The utility of intravascular ultrasound compared to angiography in the diagnosis of blunt traumatic aortic injury

Transcription

1 From the Society for Vascular Surgery The utility of intravascular ultrasound compared to angiography in the diagnosis of blunt traumatic aortic injury Ali Azizzadeh, MD, a Jaime Valdes, MD, a Charles C. Miller III, PhD, b Louis L. Nguyen, MD, MBA, MPH, c Anthony L. Estrera, MD, a Kristofer Charlton-Ouw, MD, a Sheila M. Coogan, MD, a John B. Holcomb, MD, a and Hazim J. Safi, MD, a Houston and El Paso, Tex; and Boston, Mass Background: Blunt traumatic aortic injury (TAI) refers to a spectrum of pathology that ranges from intimal tears to aortic rupture. Computed tomography angiography (CTA) has been widely used as a diagnostic tool in this setting. Additional imaging is required when CTA studies are equivocal. The purpose of this study is to evaluate the utility of intravascular ultrasound (IVUS) versus angiography in the diagnosis of TAI. Methods: We performed an analysis of prospectively collected trauma registry data. CTA was used as the initial screening test. Patients with a positive or equivocal CTA underwent angiography and IVUS. Injuries were classified into Grades 1 to 4 (intimal tear, intramural hematoma, pseudoaneurysm, and rupture). Patients with Grade 1 injuries were managed medically. Patients with Grade 2 to 4 injuries underwent repair. A blinded randomized retrospective review of positive and equivocal imaging studies was performed. Standard screening test assessments (sensitivity, specificity), inter-rater agreement (Kappa), and frequency (Chi-square for the difference) were computed to evaluate the measurement characteristics of the multiple imaging techniques. Results: Between May 2008 and August 2009, 7961 patients were admitted to our trauma center, and 2153 (27%) underwent a chest CTA. Twenty-five (0.3%) patients (21 males, mean age 21.9 years) had a positive or equivocal study for TAI. The mean Injury Severity Score was Ten patients underwent repair (nine endovascular, one open), and 15 patients were managed medically. The 30-day mortality, paraplegia, and stroke rates were zero. Equivocal results were more common with CTA images than with either IVUS or angiography (27% vs 2.5 and 5%, respectively; overall P.0002). Compared with angiography, IVUS changed the diagnosis in 13% of cases; identifying injuries in 11% and ruling them out in 2%. Sensitivity and specificity of angiography with respect to IVUS was 38% and 89%, respectively. Conclusions: CTA is useful as a screening test in suspected TAI. When additional imaging is required after an equivocal CTA, IVUS is better than angiography. Therefore, we advocate the use of IVUS in potential TAI patients in whom angiography is being considered. (J Vasc Surg 2011;53: ) Blunt traumatic aortic injury (TAI) is the second most common cause of death after blunt trauma. 1,2 The mechanism is likely related to a complex combination of both relative motion of the structures within the thorax and local loading of the tissues, either as a result of the anatomy or the nature of the impact. 3 In the aorta, the greatest strain occurs at the isthmus. 4 A 1958 article by Parmley reported an 85% pre-hospital mortality in patients with TAI. 5 Patients who survive transportation to the hospital may present with a spectrum of pathologies that include intimal tear, intramural hematoma, pseudoaneurysm, and rupture. From the University of Texas Medical School Houston, Memorial Hermann Heart and Vascular Institute; a the Texas Tech University Health Sciences Center, Paul L. Foster School of Medicine; b and the Harvard Medical School, Brigham and Women s Hospital. c Competition of interest: none. Presented at the 2010 Vascular Annual Meeting, Society for Vascular Surgery, Boston, Mass, June 10-13, Reprint requests: Ali Azizzadeh, MD, Associate Professor, 6400 Fannin, Ste. 2850, Houston, TX ( Ali.Azizzadeh@uth.tmc.edu). The editors and reviewers of this article have no relevant financial relationships to disclose per the JVS policy that requires reviewers to decline review of any manuscript for which they may have a competition of interest /$36.00 Copyright 2011 by the Society for Vascular Surgery. doi: /j.jvs Computed tomography angiography (CTA) has been widely used as a diagnostic tool in this setting. 6 Although CTA provides great sensitivity for diagnosis of TAI, there are a significant number of equivocal scans that require additional imaging. 