Validity of Computerized Tomography in Blunt Renal Trauma

0022-5347/03/1706-2475/0 THE JOURNAL OF UROLOGY® Copyright © 2003 by AMERICAN UROLOGICAL ASSOCIATION

Vol. 170, 2475–2479, December 2003 Printed in U.S.A.

DOI: 10.1097/01.ju.0000095967.26733.2f

VALIDITY OF COMPUTERIZED TOMOGRAPHY IN BLUNT RENAL TRAUMA THOMAS BSCHLEIPFER, DIMITRIOS KALLIERIS, PETER HALLSCHEIDT, EKKEHARD W. HAUCK,* WOLFGANG WEIDNER AND REINER A. PUST From the Department of Urology, Ulm Military Hospital, (TB, RAP), Ulm/Donau, Institute of Forensic Medicine (DK) and Department of Diagnostic Radiology (PH), Ruprecht Karls University, Heidelberg and Department of Urology, Justus Liebig University (EWH, WW), Giessen, Germany

ABSTRACT

Purpose: Improved imaging techniques and new therapeutic possibilities require rethinking the indication for laparotomy with regard to blunt renal trauma. Refined classification systems would facilitate the decision relating to therapy but they are based on knowledge of the imaging accuracy of computerized tomography (CT). We evaluated the validity of the CT depiction of renal injuries. Materials and Methods: A total of 42 porcine kidneys were subjected to traumatization of various degrees. They then underwent CT examination and were subsequently cross-dissected into slices 3 mm thick. The comparative evaluation involved 2,080 CT images and 1,819 macroscopic sectional views, which showed 3,521 and 3,778 individual lesions, respectively. Results: Using CT the overall extent of injury in renal trauma was only slightly overrated at an average of 15% higher than that seen on macroscopy. Simple linear lesions tended to be over assessed and parenchymal destruction tended to be under assessed. Central lesions were depicted more frequently than peripheral lesions. CT of medullary lesions and parenchymal detachment was not feasible. Conclusions: CT of the kidney enables the distinction of different kinds of lesions and their localization well. Pelvic structures or vessels can imitate linear lesions. However, this imaging procedure can be used as a basis for refining categorization systems for blunt renal trauma. It can also be used to obtain a large quantity of lesion data for biomechanical investigations. KEY WORDS: kidney; swine; wounds and injuries; tomography, emission-computed

Computerized tomography (CT) is the diagnostic modality of choice with regard to blunt renal trauma.1, 2 It is regarded as the most informative, most sensitive and most specific method, enabling the simultaneous assessment of all abdominal organs and the detection of concomitant injuries.3 The use of improved imaging techniques and the application of new therapeutic methods, such as minimally invasive techniques and laparoscopy, require rethinking indications for laparotomy.4 – 8 Trauma to the kidney and its appearance on CT is categorized mostly according to the organ injury scale for the kidney or modifications of it.1, 9 –12 Five grades of trauma severity are distinguished, whereas more cases are being reported in which not only grades I to III injuries are treated conservatively, but also grade IV and in some instances even grade V.6 In this respect a further subdivision of existing classification systems relating to blunt renal trauma would facilitate the decision of whether to operate.3 Such refinement can only be performed as far as the new classifications can be distinguished by CT. To date in the literature no study has shown how exact individual lesions of the parenchyma can be depicted. In regard to biomechanical investigations it would be useful if a high degree of congruence between macroscopic and CT depiction of lesions could be proved. The time-consuming dissection of organs could be omitted and within a short period it would be possible to obtain large quantities of data. It could also be used for the further development of computer

assisted trauma simulation programs, which currently attempt to calculate the behavior of isolated organs using finite element analysis13, 14 but 3-dimensional imaging is not yet feasible. We critically reviewed the validity of CT and compared imaging with macroscopic findings actually available at different degrees of parenchymal traumatization. MATERIALS AND METHODS

