| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Original Research |
1 Department of Radiology, David Geffen School of Medicine at the University of
California at Los Angeles, BL-428 CHS/Box 951721, Los Angeles, CA
90095-1721.
4 Department of Urology, University of California at Los Angeles, Los Angeles,
CA.
Received June 8, 2005;
accepted after revision August 12, 2005.
Address correspondence to S. Raman
(sraman{at}mednet.ucla.edu).
Abstract
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
METHODS AND MATERIALS. Fifty-five consecutive renal donors (25 men and 30 women) underwent 16-MDCT angiography followed by donor nephrectomy. In the arterial and nephrographic phases, images were acquired with 60% overlap and 0.6-mm reconstruction in both phases after 120 mL of iohexol was injected at 4 mL/sec. On a 3D workstation, images were evaluated retrospectively by two abdominal imagers blinded to surgical results with respect to number and branching pattern of renal arteries and major and minor renal veins. These CT angiography results were compared with surgical findings.
RESULTS. The surgically confirmed sensitivity of both reviewers (1 and 2) using the MDCT data for detection of renal arteries was 98.5% (65 of 66), and accuracies were 97.0% for reviewer 1 and 95.5% for reviewer 2. Sensitivity and accuracy detection of renal veins was 97% (61 of 63) and 98% (62 of 63) for reviewer 1 and reviewer 2, respectively. Sensitivity and accuracy detection of early arterial bifurcation (< 2 cm from aorta) was 100% (14 of 14), and sensitivity in detection of late venous confluence (< 1.5 cm from aorta) was 100% (8 of 8). All major renal venous variants were identified; reviewer 1 identified 78% (18 of 23) minor venous variants, and reviewer 2 identified 83% (19 of 23) minor venous variants. There were no hemorrhagic complications at surgery. Excellent agreement between reviewers (kappa = 0.92-0.97) was achieved for detection of normal and variant anatomy.
CONCLUSION. 16-MDCT angiography enabled excellent preoperative detection of arterial anatomy and venous laparoscopic donor nephrectomy.
Keywords: angiography • CT • CT imaging • MDCT • kidney • renal transplantation
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
In preoperative planning, helical CT angiography (CTA) initially supplanted catheter angiography with reasonably good accuracy in depiction of healthy and variant renal donor anatomy, primarily in the context of open nephrectomy [10-18]. Several more-recent studies that evaluate the utility of 4-MDCT angiography report similar results with respect to arterial anatomy [19-25].
However, in laparoscopically performed donor nephrectomy, venous variants relevant to the presence and insertion of prominent lumbar and/or gonadal veins (minor variants) may result in inadvertent avulsion of these sometimes confusing vessels and lead to unanticipated hemorrhage in the laparoscopic field. Therefore, our surgeons believe that preoperative CT knowledge of major and minor venous variants helps them anticipate potentially confusing venous anastomoses and facilitates laparoscopic nephrectomy. By themselves, these venous variants generally do not preclude donation. The ability of CTA to help detect these venous variants has not been studied extensively. In prior studies using 4-MDCT angiography, sensitivity for detection of these variants varied between 75% using a standard open donor nephrectomy [21] and 92% using a laparoscopic nephrectomy procedure [22]. Recently, Sahani et al. [23] reported the sensitivity and specificity of 4-MDCT were 75% and 100%, respectively, for the identification of variant anatomy of renal veins. However, only major venous variants were evaluated.
