does not use contrast media. This study aims to identify findings that are useful for diagnosing liver injury with PMCT. This study included 97 high-energy trauma cases that underwent both PMCT without contrast and forensic autopsy between July 2009 and November 2020. PMCT findings in and around the liver in cases of liver injury were collected. The diagnostic accuracy of each finding was calculated. Of 97 cases, 62 had liver injury on autopsy. PMCT detected hepatic surface gas in 31 cases, intrahepatic focal gas in 19 cases, high-density fluid accumulation around the liver in 12 cases, and fracture of a right lower (9th-12th) rib in 48 cases. Abnormal PMCT findings in and around the liver had insufficient diagnostic sensitivity, ranging from 19.4% to 38.7%. By contrast, the finding of a right lower rib fracture was more sensitive for suspected liver injury. Our results indicate that abdominal liver findings (hepatic surface gas, intrahepatic focal gas, and high-density fluid accumulation around the liver) are of limited utility for the diagnosis of liver injury, whereas a right lower rib fracture is a useful indirect finding.

Introduction

Postmortem computed tomography (PMCT) is widely performed and plays an important role in forensic medicine (Poulsen and Simonsen 2007; Bolliger et al. 2008; Usui et al. 2012). As the number of autopsies is decreasing worldwide (Shojania and Burton 2008), the need to obtain accurate information based only PMCT images is expected to increase. The liver is the most commonly injured abdominal organ due to blunt trauma. Some studies in PMCT suggest that liver injuries occurred frequently in cases of traffic accidents and falls from height (Cichon and Schön 2012), which are considered types of high-energy trauma (Di Paolo et al. 2020). Contrast-enhanced CT scans can accurately detect damage to the liver parenchyma (Yoon et al. 2005). Lacerations, contusions, and hematomas appear as non-contrast-enhanced areas (Shanmuganathan and Mirvis 1998; Romano et al. 2004). Active hemorrhage can be assessed according to the collection of extravasated contrast material (Fang et al. 1998). However, because contrast media are rarely used in PMCT, detecting damage with PMCT might be difficult even when liver damage is obvious at autopsy. In addition, contrast-enhanced CT is difficult to perform on critically injured bodies due to the significant decrease in circulating blood and injured blood vessels. However, obtaining accurate information about liver injury, which is commonly caused by trauma, is considered important in determining the cause of traumatic deaths.

With regards to PMCT diagnosis, there have been reports about liver injury (Christe et al. 2009; Carballeira Álvarez et al. 2018). No analyses of the diagnostic ability of PMCT findings in the liver have been performed. Incidentally, some clinical reports have examined the relationship between liver injury and rib fractures (Shweiki et al. 2001; Al-Hassani et al. 2010; Bhattacharya et al. 2015; Rostas et al. 2017). In a PMCT study involving falls, the presence of a lower right (9th-12th) rib fracture was reported to be strongly associated with liver injury (Kasagawa et al. 2020). In this study, we evaluated the diagnostic ability of liver injury based on PMCT and evaluated findings that might be useful for improving diagnostic ability.

Materials and Methods

We retrospectively investigated 97 high-energy trauma cases that underwent PMCT before forensic autopsy between July 2009 and November 2020 at our institution. High-energy trauma in this study included traffic accidents (drivers, motorcyclists, cyclists, and pedestrians) and falls from height (Fig. 1). Of the 97 cases included in this study, 62 had liver injury at autopsy. Diagnostic ability was calculated by comparing the presence or absence of PMCT findings between the 62 cases with liver damage confirmed at autopsy and the 35 cases with no liver damage based on autopsy.

At our institution, PMCT scanning has been performed as a part of pre-autopsy screening using a multi-detector row CT scanner (Aquilion; Canon Medical Systems, Otawara, Japan) since 2009. By December 2021, we had performed PMCT on 2,260 bodies. All were scanned while wrapped in body bags, with the upper extremities in a lower position due to the effects of rigor mortis. Contrast media was not used. The tube voltage was 135 kV. The tube current was changed depending on body size by an automated tube current modulation system. A series of 2-mm-thick images were reconstructed without beam hardening correction. An expert radiological technologist with more than 10 years of PMCT experience performed all of the CT scans.