A total of 42 kidneys from slaughtered pigs were removed from the adipose capsule, maintaining the renal capsule intact. Immediately after removal the vascular system was perfused using a heparin solution to avoid thrombogenesis or dissolve already existing thrombi. The kidneys were traumatized by a gravitational impactor (Institute of Forensic Medicine, Ruprecht Karls University, Heidelberg, Germany) about 12 to 24 hours postmortem under almost physiological conditions, as described previously.15 Infusion solution perfused the vascular system at a pressure of approximately 100 mm Hg and was passed into the renal pelvis and through the ureter. The renal pelvis was filled with fluid throughout the experiment. Energy application was performed using a gravitational impactor weighing 1,450 gm with a disc 10 cm in diameter. Loading was done on the dorsal side of the kidneys bluntly and only once per organ. Groups of 14 organs each were subjected to a load of 4.3, 8.5 and 12.8 J. The different loadings were enabled by dropping the impactor from different heights (0.3, 0.6 and 0.9 m, respectively. Imaging diagnosis was subsequently performed by a Somatom Plus (Siemens AG, Erlangen, Germany) CT unit. The gentry angle was 0 degrees with a table advance rate of 3 mm. The voltage

Accepted for publication June 20, 2003. * Corresponding author: Department of Urology, Justus Liebig University, Giessen, Rudolf-Buchheim-Str. 7, 35385 Giessen, Germany (telephone: ⫹49-641-99-44501; FAX: ⫹49-641-99-44509; e-mail: [email protected]). 2475

2476

COMPUTERIZED TOMOGRAPHY FOR BLUNT RENAL TRAUMA

calculated for each sectional view was 137 kV and the voltage of electricity was 275 mAs. The organs were then cross-dissected into disciform slices with a thickness of 3 mm, which enabled direct comparison between macroscopic visible lesions and CT images. A total of 2,080 CT images and 1,819 macroscopic sectional views were evaluated. The type and localization of each lesion were recorded as well as the extent of their depth within the parenchyma using a standardized protocol. The severity of each lesion was determined by a score calculated by a basic severity score for each type of lesion and a factor based on injury depth in relation to renal cortex thickness (tables 1 and 2). The lesion severity score was between 1 and 20. For each kidney a summation score was calculated that defined the total extent of macroscopically detected and CT depicted injuries. It was also possible to calculate such a summation score for each type or localization of lesions. It enabled highly detailed comparison between the actual lesions and their CT depiction. Direct injection of contrast medium into the vascular system of isolated, traumatized kidneys did not result in complete organ perfusion. Injected contrast medium escaped through superficial lesions. This problem was circumvented by 2 procedural techniques. In advance of traumatization 7 organs of each load category group were injected with a mixture of contrast medium and gelatin. In the second half of the group shortly prior to CT but after traumatization contrast medium was injected directly into the vascular system by a system specially developed to imitate a perirenal tamponade. The mixture of contrast medium and gelatin was prepared from 15 gm gelatin powder (Ernst Mu¨ ller u. Co., Neutraubling, Germany) and 40 ml Ultravist 300 (Schering, Berlin, Germany) (0.623 iopromide per ml) per l infusion solution. After injection into the renal artery brief refrigeration ensured rapid curing of the gelatin and its remaining inside the kidney vascular system. Imitating a perirenal tamponade enabled the direct injection of contrast medium into the vascular system of traumatized kidneys. An incoming and outgoing infusion system was inserted through the lid of a sealable preserving jar and attached to cannulas bound to the renal artery and renal vein of the organ used in the experiment. The preserving jar (volume 3,000 ml) was pre-filled with 2,000 ml infusion solution and 200 ml (Ultravist 300) contrast medium. After the preserving jar was sealed the contrast medium was injected via the incoming leg of the tubular system. The leakage of injected fluid through the superficial dehiscence led to compression of the air trapped in the preserving jar, thus, producing an increase in perirenal pressure. It enabled perfusion of the whole organ. Since we were using isolated kidneys, it was not possible to simulate systemic application and distribution of contrast medium. Thus, higher concentrations of contrast medium, the absence of different phases in opacification and the lack of concentration ability of the kidneys had to be accepted. All CT images were read by the same examiner. The macroscopic investigation was done at a different time. Reference studies showed that no additional lesions were caused by the preparatory measures, particularly the injection of contrast medium or contrast medium with gelatin. TABLE 1. Basic severity score for each renal lesion type Lesion Type