|
|
|
This study evaluates the diagnostic performance of two experienced, blinded interpreters, using 2D and 3D data from 16-MDCT angiography, in preoperative detection of arterial and venous anatomy and variants of renal donors undergoing laparoscopic nephrectomy.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
CT Protocol
All CTA studies were performed on a 16-MDCT scanner (Somatom Sensation 16,
Siemens Medical Solutions). As part of our CT protocol, unenhanced and IV
contrast-enhanced multiphasic scans (arterial, nephrographic, and excretory
phases) were acquired. After fasting for at least 3 hr, each donor ingested
500 mL of water over 15-20 min before the scan in an attempt to improve
hydration for the excretory phase. Unenhanced CT scans were obtained with 5-mm
collimation to locate the kidneys, detect renal calculi, and use as the
baseline attenuation measurement of renal mass. After unenhanced CT scans,
100-150 mL of nonionic IV iohexol (350 mg of iodine/mL) (Omnipaque 350, GE
Healthcare) was injected into an antecubital vein through an 18-gauge
peripheral IV line using a power injector at a rate of 4.0 mL/sec. The
estimated dose-to-weight amounts were 100 mL/< 100 lb (45.4 kg) (35 g
iodine), 125 mL/100-200 lb (45.4-90.7 kg) (43.8 g iodine), 150 mL/> 200 lb
(90.7 kg) (53.5 g iodine). The arterial phase scans were initiated using an
automatic bolus-tracking program (CARE Dose, Siemens Medical Solutions). A
region of interest was placed in the abdominal aorta just above the kidneys.
Scanning was triggered 5 sec after a threshold of 150 H in the region of
interest was reached. Volumetric scans were acquired from the level of the
celiac axis to the common iliac artery bifurcation using the following
parameters: 120 kVp, 200-240 mAs, 12-mm table speed, 0.5-sec rotation speed,
0.75-mm collimation, and reconstruction with 60% overlap. Nephrographic phase
images were then acquired 85 sec after the arterial phase covering the same
area described with similar parameters. For both the arterial phase and the
nephrographic phase, the images were reconstructed at 0.75-mm thickness with a
0.6-mm interval using a standard body filter without edge enhancement.
After IV injection of 250 mL of 0.9% saline, excretory phase images were acquired 5 min after the nephrographic phase and extended from above the kidneys to the bladder base using similar parameters as described in the previous paragraph.
Image Processing and Analysis
Two experienced abdominal imagers (reviewer 1, 10 years; reviewer 2, 7
years), who had been trained for preoperative evaluation of a living
laparoscopic kidney, reviewed all images from 16-MDCT angiography images at an
independent workstation with 3D capability (Vitrea 2, Vital Images). The
reviewers had no knowledge of the dictated CT report or laparoscopic surgical
result. For each CT examination, the reviewers used axial images, supplemented
by 2D and 3D postprocessing techniques including multiplanar reformations,
maximum intensity projections, and volume rendering according to individual
preferences. The reviewers edited CT volume data sets to create 3D CTA images
in real time at frame rates of 10-30 frames/sec. Each reviewer subjectively
adjusted 3D display parameters, including width, level, opacity, and
brightness.
Renal arterial anatomy was evaluated on the arterial phase images, whereas venous anatomy was evaluated in the arterial phase, supplemented by images in the nephrographic phase, especially for assessment of accessory draining renal veins including lumbar and gonadal veins, which some-times enhanced later.
Each reviewer assessed the number and size of arteries, presence of early branching arteries, and presence of accessory or capsular arteries. Surgically relevant renal arteries were those generally greater than 4 mm in diameter. Vessels 0.5-3 mm in diameter may be sacrificed at surgery if they are deemed by the surgeon to supply a small clinically insignificant capsular or polar vascular territory when clamped (with resulting blanching of renal surface). An early branching renal artery was diagnosed when any branch diverged within 2.0 cm from the lateral wall of the aorta (left kidney) or in the retrocaval segment (right kidney). Reviewers also assessed the number of main renal veins and their anastomotic pattern.
Venous anomalies were classified as major or minor. Major anomalies were those that resulted in altered surgical management, including venous anastomoses in the recipient. Major renal venous anomalies included the supernumerary renal veins and the presence of a late venous confluence. At our institution, a late venous confluence was diagnosed when renal venous branches coalescing within 1.5 cm from the left lateral margin of the aorta (left kidney) or within 1.5 cm of the anastomosis with inferior vena cava (IVC) (right kidney). Major left-sided renal venous anomalies also included circumaortic or retroaortic renal veins and left-sided or duplicated IVC. Minor venous anomalies were either anticipated or unanticipated and recognized at laparoscopy. Minor renal venous anomalies included the presence of draining gonadal and lumbar veins (especially those 5 mm in diameter) and drainage of any associated renal tributaries into these veins (Fig. 1A, 1B and 1C). These variants did not generally involve additional postharvest anastomoses and were ligated laparoscopically.