7 The purpose of this study is to evaluate the utility of intravascular ultrasound (IVUS) versus angiography in the diagnosis of TAI after a positive or equivocal CTA. METHODS We performed an analysis of prospectively collected trauma registry data at an urban Level I center. This study was approved by the Committee for the Protection of Human Subjects, which acts as the institutional review board. TAI patients were screened with CTA on arrival. Initial management included resuscitation, blood pressure control, and treatment of associated injuries. Patients with a CTA read as positive or equivocal in the report of the attending radiologist underwent both angiography and IVUS after stabilization. Therefore, the data analyzed arise only from cases where work-up beyond CTA was clinically indicated. The population of patients with negative CTA did not receive further studies and is not included in the data analysis. Angiography was performed in a hybrid

2 JOURNAL OF VASCULAR SURGERY Volume 53, Number 3 Azizzadeh et al 609 Fig 1. Classification of traumatic aortic injury. endovascular operating room using a fixed Siemens Axiom Artus (Munich, Germany) system. Intravascular ultrasound was performed concurrently using a Volcano s5 imaging system and a Visions PV 8.2 French Catheter (San Diego, Calif). Based on the combination of imaging modalities, injuries were classified into Grade 1, intimal tear; Grade 2, intramural hematoma; Grade 3, aortic pseudoaneurysm; and Grade 4, free rupture (Fig 1). 8 Patients with Grade 1 injuries were managed medically with a follow-up CTA study at 6 weeks. Patients with Grade 2 to 4 injuries underwent endovascular repair with the off-label use of a Food and Drug Administration (FDA) approved thoracic device. The suitability of a patient for endovascular repair was based on aortic diameter according to the manufacturer s sizing recommendations for thoracic devices as well as the location of the injury. Our technique for both open and thoracic endovascular repair of TAI has been described previously. 8 The primary outcome measures were mortality, stroke, and paraplegia. A retrospective review of positive and equivocal imaging studies (CTA, angiography, and IVUS) was performed by a two-member panel consisting of a vascular surgeon and a cardiothoracic surgeon, with the review conducted in a blinded fashion, and the order in which studies were presented was randomized. Uninjured control studies were also introduced into the case review as described below. Each reviewer was initially presented with the CTA images of a patient. Following that, either the angiography or the IVUS studies of the same patient were presented. The order of the patients, as well as the order of the secondary and tertiary studies (IVUS vs angiography) was randomized. Results were recorded as positive, negative, or equivocal by each reviewer. Reviewers were instructed to respond positive when an injury was present, negative when no injury was identified, and equivocal when an injury could not be ruled in or ruled out. Equivocality of ratings was therefore a clinical judgment about the interpretation of an image and not a statistically-derived determination. Images from 10 patients without traumatic aortic injury were used as control. Since IVUS is the only measure of the three able to image all layers of the aorta simultaneously, it was used as the standard against which the other two tests were judged for purposes of estimating sensitivity and specificity. Agreement by the expert raters on the diagnostic imaging tests was assessed two ways; by Kappa statistic as a measure of inter-rater agreement, and by frequency test, to determine which tests most often caused raters to change

3 610 Azizzadeh et al JOURNAL OF VASCULAR SURGERY March 2011 Fig 2. Treatment algorithm describing the study patients. CTA, Computed tomography angiography; TEVAR, thoracic endovascular aortic repair. their ratings. The effect of order in which the tests were presented to the raters was assessed by Kappa as well. All computations were performed using SAS software version (SAS Institute, Inc, Cary, NC). The null hypothesis was rejected at a nominal P.05. RESULTS Between May 2008 and August 2009, 7961 patients were admitted to our Level I Trauma Center, and 2153 (27%) underwent a chest CTA. Of the 2153 CTA exams performed, 2128 were deemed negative by the attending radiologist. Positive or equivocal results