Location of Resuscitative Endovascular Balloon Occlusion of the Aorta (REBOA) in Zone III Using External Landmark Without Fluoroscopy: A CTA-based Retrospective Study

Resuscitative endovascular balloon occlusion of the aorta (REBOA) can timely prevent the wounded from massive hemorrhage. We aim to study whether the combination of the xiphoid process and the umbilicus could guide the placement of REBOA in zone III without uoroscopy. We conducted a retrospective study in a university hospital that included 57 subjects who underwent contrast-enhanced computed tomography (CT) from April to December in 2019. External distances and intravascular lengths were measured by three-dimensional reconstruction CT images on the workstation, including the distances from the femoral artery to the xiphoid process (FA-Xi), the midpoint between the xiphoid process and the umbilicus (FA-mXU), the umbilicus (FA-Ui), the midpoint of the zone III (FA-mZIII), the lowest renal artery (FA-LRA), and aortic bifurcation (FA-AB). The relationship between external landmarks and REBOA catheter positioning in zone III was studied, involving the quartering distances between the xiphoid process and the umbilicus, the distance below the xiphoid process and that above the umbilicus. The predicted accuracy and safety margin of the balloon (distal and proximal) were calculated by curvature plane reconstruction. The probability of balloon positioning in zone III using these three methods was compared by Chi square test.


Abstract Background
Resuscitative endovascular balloon occlusion of the aorta (REBOA) can timely prevent the wounded from massive hemorrhage. We aim to study whether the combination of the xiphoid process and the umbilicus could guide the placement of REBOA in zone III without uoroscopy.

Methods
We conducted a retrospective study in a university hospital that included 57 subjects who underwent contrast-enhanced computed tomography (CT) from April to December in 2019. External distances and intravascular lengths were measured by three-dimensional reconstruction CT images on the workstation, including the distances from the femoral artery to the xiphoid process (FA-Xi), the midpoint between the xiphoid process and the umbilicus (FA-mXU), the umbilicus (FA-Ui), the midpoint of the zone III (FA-mZIII), the lowest renal artery (FA-LRA), and aortic bifurcation (FA-AB). The relationship between external landmarks and REBOA catheter positioning in zone III was studied, involving the quartering distances between the xiphoid process and the umbilicus, the distance below the xiphoid process and that above the umbilicus. The predicted accuracy and safety margin of the balloon (distal and proximal) were calculated by curvature plane reconstruction. The probability of balloon positioning in zone III using these three methods was compared by Chi square test.

Results
The average length of aortic zone III was 9.4 cm (SD = 10.0), and that of FA-mZIII on the right and left sides were 24.4 cm (SD = 2.1) ,23.8 cm (SD = 2.1), respectively. FA-Xi was signi cantly longer than FA-LRA, and FA-Ui was signi cantly shorter than FA-AB (P < 0.05). Using the quartering distances between the xiphoid and the umbilicus, the distance below the xiphoid, the distance above the umbilicus to predict the length of REBOA catheter positioning in zone III showed no statistically signi cant difference. Using FA-mXU to predict the accuracy of catheter positioning in zone III on the left and right sides were 84.2% and 86%. Although there was a little difference between the left side of FA-mZIII and FA-mXU, there were no statistical differences on the right side.