Basic Score

Longitudinal disruption Medullary lesion Shearing lesion Lesion cluster Transverse disruption Parenchymal destruction Parenchymal detachment from pelvis

1 1 2 3 4 4 4

TABLE 2. Factors based on injury depth relative to renal cortex thickness Renal Injury Depth Lesions 1⁄3 or less parenchymal thickness Lesions greater than 1⁄3 and 2⁄3 or less parenchymal thickness Lesions greater than 2⁄3 parenchymal thickness ⫹ no complete renal cortex penetration Complete renal cortex penetration down to medullary pyramid base Complete renal cortex penetration down to renal pelvis

Factor 1 2 3 4 5

RESULTS

A total of 3,778 macroscopically detected lesions were compared with 3,521 lesions depicted by CT. Macroscopically it was possible to distinguish longitudinal disruptions (fig. 1, a1), transverse disruptions (fig. 1, b1), shearing lesions (fig. 1, c1), lesion clusters (fig. 1, d1), parenchymal destruction (fig. 1, e1), detachments of the parenchyma from the pelvis (fig. 1, f1) and medullary lesions (fig. 1, g1). CT also showed longitudinal disruptions (fig. 1, a2), transverse disruptions (fig. 1, b2) and parenchymal destruction (fig. 1, e2) well. Shearing lesions appeared more as longitudinal lesions (fig. 1, c2) and lesion clusters could not be differentiated from slight parenchymal destruction (fig. 1, d2). Detachments of the parenchyma from the pelvis (fig. 1, f2) and medullary lesions (fig. 1, g2) were not depicted by CT. When contrast medium was injected directly, the total extent of injury was assessed as an average of 15% greater on CT than that detectable macroscopically. In this context kidneys subjected to a lesser load were depicted by CT as more severely injured (⫾28%) than kidneys subjected to major loading, which showed an assessment accuracy of 99% (ie ⫺1%). The severity of injury caused by longitudinal lesions was over interpreted by ⫾55%. In contrast, the extent of renal damage based on parenchymal destruction, transverse lesions, shearing lesions or lesion clusters was less readily discernible (⫺37%). Central lesions were over interpreted (⫾72%), while the extent of peripheral trauma could not be depicted completely (⫺19%). Injuries penetrating the whole parenchymal depth tended to be over interpreted in organs subjected to lesser loading, whereas with high energy application the feasibility of imaging was only 67% (ie ⫺33%). These results contrasted with the outcome of kidneys pretreated with the mixture of contrast medium and gelatin. An average of only 49% of the extent of existing lesions were depicted in the images. DISCUSSION

In the current study we ascertained the validity of CT following blunt renal trauma. Using a gravitational impactor to generate reproducible parenchymal ruptures the trauma mechanism was decreased to dorsal-ventral organ compression. It is only 1 mechanism occurring in individuals sustaining renal injury. Kidney movement relative to the fixation of its vessels, organ oscillation during impact and lateral blows could not be considered in our experiment. Thus, vascular injuries were not simulated and the quantity of transverse disruption might be larger in vivo. Imaging was enabled by 2 methods, namely injection of a mixture of contrast medium and gelatin prior to organ traumatization and imitation of a perirenal tamponade. It was found that the former experimental procedure was unsuitable for answering the specified question because in the comparison of the 2 methods it only enabled imaging of half of the lesions actually present. In contrast, imaging after direct injection of contrast medium showed the total extent of blunt renal injury relatively accurately. The severity of injury was over assessed by an average of approximately 15%. In this respect an increased