were obtained in 25 (0.31%) patients (21 males, mean age 21.9 years). The mean Injury Severity Score was The official radiology report for the 25 studies revealed: TAI (n 14), suspected TAI or TAI cannot be excluded (n 6), and periaortic mediastinal hematoma (n 5). Following stabilization, all 25 patients underwent angiography and IVUS. Aortic injury was ruled out in 10 patients with angiography and IVUS. No further intervention was done in these patients. Five patients with Grade 1 injuries were managed medically using anti-impulse therapy. This consisted of short-acting intravenous -blocker administration to maintain a systolic blood pressure of 120 mm Hg and heart rate 90 beats per min. Patients were weaned to oral therapy when possible. A follow-up CT scan at 6 weeks confirmed healing in all patients. Ten patients with Grade 2 to 4 injuries underwent repair (nine endovascular, one open). The detailed treatment algorithm is described in Fig 2. The 30-day mortality, paraplegia, and stroke rates were zero. There was no mortality from aortic rupture or any other cause in this cohort after admission to the hospital. One patient underwent open repair after failure of medical management. This was a 20-year-old man after a motor vehicle accident who had a CTA that demonstrated an intramural hematoma (Grade 2 TAI). He underwent angiography and IVUS per protocol. The angiography was negative. The IVUS confirmed the Grade 2 aortic injury. He did not meet sizing criteria for the only FDA-approved thoracic device at the time, so he was followed with serial imaging. His follow-up CTA at 1 week showed an enlargement of the injury. Our attempts to obtain an investigational device exemption for endovascular repair using a smaller diameter device were unsuccessful. He subsequently underwent open repair without complication. Agreement between raters under various conditions as assessed by Kappa is shown in the Table. Agreement between raters about whether studies were positive, negative, or equivocal was good (Kappa 0.77) when the CTA was considered by itself. Addition of IVUS to this determination did not affect rater agreement on diagnostic findings (CTA IVUS also 0.77). Addition of angiography readings reduced rater concordance regardless of how the other tests were presented in combination with angiography. Simply stated, raters were more likely to agree about the interpretation of IVUS findings than angiographic findings, so that addition of angiography tended to introduce more disagreement between raters about the diagnosis than CTA or IVUS. Comparing the findings of the imaging studies, expert readings of studies as equivocal were more common with CTA images than with either IVUS or

4 JOURNAL OF VASCULAR SURGERY Volume 53, Number 3 Azizzadeh et al 611 Table. Agreement among raters assessed by Kappa statistic Test (order) Kappa Overall 0.68 CTA alone 0.77 CTA (angio) 0.57 CTA (angio IVUS) 0.61 CTA (IVUS) 0.77 CTA (IVUS angio) 0.44 CTA, Computed tomography angiography; IVUS, intravascular ultrasound. CTA and IVUS show good agreement. Angiography reduces agreement markedly, presumably by introducing ambiguity to an otherwise generally agreed-upon finding. Notably, CTA was the most-often equivocal test, but the raters agreed that it was equivocal. Fig 3. Percentage of equivocal results with computed tomography angiography (CTA), angiography (Angio), and intravascular ultrasound (IVUS). angiography (27% vs 2.5 and 5%, respectively; overall P.0002; Fig 3). Compared with angiography, IVUS changed the diagnosis in 13% of cases; identifying injuries in 11% and ruling them out in 2%. Sensitivity and specificity of angiography with respect to IVUS was 38% and 89%, respectively. Fig 4 shows CTA, angiography, and IVUS studies of a patient after a motor vehicle accident. CTA shows irregular contour of the lesser curvature of the descending thoracic aorta at the isthmus consistent with a TAI. Although the angiogram was read as equivocal, IVUS clearly shows a Grade 2 TAI. DISCUSSION TAI is a major cause of mortality after blunt trauma. Those surviving hospital admission require rapid and accurate diagnosis, which is crucial to appropriate treatment. Diagnosis can be initially suspected based on the finding of an abnormal mediastinum on plain chest x-ray. 9 In a study of 656 patients with TAI, Woodring found that 93% had an abnormal mediastinum on initial chest radiograph. 