Conclusions
The midpoint between the xiphoid process and the umbilicus is a good external landmark to guide the placement of REBOA in zone III without uoroscopy.
Background Page 3/18 Massive hemorrhage remains an important cause of death for the injured worldwide. Despite many endeavors, the survival rate remains bleak, and 85% of hemorrhagic deaths arise within the rst six hours [1,2]. Rapid and effective hemostasis is a crucial strategy to rescue critically ill patients in prehospital settings. Resuscitative endovascular balloon occlusion of the aorta (REBOA) is a minimally invasive and convenient treatment for non-compressible torso hemorrhage. It can be used by trained physicians to achieve rapid hemostasis, increase arterial pressure, and obtain opportunities for further de nitive treatment. However, a series of factors including limited imaging equipment, transportation time and onsite emergency, make it di cult to spread its use in the pre-hospital environment [3,4].
Zone I is described as the region from the left subclavian artery to the celiac trunk, and zone III is described as the region from the lowest renal artery to the aortic bifurcation [5]. It has been suggested that the length of the REBOA catheter positioning in zone I and zone III could be predicted by torso height.
Comparing with zone I, zone III has a poor accuracy [6]. The xed balloon catheter de nes zone III as the common segment (236-256 mm) from the superior margin of symphysis pubis, which has been proposed as an alteration to increase the possibility of the placement of REBOA in zone III [7]. Despite the prediction accuracy of zone III could be optimized by this model, Meyer et al. have further veri ed that the predicted catheter lengths were located partially or completely outside the boundary of the zone III in 50% of subjects [8]. Thus, in the pre-hospital emergency, a quick and convenient localization method is urgently required to guide the balloon positioning in zone III.
Others have proposed that REBOA can be accomplished by external landmarks without uoroscopy.
Inaba et al. found that the midpoint of the sternum could be used to guide the correct placement of REBOA in zone I via the study of cadavers [9]. Then, Okada et al showed that the safe catheter length positioning in zone I should be between the distance from the femoral artery to suprasternal notch and that to the xiphoid process [10]. In comparison, the current literatures provide less guidance on good practice in regard to the placement of REBOA in zone III by external landmark. Eliason et al placed the balloon to the projection position 5.5 cm above the umbilicus with marked improvement in predicted accuracy [6]. Yet the best appropriate distance above the umbilicus is still unknown. As the position of xiphoid process is xed and that of umbilicus is variable due to contour of body, we hypothesize that the combination of the xiphoid process and the umbilicus could guide the placement of REBOA in zone III.

Methods
This study is a retrospective study, which was certi ed by the Ethics Committee of the Second A liated Hospital, Zhejiang University College of Medicine, and conducted in the radiology department. Data of all patients following no apparent vascular tortuosity by computed tomography (CT) diagnosis aged 18-70 years with contrast-enhanced CT images of chest, abdomen, and pelvis from April to December in 2019 were analyzed. Subjects whose images accompanied by vascular malformation, poor contrast enhancement, and history of aortic surgery were precluded.
Before CT scanning, 70 ml contrast agent iohexol (Omnipaque 370 mg I/ml, GE Healthcare, Shanghai, China) and 60 ml 0.9% saline were injected into the middle elbow vein by a dual-head power injector (Mallinckrodt, American) at a ow rate of 4 ml/s [11]. A 128-slice dual-energy CT scanner (SOMATOM De nition Flash, Siemens Medical Solutions, Germany) was utilized for the scanning during the arterial phase. The image thickness and interlayer distances are 0.75 mm and 0.5 mm, respectively. The range of the scan area is from subclavian to bilateral femoral artery. The images were saved to the CT database, then uploaded to a CT workstation (Syngo Via VB10B, Siemens, Germany) following screened by inclusion and exclusion criteria. Effective image reconstruction methods for blood vessels and external landmarks include multiplanar reconstruction, curvature plane reconstruction, and volume rendering.
The external landmarks were measured by volume-rendering reconstruction, referring to the distance from bilateral femoral artery to the xiphoid process (FA-Xi), the midpoint between the xiphoid process and the umbilicus (FA-mXU) and the umbilicus (FA-Ui), the distance from the xiphoid process to the umbilicus. The intravascular regions were measured along the centerline of the blood vessels by multiplanar reconstruction and curvature plane reconstruction, including the lengths from bilateral femoral artery to the left subclavian artery, celiac trunk, lowest renal artery, aortic bifurcation ( Fig. 1). Data recording was performed using a standard data record form by a trained clinician and examined for proofreading errors by an experienced imaging expert.