COMPUTERIZED TOMOGRAPHY FOR BLUNT RENAL TRAUMA

2477

FIG. 1. Lesion types and their depiction on CT. a1 and a2, longitudinal disruption. b1 and b2, transverse disruption. c1 and c2, shearing lesion. d1 and d2, lesion cluster. e1 and e2, parenchymal destruction. f1 and f2, parenchymal detachment from pelvis. g1 and g2, medullary lesion.

degree of lesion was attributed particularly to organs subjected to lesser loading. The over interpretation of CT imaging is attributable to the evaluation of mainly central, macroscopically undetectable linear lesions because end portions of the minor calices and vessels extending centrifugally were

partially indistinguishable from longitudinal tissue dehiscence (fig. 2, a1/2 and b1/2). The extent of parenchymal destruction, transverse lesions, shearing lesions and lesion clusters was not fully depicted (fig. 3, a1/2 and b1/2). Since contrast medium present in medullary lesions and detach-

2478

COMPUTERIZED TOMOGRAPHY FOR BLUNT RENAL TRAUMA

In vivo the depiction of some injuries might be complicated by artifacts from surrounding tissue and physiological processes, such as the effects of pulsation, adjacent peristalsis and respiration, but on the other hand the evaluation is also likely to be facilitated by imaging extravasation, (subcapsular) hemorrhage or contusions that were not shown by the experimental technique used in our study.19, 20 CT followup is advisable since only 1 imaging procedure is hardly a good basis for predicting patient recovery. The outcome also depends on trauma to the vascular system and concomitant injuries to the other abdominal organs, which are revealed well by CT. CONCLUSIONS

FIG. 2. CT shows 2 examples of end portions of minor renal calices imitating peripelvic longitudinal disruption.

Through a comparison of diagnostic imaging and macroscopy it was possible to prove a high degree of accuracy with regard to CT for the assessment of renal injuries. To our knowledge we report the first investigation comparing single lesions depicted on CT images with macroscopic findings. Despite some over interpretation of lesions close to the pyelon and slightly lower feasibility in the imaging of parenchymal destruction the total extent of injury could be displayed with a high degree of accuracy. The current study demonstrates that in the future through CT data relating to renal trauma behavior can be provided in large quantities for further scientific evaluation without dissecting the organs. Moreover, diagnostic imaging does not limit a further subdivision of existing categories of renal injury, which appears increasingly appropriate with the use of more recent therapeutic options. REFERENCES

FIG. 3. Two examples of decreased depiction of parenchymal destruction.

ments of the parenchyma from the pelvis was indistinguishable from the renal pelvis content, detachments of the parenchyma from the pelvis and medullary lesions could not be assessed. It remained unaffected by the low opacification of the renal pyramids caused by the lack of concentration ability in isolated kidneys. However, this study proved high congruence between CT imaging and macroscopic findings (fig. 1). According to the literature CT has been confirmed as the imaging procedure of choice in cases of blunt renal trauma.1, 2 CT staging is described to be exact16 –18 but findings are usually assigned only to current classification systems with not more than 5 grades differentiated.1, 9 –12 It was reported that on CT more grades of blunt renal trauma can be distinguished than are used in existing classification systems.3 Thus, a further subdivision of categorization systems regarding blunt renal trauma is applicable in clinical practice. It was also demonstrated that the evaluation of diagnostic imaging can replace more labor-intensive and timeconsuming macroscopic analysis in cases of biomechanic investigations. Within a short period CT could provide large quantities of data relating to renal injury patterns. It is of great significance for the further development of computer assisted simulation of renal trauma.13, 14