10 This highlights that in up to 7% of patients with a normal mediastinum on chest x-ray, additional imaging was done based on mechanism of injury or clinical suspicion. CTA has been widely adopted as a screening tool for patients suspected of having TAI. 9 This is due to the very high sensitivity (95% to 100%) and a high negative predictive value (99% to 100%). 7,11-13 However, CTA can have a relatively low specificity (40%) and positive predictive value (15%). 7 In simple terms, CTA produces few false negative results at the expense of having a large number of false positives. Therefore, patients with an equivocal CTA study often require additional imaging including aortography, IVUS, or transesophageal echocardiography. 14 In our early experience with TAI, we made the clinical observation that some patients with a positive CTA had a negative angiogram. As a result, we started relying on IVUS as an additional imaging modality in these patients. This study was designed to mirror the standard clinical algorithm where CTA is primarily used as a screening test. Patients with negative CTA did not receive additional imaging and were not included in the analysis. The study was limited to the cases where work-up beyond CTA was clinically indicated. We compared the utility of IVUS versus angiography for the diagnosis of TAI in patients who had positive or equivocal CTA. In the present study, of the 7961 patients admitted to the emergency center, 2153 (27%) underwent a CTA study, with 14 (0.18%) positive and 11 (0.14%) equivocal results. The most common CTA finding in patients with equivocal results was a periaortic mediastinal hematoma. All patients underwent angiography and IVUS. Our comparative study found equivocal results twice as often in angiography (5%) as compared with IVUS (2.5%). The assessment of agreement between raters (Kappa) showed that the experts were more likely to agree about the interpretation of IVUS findings than angiographic findings. As a result, the addition of angiography tended to introduce more disagreement between raters about the diagnosis than CT or IVUS. It is important to keep in mind that Kappa does not measure the character or quality of the diagnosis, but only consistency among raters about the call. That is, an identical Kappa (0.77) for CT and IVUS does not mean the diagnoses were the same with both modalities; it means that the raters agreed about the diagnoses for each, even if they were different between modalities. For example, two raters might have agreed that a CT was equivocal and a follow-up IVUS was positive for injury, so that their agreement is the same, but IVUS is enhancing the CT diagnosis rather than confirming it. The IVUS is adding additional diagnostic information that the CT is unable to provide by itself. Based on these findings, IVUS is very helpful in defining TAI in patients who have equivocal CTA or angiography. IVUS, first developed in the 1960s by Born et al, provides real-time 360 images of the vessels using a min-

5 612 Azizzadeh et al JOURNAL OF VASCULAR SURGERY March 2011 Fig 4. Computed tomography angiography (CTA), angiogram, and intravascular ultrasound (IVUS) images of a patient with traumatic aortic injury. CTA shows irregular contour of the lesser curvature of the descending thoracic aorta at the isthmus consistent with a traumatic aortic injury (TAI). Although the angiogram was read as equivocal, IVUS clearly shows a Grade 2 TAI. iature sonographic probe. 15 The role of IVUS in the diagnosis of TAI has been previously reported. 14,16-19 In addition, the interpretation of IVUS has yielded excellent inter- and intraobserver agreement as an adjunct to angiography for the diagnosis of TAI. 20 We utilized a 10 MHz Visions PV 8.2 French (Volcano) catheter that incorporates a cylindrical ultrasound transducer array. The catheter is introduced over a guidewire to avoid the risk of perforation of an injured aortic wall. When combined with angiography, the IVUS catheter can be positioned in the area of injury using fluoroscopy for detailed assessment. IVUS does not require contrast or radiation and can be performed concurrently using the same femoral puncture as angiography. The disadvantages of IVUS are additional cost (capital equipment and disposable catheter), larger sheath, and operating room time. However, the information that it can provide can be invaluable in properly identifying TAI. In the present series, we used a combination of angiography and IVUS to rule out aortic injury in 