A forensic pathologist with over 30 years of experience and colleagues performed conventional autopsies after PMCT. All CT images were evaluated on a two-dimensional digital imaging and communications in medicine viewer (POP-Net Server; ImageONE, Tokyo, Japan) and a three-dimensional digital imaging and communications in medicine workstation (Ziostation™ 2, version 2.4.3.3; ZIOSOFT, Tokyo, Japan).

Three observers (two radiologists and a radiological technologist) evaluated CT images. One radiologist (Rater 1) had more than 25 years of experience in clinical diagnostic radiology and 4 years in PMCT. The other radiologist (Rater 2) had over 15 years of experience in PMCT. The radiological technologist (Rater 3) had over 10 years of experience in providing diagnostic assistance for PMCT. They focused on the following three CT imaging findings suspicious of liver damage observed when reevaluating liver injury cases in our institution and right lower (9th-12th) rib fractures: (i) hepatic surface gas (Fig. 2) appearing as small bubbles, or a flat hepatic surface with gas probably under the hepatic capsule; (ii) intrahepatic focal gas (Fig. 3), which is an abnormal gas collection in the parenchyma different from the gas pattern along blood vessels and bile ducts seen as postmortem changes; and (iii) accumulation of fluid around the liver with a higher CT value than water (Fig. 4). Complete fractures of any right lower ribs were judged as the presence of a rib fracture. Obvious obsolete and incomplete fractures were excluded. In cases with incomplete agreement among observers, the presence or absence of each finding was decided based on a majority vote.

We evaluated interobserver correlation using the kappa statistic. To evaluate the diagnostic ability of each finding, sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV) were calculated based on autopsy findings. The 95% confidence intervals were calculated using the Clopper-Pearson exact method. Statistical analyses were conducted with JMP Pro 16.0.0 (SAS Institute Inc., Cary, NC, USA). This study was approved by the ethics committee of our institution (approval number, 2020-1-653).

Fig. 1.

Study flow diagram.

Fig. 2.

Hepatic surface gas.

Areas of gas were seen on the surface of the liver (arrow).

Fig. 3.

Intrahepatic focal gas.

Focal gas that does not follow the vascular tree was seen in the liver (arrow).

Fig. 4.

High-density fluid accumulation around the liver.

There were fluid accumulations around the liver with CT values higher than water (arrow).

Results

Of 2,042 consecutive cases that underwent PMCT before forensic autopsy between July 2009 and November 2020 at our institution, this study included 97 high-energy trauma cases. Seven cases were excluded: one case involved an infant, two cases involved severe decomposition in which the liver parenchyma was filled with putrefaction gas and had a sponge-like appearance, one case involving whole-body disruption with difficulty in identifying liver parenchyma, and three cases in which a few weeks had passed from death (Fig. 1). Of the 97 cases, 38 resulted from traffic accidents (13 drivers, 5 motorcyclists, 2 cyclists, and 18 pedestrians) and 59 were associated with falls from height. There were 74 males and 23 females. The mean age was 49 years (range, 13-87 years). Of the 62 cases with liver injury, 22 resulted from traffic accidents (8 drivers, 3 motorcyclists, and 11 pedestrians) and 40 were associated with falls from height. The time interval from death to PMCT was less than 24 hours in 45 cases, 24 hours to a few days in 50 cases, a few days to one week in 1 case, and unknown in 1 case. Table 1 shows the interobserver correlations.

Among cases with liver injury confirmed at autopsy, hepatic surface gas was observed in 24 cases and not observed in 38 cases. Among cases with no confirmed liver damage, hepatic surface gas was observed in 7 cases and not observed in 28 cases on PMCT. As shown in Table 2, the diagnostic performance of hepatic surface gas was as follows: sensitivity, 38.7%; specificity, 80.0%; PPV, 77.4 %; and NPV, 42.4%.

Similarly, among cases with liver injury confirmed at autopsy, intrahepatic focal gas was observed in 18 cases and not in 44 cases. Among cases with no confirmed liver damage, intrahepatic focal gas was observed in 1 case and not observed in 34 cases on PMCT. The diagnostic performance of intrahepatic focal gas was as follows: sensitivity, 29.0%; specificity, 97.1%; PPV, 94.7%; and NPV, 43.6% (Table 2).

Among cases with liver injury confirmed at autopsy, high-density fluid accumulation around the liver was observed in 12 cases and not observed in 50 cases. Among cases with no liver injury at autopsy, none had high-density fluid accumulation around the liver. The diagnostic performance of high-density fluid accumulation around the liver was as follows: sensitivity, 19.4%; specificity, 100.0%; PPV, 100.0%; and NPV, 41.2% (Table 2).