The results from retrospective review for each reviewer were compared with the dictated surgical report and surgical assessment for the laparoscopically resected kidney. Thus only one kidney per patient was evaluated because surgical correlation was available.
A consensus review was also performed to resolve any discrepancies between interpretations on CTA and surgical findings.
Surgical Correlation
Laparoscopic donor nephrectomy was performed in 53 donors, 52 with left
renal harvest and one with right renal harvest. In two donors, a right donor
nephrectomy was performed using an open extraperitoneal approach after the
laparoscopic approach was deemed unfeasible. Surgery was performed between 2
weeks to 11 months (median, 4 months) after the CTA and CT urography
examination. For each patient, the transplant surgeons recorded a surgical
result including the number of arteries, the presence of early branching
arteries, the number of renal veins, the presence of late confluence within
veins, and the presence of major or minor renal vein anomalies. Because of the
limited field of view of surgery, not all minor renal venous anomalies could
be evaluated or recognized laparoscopically in each patient. For this study,
these were listed as false-positives because laparoscopy was deemed to be the
gold standard. However, minor renal venous variants needing surgical attention
were recorded.
Data Analysis
Considering surgical findings as the standard, we evaluated sensitivity and
accuracy for the number of main and accessory renal arteries and veins, the
number of renal arteries with early branching and number of renal veins with
late confluence, and the presence of major venous anomalies. For minor venous
anomalies, we evaluated only the sensitivity in correlation to surgical
results. These analyses were based on each reviewer's evaluation.
Interobserver agreement between two reviewers was assessed from the data of
independent reviews using kappa statistics.
|
|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Renal Arteries
At surgery, a total of 66 renal arteries were identified in 55 donor
kidneys. In 45 donor kidneys, one renal artery was identified; in nine donor
kidneys, two arteries were identified; and in one donor kidney, three arteries
were identified (Table 1). Both
reviewers agreed on the number of renal arteries in 54 (98%) of 55 harvested
kidneys, and interobserver agreement was excellent (kappa = 0.94). Overall,
both reviewers detected 65 of 66 surgically identified arteries (sensitivity:
98.5%). Overall accuracy for renal artery detection was 97% (reviewer 1) and
95.5% (reviewer 2). The false-negative was a missed accessory artery, which
was a 2-mm vessel that was laparoscopically transected without complication
(Fig. 2A and
2B). Reviewer 1 had one
false-positive and reviewer 2 had two false-positive diagnoses. The
false-positive cases had capsular arteries of 1 mm or less in diameter not
reported at surgery. One case was a small retroperitoneal artery, which was
misinterpreted by reviewer 2 as a capsular artery (Fig.
3A and
3B).
|
|
|
At surgery, early branching was found in 14 (21%) of 66 renal arteries. In 6 of the 14, two anastomoses were created in the recipients, and all had early bifurcation of less than 1.5 cm from the aorta on 16-MDCT angiography. In 8 of the 14, a single anastomosis was created in the recipients, and all had early bifurcation of between 1.1 cm and 2 cm on pre-operative 16-MDCT angiography (Table 2).
|
|
|
|
Renal Vein Anomalies
In 55 kidneys, a total of 63 veins were detected at surgery. Eight
supernumerary renal veins were detected at surgery, including five retroaortic
components of circumaortic left renal veins and three accessory right renal
veins (one right kidney had three renal veins and another had two renal
veins). Both reviewers agreed on the number of major renal veins and venous
anomalies in all 55 kidneys with excellent interobserver agreement (kappa =
0.9). The sensitivity and accuracy for detection of renal veins were 98.4% (62
of 63 for reviewer 1) and 97% (61 of 63 for reviewer 2), respectively. Major
renal venous anomalies, including one case of duplicated IVC and two cases of
retroaortic left renal vein, were all detected preoperatively by both
reviewers. All three accessory renal veins were detected preoperatively. Both
reviewers detected a less than 0.5-cm retroaortic vein component of three
circumaortic left renal veins, mischaracterizing the branches as a single main
renal vein with a draining small lumbar vein (Fig.