The quartering distances between xiphoid process and umbilicus
The catheter length in this study was considered as the distance from the femoral artery to the midpoint of the balloon. The balloon length was considered to be 4 cm. The distance between xiphoid process and umbilicus was evenly divided into four parts. The distance of 1/4, 1/2, 3/4 was separately added to the distance from the bilateral femoral artery to the umbilicus as the predicted catheter length. The distal safety margin of the balloon was de ned as the distance between the superior edge of the balloon and the LRA, and the proximal safety margin of the balloon was de ned as the distance between the inferior edge of the balloon and the AB. The distance of the distal safety margin was FA-LRA minus catheter length minus 2 cm, the distance of the proximal safety margin of the balloon was catheter length minus FA-AB minus 2 cm (Fig. 2). (Bilateral balloon safety margin must be greater than or equal to 0). Simultaneously, the prediction accuracy was calculated.
The distance below the xiphoid The distance from the bilateral femoral artery to the xiphoid was measured on the workstation, then separately minus 1, 2, 3, 4, 5-12cm as the predicted catheter length. Distal and proximal balloon safety margin distances, as well as prediction accuracy, were calculated by using this catheter length.
The distance above the umbilicus The distance from the bilateral femoral artery to the umbilicus was measured on the workstation, then added 1, 2, 3, 4, 5-12cm in sequence as the predicted catheter length. Similarly, distal and proximal balloon safety margin distances, as well as prediction accuracy, were calculated with this catheter length.

Statistical Analysis
The measured distances and lengths were described as mean (standard deviation SD). Age, sex, height, and BMI were used to describe the individual characteristics. Data were inserted into an Excel database (Excel 2019, Microsoft), and statistical analysis was performed by using SPSS statistics 20.0 (SPSS Inc., Chicago, IL, USA). Tests comparing bilateral data differences were conducted by using paired-sample ttests if Shapiro-Wilk (S-W) test indicates that the data were normally distributed. The difference between the three predicted methods was compared by the Chi-Square test. Differences were declared to be signi cant when p-values were less than 0.05 (P < 0.05).

Result
The contrast-enhanced CT scans of the chest, abdominal, and pelvic were available for analysis on 77 subjects from April to December in 2019. We identi ed 70 subjects who ful lled the inclusion criteria. Of these, 7 subjects were diagnosed with aortic dissection or aortic aneurysm and 6 subjects lost general information. Ultimately, 57 subjects were enrolled in the study (Fig. 3).
The individual characteristics of the subjects were described in Table 1. The mean age of all patients was 60 years (SD = 9.4), mean height was 164 cm (SD = 8.6), and mean BMI was 23.2 kg/m 2 (SD = 3.0), 70% of the subjects were male (n = 40). The length of zone I was 24.7 cm (SD = 2.2) which is the longest region in the aorta, while zone III had only a length of 9.4 cm (SD = 1.0). The FA-AB on the right side was 5 mm longer than that of the left side. Similarly, the FA-LRA on the right side was 5 mm longer than that of the left side. In general, the right length of REBOA catheter positioning in zone III was slightly longer compared with the left side. The average length from the femoral artery to the midpoint of zone III on the right and left sides were 24.4 cm (SD = 2.1), 23.8 cm (SD = 2.1), respectively ( Table 2). The box plots can be used to describe the distribution of lengths ranging from the bilateral femoral artery to the key anatomical structure (Fig. 4a).    Safe distance was described as mean± standard deviation. The measured catheter length was de ned as the distance from the femoral artery to the midpoint of the balloon. Distal safe distance was the distance from the distal edge of the balloon to the lowest renal artery. Proximal safe distance was the distance from the proximal edge of the balloon to the aorticbifurcation. The right distance above the umbilicus is 9cm and the left distance above the umbilicus is 7cm. The bilateral distance below the xiphoid is 9.5cm. * Chi-Square test, P > 0.05,no statistically signi cant.
The distance from the bilateral femoral artery to the external landmarks were measured ( Table 3) . It was evident that the FA-Xi was longer than the FA-LRA and the FA-Ui was shorter than the FA-AB (paired two-tailed test, P < 0.05) (Fig. 4a). Although there was a little difference between the left side of FA-mZIII and FA-mXU (paired two-tailed test, P = 0.04) (Fig. 4b), there were no statistical differences on the right side (paired two-tailed test, P > 0.05) (Fig. 4b).
The 2 equal distances of xiphoid and umbilicus (the midpoint between the xiphoid process and the umbilicus) were added to the bilateral distances from the femoral artery to the umbilicus with acceptable predicted accuracy (right: 86.0%, left:84.2%). The average of distal balloon safety margin distances with standard deviation on the right and left sides were 3.0 cm (SD = 1.7) and 2.5 cm (SD = 1.7), respectively.
The proximal balloon safety margin distances on the right and left sides were 2.4 cm (SD = 1.7) and 3.1 cm (SD = 1.6). The distance from the femoral artery to 9 cm above the umbilicus on the right side was used to predict the length of REBOA catheter positioning in zone III with the highest accuracy (89.5%). The distal and proximal balloon safety margin distance on the right side were 2.7 cm (SD = 1.8) and 2.8 cm (SD = 1.6), respectively. In contrast, the distance from the femoral artery to 7 cm above the umbilicus on the left side reached an accuracy of 87.7%. The distal and proximal balloon safety margin distance on the left side were 3.4 cm (SD = 1.9) and 2.1 cm (SD = 1.5), respectively. The predicted accuracy of the distance from the femoral artery to 9.5 cm below the xiphoid on right and left sides were 82.5% and 80.7%, respectively. The distal balloon safety margin distances on the right and left sides were 3.0 cm (SD = 1.9) and 2.7 cm (SD = 1.8), respectively. The proximal balloon safety margin distances on the right and left sides were 2.4 cm (SD = 1.9) and 2.8 cm (SD = 1.9), respectively. The bilateral predicted accuracy of catheter length positioning in zone III were no signi cant difference in statistical analysis using the above three predicted methods (Chi-Square test, p > 0.05) ( Table. 4) (Fig. 5).