Among cases with liver injury confirmed at autopsy, right-sided lower rib fracture was observed in 41 cases and not in 21 cases. Among cases with no liver injury at autopsy, a right-sided lower rib fracture was observed in 7 cases and not observed in 28 cases on PMCT. The diagnostic performance of a right lower rib fracture was as follows: sensitivity, 66.1%, specificity, 80.0%; PPV, 85.4%; and NPV, 57.1% (Table 2).

Table 1.

Interobserver correlation coefficients and 95% confidence intervals for each postmortem finding.

Table 2.

Diagnostic performance for each postmortem finding.

CI, confidence interval; PPV, positive predictive value; NPV, negative predictive value.

Discussion

PMCT had insufficient diagnostic sensitivity based on abnormal findings in and around the liver. By contrast, the finding of a right lower rib fracture was more sensitive for suspected liver injury.

In clinical practice, ectopic gas is a useful imaging finding for the diagnosis of potential lesions (Leturia Etxeberria et al. 2020). In PMCT images, abnormal gas distribution in the parenchyma of the liver that does not follow the vessel tree is an indicator of suspected liver injury (Carballeira Álvarez et al. 2018; Chatzaraki et al. 2021). Focal intrahepatic gas in this study was considered the same as this type of abnormal gas distribution. Both the specificity and PPV of intrahepatic focal gas for the diagnosis of liver injury exceeded 90%, suggesting that abnormal gas distribution in the liver parenchyma is a suspicious finding for liver injury, as in previous studies (Carballeira Álvarez et al. 2018; Chatzaraki et al. 2021). Although there are no reports of similar findings with regards to hepatic surface gas, this finding also suggests liver injury, with specificity 80%. One of the characteristic imaging findings for bleeding source determination in hemoperitoneum is hyperdense fluid accumulation around the abdominal organs (Chatzaraki et al. 2021). This finding appears to be consistent with high-density fluid accumulation around the liver in this study. Liver injury was confirmed in all patients with high-density fluid accumulation around the liver; thus, this finding is considered a suspicious finding for liver injury. However, in this study, these three liver findings were not sufficiently sensitive; many liver injuries were missed based on PMCT. Furthermore, interobserver correlation was not high. This result might have been due to intravascular gas or artifacts that prevent the detection of findings on PMCT. Intravascular gas reportedly results from postmortem changes and cardiopulmonary resuscitation (CPR) (Zenda et al. 2011). Hepatic surface gas and intrahepatic focal gas are extravascular gas, which is distinguishable from intravascular gas. When scrolling and observing the anterior and posterior slices of the gas in a CT image plane on workstation, continuity of gas is observed for intravascular gas, but not for extravascular gas. However, it might be difficult for observers inexperienced in the diagnosis of PMCT to make this distinction. The presence of gas is difficult to confirm during conventional autopsy (Femia et al. 2021). Artifacts from the upper limbs and ribs might interfere with the detection of high-density fluid accumulation around the liver.

The presence of right lower rib fractures on PMCT is highly suspicious of concomitant liver injury (Kasagawa et al. 2020). In this study, the finding of a right lower rib fracture was more sensitive to liver damage than the three liver findings; moreover, interobserver correlation was excellent. Rib fractures caused by CPR occur mainly in the 2nd-7th ribs (Yang et al. 2011). Therefore, the right lower ribs, the focus of this study, are unlikely to be affected.

Although there have been reports of contrast media use in PMCT (Chevallier et al. 2013; Ross et al. 2014; Grabherr et al. 2015), this method is rarely performed in practice because blood circulation has ceased and there is often postmortem clotting in the vessels. Therefore, the finding of a right lower rib fracture, which is not affected by postmortem changes or artifacts and can be detected even without contrast, might be important in the diagnosis of liver injury with PMCT. In clinical cases, organ damage might also be predicted by the presence of rib fractures (Shweiki et al. 2001; Al-Hassani et al. 2010; Bhattacharya et al. 2015; Rostas et al. 2017). In postmortem imaging, comparison with autopsy results is very important for developing future diagnostic techniques, although liver damage can be easily imagined or estimated at a glance without PMCT during autopsy.