10 of the 25 patients with equivocal CTAs. These patients did not undergo any further follow-up imaging or intervention. The complementary use of both modalities was helpful in ruling out TAI with a high degree of certainty. Reporting equivocal results to patients is very troublesome and frustrating. This is especially important in the setting of TAI when a missed injury can have devastating consequences. Five of the 25 patients were diagnosed with a Grade 1 aortic injury. Two of these patients had a negative angiogram. In other words, if angiography had been used as the gold standard, two Grade 1 TAIs would have been missed. The disadvantages of angiography for identification of aortic injuries have been previously described. 14,16,21 Malhotra et al reported a sensitivity of 37.5% for angiographic diagnosis of minimal aortic injury (defined as a 1-cm intimal flap with no or minimal peri-aortic hematoma) comparable to a Grade 1 injury in the present series. Remarkably, the sensitivity of angiography in the present series was also 38%. TAI lesions that do not cause an

6 JOURNAL OF VASCULAR SURGERY Volume 53, Number 3 Azizzadeh et al 613 abnormality in the contour of the aortic wall (Grade 1 and some Grade 2) are inherently difficult to see on angiography. In addition, several normal variants including ductus diverticulum, bronchial artery diverticulum, aortic spindle, and penetrating aortic ulcer can produce false-positive results on angiography. 14,21-23 IVUS, on the other hand, provides a two-dimensional axial view of the lumen and the aortic wall without the interference of intraluminal contrast, motion artifacts, or volume averaging. All five patients with Grade 1 TAI were managed medically with intravenous anti-impulse therapy. Follow-up CTA at 4 to 6 weeks demonstrated healing in all five patients. CTA was used as the only follow-up imaging modality in this group of patients due to cost constraints. The medical management of minor TAIs (Grade 1) has been previously reported. 14,16,24 Several animal studies have demonstrated spontaneous healing of arterial injuries that are limited to the intima and the internal elastic lamina (inner media) The current data support non-operative treatment of selected patients with Grade 1 TAI who are compliant with medical therapy and follow-up imaging. The remaining 10 patients with Grades 2 to 4 TAI underwent repair. The utility of angiogram and IVUS in this group of patients who had a definitive diagnosis on CTA is limited. The additional imaging studies were usually done at the time of definitive repair. The only exception was the case of the patient with a Grade 2 TAI who failed medical therapy. If angiography had been used as the gold standard in this patient, a Grade 2 TAI would have been missed. The remainder of the patients in the study with Grades 2 to 4 TAI underwent endovascular repair with the off-label use of approved devices. Limitations of this series include the relatively small sample size and the non-randomized nature of the study. A future study design could prospectively randomize patients with equivocal CTA studies to two arms: IVUS first versus angiography first. The images of the second study would then have to be reviewed prior to performance of the third study. From a procedural standpoint, however, such a study would be somewhat cumbersome to carry out. Another limitation is the lack of pathologic diagnosis, as 24 of the 25 patients did not have an open surgical exploration. CONCLUSIONS CTA is useful as a screening test in patients suspected of having TAI. When additional imaging is required after an equivocal reading for CTA, IVUS is better than angiography. If angiography had been used as the gold standard in this study, three TAIs would have been missed. Therefore, we advocate the use of IVUS in potential TAI patients in whom angiography is being considered. The authors would like to thank G. Ken Goodrick for editing, Chris Akers for illustrations, and Edmundo Dipasupil for trauma registry. AUTHOR CONTRIBUTIONS Conception and design: AA, JV, CM, AE, KCO, SC Analysis and interpretation: AA, CM, LN Data collection: AA, JV, KCO, SC Writing the article: AA, CM Critical revision of the article: AE, LN, KCO, SC, JH, HS Final approval of the article: AA, HS Statistical analysis: CM, LN Obtained funding: Not applicable Overall responsibility: AA REFERENCES 1. Clancy TV, Gary MJ, Covington DL, Brinker CC, Blackman D. A statewide analysis of