Discussion
In our study, the xiphoid process and the umbilicus has been demonstrated to be involved in REBOA positioning in zone III with acceptable predicted accuracy (right: 86.0%, left:84.2%). FA-Xi was longer than FA-LRA and FA-Ui was shorter than the FA-AB. Hence, the length of REBOA catheter positioning in zone III should range from FA-Xi to FA-Ui. The combination of xiphoid process and umbilicus can be used to guide the placement of REBOA in zone III without uoroscopy.
At present, studies have shown that severe pelvic fractures are regularly accompanied by arterial injury, and the mortality rate can achieve 32% when hemorrhagic shock occurs [12]. Timely and effective treatment of vascular hemorrhage is critical to prevent death from hemorrhagic shock. Clinically, it is possible to intervene early through pelvic packing, and then be treated with interventional embolization to further control catastrophic hemorrhage [13]. However, it is almost unrealistic to achieve complete hemostasis by means of vascular embolization in the pre-hospital environment. REBOA can achieve rapid hemostasis in the pre-hospital settings to maintain central arterial pressure and stabilize vital signs. But due to the limited transport times and pre-hospital environment, the placement of REBOA in zone III is still not ideal at present.
Although it has been proposed the estimation of the catheter length positioning in zone III by torso height, the result was just passable [6]. Wessels et al tried to apply radiofrequency identi cation to predict the location of balloon in zone III on the cadaver with an accuracy of only 16% [14]. It is also recommended to assemble various individual information (gender, smoking, diabetes mellitus, etc.) to estimate catheter length through multi-factor correlation analysis [15]. Considering the integrity of information and the accuracy of calculation, it may hamper its wide application in the pre-hospital environment. Recently, New technologies such as Gamma probe and the magnetically trackable device have been applied into the balloon placement of the REBOA without uoroscopy, yet these studies are still in the animal experimental stage [16,17]. Thus, it is unclear whether these ndings are applicable to human being. The positioning of the balloon guided by anatomical structures has attracted much attention. Measuring the distance between external landmarks can directly determine the inserted catheter length for the sake of balloon positioning, which is a practical approach.
Some researchers have struggled to utilize anatomical structures or other strategies to guide the right placement of REBOA. Martinelli et al. proposed that the catheter was inserted 50 cm toward the cephalic, lled with the saline, withdrawn until it was wedged in the aortic bifurcation, and then moved 5 cm to the cephalic [18]. However, it should be noted that the withdrawal process may provoke vascular endothelial damage and the relevant imaging equipment is required for real-time positioning. Recently, It has been proposed to estimate the balloon length by torso height. The accuracy of balloon placement in the target region is appreciable-the accuracy of zone I and zone III are 100%,74%, respectively [6]. We have to admit, however, that the accuracy of the catheter length positioning in zone III is not very satisfactory. In our study, the catheter length positioning in zone III can be quickly estimated according to the external landmarks. Of these, we found that the aortic length was shorter in those patients with short stature. If the xed catheters continued to be applied in zone III at this time, it may cause long-term ischemic necrosis of vital organs. Inaba and Okada proposed that sternum was likely to be used to guide the catheter length positioning in zone I [9,10], yet there is no further study on the relevant anatomical structure of zone III.