5A,
5B and
5C). In all three cases, the
small retroaortic vein branch was safely transected at surgery and a single
venous anastomosis was performed in the recipient.
|
|
|
Minor renal venous anomalies were those relating to the lumbar-gonadal axis. Surgically or radiologically described gonadal or lumbar veins were not seen in three of three right donor kidneys. On 16-MDCT angiography, both reviewers detected at least one gonadal vein in 45 kidneys (kappa, 0.71) and a draining lumbar vein in 36 kidneys (reviewer 1) and 38 kidneys (reviewer 2) (kappa, 0.92).
Twenty three minor left-sided venous anomalies were detected in 18 left kidneys at surgery including prominent (> 5 mm) lumbar veins draining into the main renal vein (11 patients), dual gonadal veins draining into a single main renal vein (three patients), prominent (> 5 mm) gonadal veins draining into left renal vein (two patients), and accessory renal veins draining into the lumbar-gonadal vein complex (seven patients). On 16-MDCT angiography, reviewer 1 detected in 19 of 23 (sensitivity, 83%) and reviewer 2 detected in 18 of 23 (sensitivity, 78%) minor venous anomalies. One accessory renal vein draining into the lumbar-gonadal complex was missed by both reviewers (Fig. 6A and 6B). Each reviewer missed two accessory renal veins individually. Both reviewers missed one case of a small vein connecting the gonadal vein with the inferior renal venous branch (Fig. 7). Reviewer 2 missed one case of a prominent (> 5 mm) lumbar vein draining into the renal vein. In this case, the donor was thin and the CT examination was performed in the late nephrographic phase because of scanner malfunction.
|
|
|
Surgical Morbidity in Donors
All 55 living renal donors successfully underwent nephrectomy without
complication. Minimal blood loss was described in surgical reports in all
patients including the two donors who underwent open nephrectomy.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
The left kidney is preferred for laparoscopic living donor nephrectomy because resection is easier than on the right. Also, transplantation of the donor kidney is facilitated by the longer artery and vein in the left kidney and less potential exists for vascular complications and eventual graft loss [27]. However, the reported incidence of renal venous anomalies and intrarenal collateral venous drainage from the renal-lumbar-azy-gos-vertebral axis [10] is more frequent in the left kidney than the right. Recent studies show that the presence of multiple left renal arteries and major anomalies of the left renal vein, including circumaortic and retroaortic left renal veins, if detected on preoperative imaging may not necessarily preclude a laparoscopic resection or increase morbidity to the donor [27-30]. Thus our urologists regard precise prior knowledge of anatomy and variants in the vascular system and urinary tract to be essential before performing a laparoscopic nephrectomy.
Cross-sectional imaging primarily by CT replaced other methods of preoperative imaging in the 1990s, especially with regard to open nephrectomy [11-19]. However, up to 15% of small arteries were missed by single and 4-MDCT angiography [25]. Unanticipated bleeding from accidental transection of small vessels is more difficult to control during laparoscopic harvest, although they are usually not relevant for graft function. Only one study previously reported the performance of cross-sectional angiographic techniques using single-detector CT with respect to some venous anatomy and variants using an open nephrectomy standard [19].