This study has some limitations. The sample size was relatively small. Future studies with more cases might be necessary. Nevertheless, the number of forensic autopsies is declining worldwide (Shojania and Burton 2008), making it difficult to accumulate a sufficient number of cases involving high-energy trauma with liver injury. Furthermore, liver injury in this study was considered to be the result of high-energy trauma. The diagnosis of liver injury caused by other factors should be considered separately. In addition, this study was performed retrospectively.

In conclusion, findings that indirectly suggest liver injury, such as a right lower rib fracture, might be important in the diagnosis of liver injury in high-energy trauma. Such sensitive information is helpful not only before autopsy, but also when autopsy cannot be performed due to religious or cultural reasons and the cause of death must be determined based only on imaging.

Acknowledgments

The authors thank the staff of the Department of Forensic Medicine, Tohoku University Graduate School of Medicine, for providing the results of the forensic autopsy.

This work was supported by Japan Society for the Promotion of Science (JSPS) KAKENHI grant number JP20K12591.

Conflict of Interest

The authors declare no conflict of interest.

References

•  Al-Hassani,  A.,  Abdulrahman,  H.,  Afifi,  I.,  Almadani,  A.,  Al-Den,  A.,  Al-Kuwari,  A.,  Recicar,  J.,  Nabir,  S. &  Maull,  K.I. (2010) Rib fracture patterns predict thoracic chest wall and abdominal solid organ injury. Am. Surg., 76, 888-891.
•  Bhattacharya,  B.,  Fieber,  J.,  Schuster,  K.,  Davis,  K. &  Maung,  A. (2015) “Occult” rib fractures diagnosed on computed tomography scan only are still a risk factor for solid organ injury. J. Emerg. Trauma Shock, 8, 140-143.
•  Bolliger,  S.A.,  Thali,  M.J.,  Ross,  S.,  Buck,  U.,  Naether,  S. &  Vock,  P. (2008) Virtual autopsy using imaging: bridging radiologic and forensic sciences. A review of the Virtopsy and similar projects. Eur. Radiol., 18, 273-282.
•  Carballeira Álvarez,  A.,  Mancini,  J.,  Tuchtan-Torrents,  L.,  Gach,  P.,  Bartoli,  C.,  Desfeux,  J.,  Piercecchi,  M.D. &  Gorincour,  G. (2018) Diagnostic value of unenhanced postmortem computed tomography in the detection of traumatic abdominal injuries. Diagn. Interv. Imaging, 99, 397-402.
•  Chatzaraki,  V.,  Thali,  M.J. &  Ampanozi,  G. (2021) Diagnostic accuracy of postmortem computed tomography for bleeding source determination in cases with hemoperitoneum. Int. J. Legal Med., 135, 593-603.
•  Chevallier,  C.,  Doenz,  F.,  Vaucher,  P.,  Palmiere,  C.,  Dominguez,  A.,  Binaghi,  S.,  Mangin,  P. &  Grabherr,  S. (2013) Postmortem computed tomography angiography vs. conventional autopsy: advantages and inconveniences of each method. Int. J. Legal Med., 127, 981-989.
•  Christe,  A.,  Ross,  S.,  Oesterhelweg,  L.,  Spendlove,  D.,  Bolliger,  S.,  Vock,  P. &  Thali,  M.J. (2009) Abdominal trauma--sensitivity and specificity of postmortem noncontrast imaging findings compared with autopsy findings. J. Trauma, 66, 1302-1307.
•  Cichon,  K. &  Schön,  C.A. (2012) Influences on the amount of intraperitoneal haemorrhage after blunt liver injury: a retrospective autopsy study. Eur. J. Gastroenterol. Hepatol., 24, 1333-1340.
•  Di Paolo,  M.,  Maiese,  A.,  dell’Aquila,  M.,  Filomena,  C.,  Turco,  S.,  Giaconi,  C. &  Turillazzi,  E. (2020) Role of post mortem CT (PMCT) in high energy traumatic deaths. Clin. Ter., 171, e490-e500.