level I and II trauma centers for patients with major injuries. J Trauma 2001;51: Richens D, Field M, Neale M, Oakley C. The mechanism of injury in blunt traumatic rupture of the aorta. Eur J Cardiothorac Surg 2002;21: Pearson R, Philips N, Hancock R, Hashim S, Field M, Richens D, et al. Regional wall mechanics and blunt traumatic aortic rupture at the isthmus. Eur J Cardiothorac Surg 2008;34: Schmoker JD, Lee CH, Taylor RG, Chung A, Trombley L, Hardin N, et al. A novel model of blunt thoracic aortic injury: a mechanism confirmed? J Trauma 2008;64: Parmley LF, Mattingly TW, Manion WC, Jahnke EJ Jr. Nonpenetrating traumatic injury of the aorta. Circulation 1958;17: Steenburg SD, Ravenel JG, Ikonomidis JS, Schönholz C, Reeves S. Acute traumatic aortic injury: imaging evaluation and management. Radiology 2008;248: Bruckner BA, DiBardino DJ, Cumbie TC, Trinh C, Blackmon SH, Fisher RG, et al. Critical evaluation of chest computed tomography scans for blunt descending thoracic aortic injury. Ann Thorac Surg 2006;81: Azizzadeh A, Keyhani K, Miller CC 3rd, Coogan SM, Safi HJ, Estrera AL. Blunt traumatic aortic injury: initial experience with endovascular repair. J Vasc Surg 2009;49: O Conor CE. Diagnosing traumatic rupture of the thoracic aorta in the emergency department. Emerg Med J 2004;21: Woodring JH. The normal mediastinum in blunt traumatic rupture of the thoracic aorta and brachiocephalic arteries. J Emerg Med 1990;8: Mirvis SE, Bidwell JK, Buddemeyer EU, Diaconis JN, Pais SO, Whitley JE, et al. Value of chest radiography in excluding traumatic aortic rupture. Radiology 1987;987: Gavant ML, Menke PG, Fabian T, Flick PA, Graney MJ, Gold RE. Blunt traumatic aortic rupture: detection with helical CT of the chest. Radiology 1995;197: Wicky S, Capasso P, Meuli R, Fischer A, Segesser L, Schnyder P. Spiral aortography: an efficient technique for the diagnosis of traumatic aortic injury. Eur Radiol 1998;8: Patel NH, Hahn D, Comess KA. Blunt chest trauma victims: role of intravascular ultrasound and transesophageal echocardiography in cases of abnormal thoracic aortogram. J Trauma 2003;55: Born N, Lancee CT, Van Egmond FC. An ultrasonic intracardiac scanner. Ultrasonics 1972;10: Malhotra AK, Fabian TC, Croce MA, Weiman DS, Gavant ML, Pate JW. Minimal aortic injury: a lesion associated with advancing diagnostic techniques. J Trauma 2001;51: Marty B, Tozzi P, Ruchat P, Huber C, Doenz F, von Segesser LK. An IVUS-based approach to traumatic aortic rupture, with a look at the lesion from inside. J Endovasc Ther 2007;14: Williams DM, Simon HJ, Marx MV, Starkey TD. Acute traumatic aortic rupture: intravascular US findings. Radiology 1992;182: Uflacker R, Horn J, Phillips G, Selby JB. Intravascular sonography in the assessment of traumatic injury of the thoracic aorta. AJR Am J Roentgenol 1999;173:

7 614 Azizzadeh et al JOURNAL OF VASCULAR SURGERY March Lee DE, Arslan B, Queiroz R, Waldman DL. Assessment of inter- and intraobserver agreement between intravascular US and aortic angiography of thoracic aortic injury. Radiology 2003;227: Fisher RG, Sanchez-Torres M, Whigham CJ, Thomas JW. Lumps and bumps that mimic acute aortic and brachiocephalic vessel injury. RadioGraphics 1997;17: Morse SS, Glickman MG, Greenwood LH, Denny DF Jr, Strauss EB, Stavens BR, et al. Traumatic aortic rupture: false-positive aortographic diagnosis due to atypical ductus diverticulum. AJR Am J Roentgenol 1988;150: Goodman PC, Jeffrey RB, Minagi H, Federic MP, Thomas AN. Angiographic evaluation of the ductus diverticulum. Cardiovasc Intervent Radiol 1982;5: Wigle RL, Moran JM. Spontaneous healing of a traumatic thoracic aortic tear: case report. J Trauma 1991;31: Pederson DC, Bowyer DE. Endothelial injury and healing in vitro: studies using an organ culture system. Am J Pathol 1985;119: Chemnitz J, Christensen BC. Repair in arterial tissue 2 years after a severe single dilatation injury: the regenerative capacity of the rabbit aortic wall the importance of endothelium and the state of subendothelial connective tissue to reconstitution of the intimal barrier. Virchow Arch A Pathol Anat Histopathol 1991;418: Hsiang YN, Fragoso M, Lundkist A, Weis M. The natural history of intimal tears caused by angioscopy. Ann Vasc Surg 1992;6: Neville RF, Padberg FT Jr, DeFouw D, Hernandez J, Duran W, Hobson RW II. The arterial wall response to intimal injury in an experimental model. Ann Vasc Surg 1992;6:50-4. Submitted Jun 10, 2010; accepted Sep 25, 2010.