Through analysis of the CTA measurement data, our study further veri ed that the accuracy of the catheter length positioning in zone I was up to be 100% by the midpoint of the sternum. Besides, we utilized three measurement methods to compare the difference of predicted accuracy in zone III and found that the midpoint of the xiphoid process and the umbilicus, 7 cm or 9 cm above the umbilicus, and 9.5 cm below the xiphoid process were proven no signi cant difference. Considering the characteristics of simple operation and quick location in the pre-hospital environment, the midpoint of the xiphoid process and the umbilicus is the preferred method to predict the appropriate catheter length. The optimal distances above umbilicus on the left and right sides are not the same, which may be explained by the difference in bilateral anatomical structures. Interestingly, the optimal distance below the xiphoid process and above the umbilicus is close to half of the distance between the xiphoid process and the umbilicus (17.9 cm (SD = 2.1)). Thus, we have reason to believe that the midpoint of the umbilicus and xiphoid process may be a useful external landmark for the rapid judgment of catheter length positioning in zone III.
Rapid determination of the catheter length positioning in zone III is signi cantly imperative for the prehospital patients who suffered from catastrophic hemorrhage due to the lower limb and pelvis. Although the position of the balloon can be con rmed by X-ray and other imaging examinations after REBOA balloon being lled, it is impractical to achieve this strategy in the pre-hospital environment. For patients with non-compressible torso hemorrhage, the long-time malposition of balloon not only fail to achieve effective hemostasis in the target region, but also largely cause balloon-related complications. As the White Book states, the maximum transportation time for a major trauma patient from the scene should not exceed 30 minutes [19]. We found that the right FA-mZIII and FA-mXU were shown no signi cant difference in statistical analysis, and FA-mXU could be used to guide the inserted catheter length positioning in zone III. It is well known that the length of zone III is about 8-9 cm without a large faulttolerance range as zone I. "Too long" and "too short" are likely to lead to disastrous consequences. "Too long" means that the inferior edge of the catheter balloon is easy to cross the lowest renal artery and block the blood supply of vital organs, which immensely reduces the blocking time of REBOA and drives visceral ischemic necrosis and reperfusion injury in severe cases. "Too short" means that the superior edge of the catheter balloon fails to exceed the aortic bifurcation, which can easily lead to vascular dissection or even amputation, thus fail to take full advantage of REBOA in the rapid control of catastrophic hemorrhage. As a result, the appropriate catheter length positioning in zone III is essential to rapid hemostasis in a critical situation.
Finally, several limitations to this study should be considered. First, these conclusions were based on the responses of CT three-dimensional reconstruction and might not vividly re ect measurement processes on the intact body. The positioning of body surface markers (such as umbilicus) may be affected by body shape and body position.
For special groups, such as obesity and pregnant patients, some extra imaging methods must be equipped to avoid possible injuries and con rm precise positioning before in ation. Second, the catheter length estimated by the external landmark may be different from the actual length running in the vascular. The actual catheter length may also be affected by the blood ow velocity and the degree of vascular curvature. Third, our subjects were Chinese population, so it is not known whether our results apply to other groups. Larger studies with longer follow-up are needed to study whether exists other factors affecting the prediction of catheter length in zone III.

Conclusion
The midpoint of the xiphoid process and the umbilicus is a good external landmark for predicting the balloon position in zone III. We are convinced that it could be a relatively reliable and convenient way to guide the right placement of REBOA in zone III without uoroscopy in the pre-hospital environment. Multiplanar reconstruction, curvature plane reconstruction and volume-renderingimages.(a)Multiplanar reconstruction and curvature plane reconstruction were used to measure the length from the bilateral femoral artery (FA) to the lowest renal artery (LRA) and the aortic bifurcation (AB). (b)Volume-rendering was used to measure the distance from the bilateral femoral artery (FA) to the xiphoid process (Xi), the midpoint of the distance from xiphoid process to umbilicus (mXU) and the umbilicus (Ui).(c d). Threedimensional vascular imaging technology was used to depict the front and side images of the total aorta.