1. Kawashima, A., Sandler, C. M., Corl, F. M., West, O. C., Tamm, E. P., Fishman, E. K. et al: Imaging of renal trauma: a comprehensive review. Radiographics, 21: 557, 2001 2. Herschorn, S., Radomski, S. B., Shoskes, D. A., Mahoney, J., Hirshberg, E. and Klotz, L.: Evaluation and treatment of blunt renal trauma. J Urol, 146: 274, 1991 3. Lent, V.: What classification is appropriate in renal trauma? Eur Urol, 30: 327, 1996 4. Baumann, L., Greenfield, S. P., Aker, J., Brody, A., Karp, M., Allen, J. et al: Nonoperative management of major blunt renal trauma in children: in-hospital morbidity and long-term followup. J Urol, 148: 691, 1992 5. Russell, R. S., Gomelsky, A., McMahon, D. R., Andrews, D. and Nasrallah, P. F.: Management of grade IV renal injury in children. J Urol, 166: 1049, 2001 6. Altman, A. L., Haas, C., Dinchman, K. H. and Spirnak, J. P.: Selective nonoperative management of blunt grade 5 renal injury. J Urol, 164: 27, 2000 7. Matthews, L. A., Smith, E. M. and Spirnak, J. P.: Nonoperative treatment of major blunt renal lacerations with urinary extravasation. J Urol, 157: 2056, 1997 8. Smith, E. M., Elder, J. S. and Spirnak, J. P.: Major blunt renal trauma in the pediatric population: is a nonoperative approach indicated? J Urol, 149: 546, 1993 9. Moore, E. E., Shackford, S. R., Pachter, H. L., McAninch, J. W., Browner, B. D., Champion, H. R. et al: Organ injury scaling: spleen, liver, and kidney. J Trauma, 29: 1664, 1989 10. Carpio, F. and Morey, A. F.: Radiographic staging of renal injuries. World J Urol, 17: 66, 1999 11. Becker, C. D., Mentha, G., Schmidlin, F. and Terrier, F.: Blunt abdominal trauma in adults: role of CT in the diagnosis and management of visceral injuries. Part 2: Gastrointestinal tract and retroperitoneal organs. Eur Radiol, 8: 772, 1998 12. Kawashima, A., Sandler, C. M., Corl, F. M., West, O. C., Tamm, E. P., Fishman, E. K. et al: Imaging evaluation of posttraumatic renal injuries. Abdom Imaging, 27: 199, 2002 13. Schmidlin, F. R., Schmid, P., Kurtyka, T., Iselin, C. E. and Graber, P.: Force transmission and stress distribution in a computer-simulated model of the kidney: an analysis of the injury mechanisms in renal trauma. J Trauma, 40: 791, 1996 14. Schmidlin, F., Farshad, M., Bidaut, L., Barbezat, M., Becker, C., Niederer, P. et al: Biomechanical analysis and clinical treat-

COMPUTERIZED TOMOGRAPHY FOR BLUNT RENAL TRAUMA ment of blunt renal trauma. Swiss Surg, 5: 237, 1998 15. Bschleipfer, T., Kallieris, D., Hauck, E. W., Weidner, W. and Pust, R. A.: Blunt renal trauma: biomechanics and origination of renal lesions. Eur Urol, 42: 614, 2002 16. Vasile, M., Bellin, M. F., Helenon, O., Mourey, I. and Cluzel, P.: Imaging evaluation of renal trauma. Abdom Imaging, 25: 424, 2000 17. Santucci, R. A. and McAninch, J. M.: Grade IV renal injuries: evaluation, treatment, and outcome. World J Surg, 25: 1565, 2001 18. Bretan, P. N., Jr., McAninch, J. W., Federle, M. P. and Jeffrey,

2479

R. B., Jr.: Computerized tomographic staging of renal trauma: 85 consecutive cases. J Urol, 136: 561, 1986 19. Tong, Y.-C., Chun, J.-S., Tsai, H.-M., Yu, C.-Y. and Lin, J. S. N.: Use of hematoma size on computerized tomography and calculated average bleeding rate as indications for immediate surgical intervention in blunt renal trauma. J Urol, 147: 984, 1992 20. Harris, A. C., Zwirewich, C. V., Lyburn, I. D., Torreggiani, W. C. and Marchinkow, L. O.: CT findings in blunt renal trauma. Radiographics, 21: S201, 2001