In this study, we have showed that using a multiphasic 16-MDCT angiography protocol and 3D reviewing on a dedicated workstation, very high sensitivity and accuracy is possible preoperatively for a range of laparoscopically relevant renal arterial and venous variants. We showed excellent agreement between both blinded reviewers with each other (kappa = 0.97) and excellent sensitivity (98%) and accuracy (> 95%) in evaluation of the number of renal arteries. Only a 2-mm supernumerary renal artery was missed by both reviewers on blinded review, and it was transected safely during surgery without complication. Even with a subjective gold standard, we had excellent sensitivity for detection of early renal arterial branching, with sensitivity and accuracy of 100%, compared with previous reports of 86-100% and 95-98%, respectively [21, 22]. Unlike a surgically determined gold standard with regard to the number of renal arteries or veins, which tends to be more objective, early arterial branching tends to be subjective and varies among surgeons and institutions.
Our sensitivity and accuracy for detection of supernumerary veins and major venous anomalies approached 100%, similar to results achieved with arteries. These results compare favorably to previous studies with regard to accuracy and interobserver agreement (kappa = 0.9 vs kappa = 0.85) [21, 22]. Also, 16-MDCT angiography enabled excellent results for detecting renal venous branching patterns and the concept of late venous confluence. Although this result is subjective and varies among both institutions and surgeons (analogous to early arterial bifurcation), we achieved 100% sensitivity and accuracy for this finding. Also, the laparoscopic surgeons were able to create a single vein anastomosis without significant bleeding.
To our knowledge, detection and characterization of minor venous anomalies related to the confluence of the renal veins with both lumbar and gonadal veins have not been well characterized by imaging. The relationships among the renal, lumbar, and gonadal veins are variable. Accessory renal veins often anastomose to lumbar or gonadal veins before joining the main venous pedicle. One or more lumbar veins communicate with the left main renal vein posteriorly in approximately 43-75% of patients [21, 31-33]. Because the posterior aspect of the renal vein cannot usually be directly visualized during laparoscopic nephrectomy, this information may be important preoperatively. However, the need for preoperative detection and characterization of these anomalies varies institutionally. Using our 16-MDCT angiography protocol, we achieved sensitivities and accuracies of 78-83% in detection of these anomalies. These results reflect the difficulty of identifying some variants for a variety of reasons including their complexity and variability, our learning curve in detection, and optimizing CT protocols to maximize enhancement of these small vessels.
Our study has limitations that must be acknowledged. By design, the study is biased against complex variants and right-sided variants because these did not undergo laparoscopic nephrectomy. The surgical gold standard is variable between arteries and veins, especially when related to branching pattern and distance. Surgeons were unable to record an exact distance between the aorta and the first branch of the renal artery or the confluence of renal veins because of a limited field of view. Some minor venous anomalies detected on CTA were not proved because of the limited surgical field of view. Therefore, if these findings did not affect the laparoscopic nephrectomy or if the surgeons did not visualize these anomalies, they were not described in the surgical report and for the purposes of this study were recorded as a false-positive. This likely decreased the specificity and accuracy for detection of these lumbar and gonadal variants.
Laparoscopic donor nephrectomy is now an established but evolving technique with a steep learning curve for both surgeons and radiologists. However excellent preoperative imaging results are possible with a good MDCT angiography technique and with experienced reviewers, as we have shown in this study. We need to improve our understanding of variants in the lumbar-gonadal vein axis to improve our performance in characterizing these anomalies.
In conclusion, we have shown that 16-MDCT angiography combined with 3D-workstation reviewing provides excellent preoperative arterial and venous anatomic information for laparoscopic renal donor nephrectomy.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
This article has been cited by other articles:
|
|
 |
|
  J. Burrill, Z. Dabbagh, F. Gollub, and M. Hamady Multidetector computed tomographic angiography of the cardiovascular system Postgrad. Med. J., November 1, 2007; 83(985): 698 - 704. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. S. Raman, S. Pojchamarnwiputh, K. Muangsomboon, P. G. Schulam, H. A. Gritsch, and D. S. K. Lu Surgically Relevant Normal and Variant Renal Parenchymal and Vascular Anatomy in Preoperative 16-MDCT Evaluation of Potential Laparoscopic Renal Donors Am. J. Roentgenol., January 1, 2007; 188(1): 105 - 114. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