•  Fang,  J.F.,  Chen,  R.J.,  Wong,  Y.C.,  Lin,  B.C.,  Hsu,  Y.B.,  Kao,  J.L. &  Kao,  Y.C. (1998) Pooling of contrast material on computed tomography mandates aggressive management of blunt hepatic injury. Am. J. Surg., 176, 315-319.
•  Femia,  G.,  Langlois,  N.,  Raleigh,  J.,  Gray,  B.,  Othman,  F.,  Perumal,  S.R.,  Semsarian,  C. &  Puranik,  R. (2021) Comparison of conventional autopsy with post-mortem magnetic resonance, computed tomography in determining the cause of unexplained death. Forensic Sci. Med. Pathol., 17, 10-18.
•  Grabherr,  S.,  Grimm,  J.,  Baumann,  P. &  Mangin,  P. (2015) Application of contrast media in post-mortem imaging (CT and MRI). Radiol. Med., 120, 824-834.
•  Kasagawa,  A.,  Hirakawa,  N.,  Usui,  A.,  Kawasumi,  Y.,  Funayama,  M. &  Saito,  H. (2020) Relationship between rib fractures and thoracoabdominal organ injury in cases of fatal falls from heights. Res. Pract. Forens. Med., 63, 55-62 (abstruct in English).
•  Leturia Etxeberria,  M.,  Biurrun Mancisidor,  M.C.,  Ugarte Nuno,  A.,  Arenaza Choperena,  G.,  Mendoza Alonso,  M.,  Esnaola Albizu,  M.,  Serdio Mier,  A.,  Gredilla Saenz,  M. &  Gomez Usabiaga,  V. (2020) Imaging assessment of ectopic gas collections. Radiographics, 40, 1318-1338.
•  Poulsen,  K. &  Simonsen,  J. (2007) Computed tomography as routine in connection with medico-legal autopsies. Forensic Sci. Int., 171, 190-197.
•  Romano,  L.,  Giovine,  S.,  Guidi,  G.,  Tortora,  G.,  Cinque,  T. &  Romano,  S. (2004) Hepatic trauma: CT findings and considerations based on our experience in emergency diagnostic imaging. Eur. J. Radiol., 50, 59-66.
•  Ross,  S.G.,  Bolliger,  S.A.,  Ampanozi,  G.,  Oesterhelweg,  L.,  Thali,  M.J. &  Flach,  P.M. (2014) Postmortem CT angiography: capabilities and limitations in traumatic and natural causes of death. Radiographics, 34, 830-846.
•  Rostas,  J.W.,  Lively,  T.B.,  Brevard,  S.B.,  Simmons,  J.D.,  Frotan,  M.A. &  Gonzalez,  R.P. (2017) Rib fractures and their association with solid organ injury: higher rib fractures have greater significance for solid organ injury screening. Am. J. Surg., 213, 791-797.
•  Shanmuganathan,  K. &  Mirvis,  S.E. (1998) CT scan evaluation of blunt hepatic trauma. Radiol. Clin. North Am., 36, 399-411.
•  Shojania,  K.G. &  Burton,  E.C. (2008) The vanishing nonforensic autopsy. N. Engl. J. Med., 358, 873-875.
•  Shweiki,  E.,  Klena,  J.,  Wood,  G.C. &  Indeck,  M. (2001) Assessing the true risk of abdominal solid organ injury in hospitalized rib fracture patients. J. Trauma, 50, 684-688.
•  Usui,  A.,  Kawasumi,  Y.,  Hosokai,  Y.,  Hayashizaki,  Y.,  Funayama,  M. &  Saito,  H. (2012) Usefulness of postmortem computed tomography before forensic autopsy for alerting forensic personnel to tuberculosis infection. Jpn. J. Radiol., 30, 612-615.
• Yang, K.M., Lynch, M. & O’Donnell, C. (2011) “Buckle” rib fracture: an artifact following cardio-pulmonary resuscitation detected on postmortem CT. Leg. Med. (Tokyo), 13, 233-239.
•  Yoon,  W.,  Jeong,  Y.Y.,  Kim,  J.K.,  Seo,  J.J.,  Lim,  H.S.,  Shin,  S.S.,  Kim,  J.C.,  Jeong,  S.W.,  Park,  J.G. &  Kang,  H.K. (2005) CT in blunt liver trauma. Radiographics, 25, 87-104.
•  Zenda,  T.,  Takayama,  T.,  Miyamoto,  M.,  Yamaguchi,  S.,  Endo,  T. &  Inaba,  H. (2011) Intravascular gas in multiple organs detected by postmortem computed tomography: effect of prolonged cardiopulmonary resuscitation on organ damage in patients with cardiopulmonary arrest. Jpn. J. Radiol., 29, 148-151.