Comparison of computed tomographic angiography and noncontrast magnetic resonance angiography in preoperative evaluation of living renal donors.

INTRODUCTION: The computed tomographic angiography (CTA) renal donor protocol is an established method of preoperative renal vascular pedicle evaluation in prospective renal donors. However, CTA is associated with significant radiation exposure and intravenous contrast administration. The newer noncontrast-enhanced magnetic resonance angiography (NCE-MRA) techniques, especially arterial spin labeling (ASL) with steady-state free precession (SSFP) hold promise as an effective alternative. We prospectively compared CTA with NCE MRA for accuracy in the evaluation of renal arterial anatomy in prospective renal donors.
METHODS: A total of 43 subjects underwent CTA followed by NCE MRA in a prospective comparative study. The number of renal arteries and early branching of renal arteries were noted in both kidneys in all subjects. Intermodality agreement was calculated using "K" (Kappa) statistics and 95% confidence interval for both modalities.
RESULTS: A total of 63 single, 21 double, and 2 triple arteries were detected in 43 subjects on CTA. CTA showed an early branch in 17 kidneys. NCE MRAshowed 64 single arteries, 20 double arteries, and 2 triple arteries. A total of 14 kidneys showed an early branch. Unweighted Kappa statistic of agreement between CTA and NCE MRA for number of renal arteries and for frequency of early branching was 0.9707 and 0.8822, respectively.
CONCLUSIONS: The newer NCE MRA techniques such as ASL with SSFP among others are potential alternatives for CTA, in the evaluation of prospective renal donors.

INTRODUCTION

Renal transplant continues to be the best therapeutic choice for end-stage renal disease.[1234] Anatomic assessment of the donor kidney is performed to aid in the selection of which kidney to use and to plan the surgical approach. Computed tomographic (CT) or magnetic resonance angiography (MRA) is the currently used for this purpose.[56789]

CT involves high doses of radiation, resulting in a marked increase in radiation exposure in the population. Most of the candidates for renal donation are healthy and young individuals. Magnetic resonance imaging (MRI) has the advantage of relying on the intrinsic magnetic properties of body tissues and blood to produce an image, without the need of ionizing radiation or nephrotoxic contrast agents. While the contrast-enhanced (CE) approach has shown excellent diagnostic performance in diverse applications,[10111213] intravenous administration of contrast agents increases patient discomfort as well as examination costs and limits the achievable spatial resolution and artery-to-background contrast. In addition, the risk of nephrogenic systemic fibrosis has not been completely cleared, particularly in patients with renal failure.[12]

Non-CE (NCE) MRA is free from the aforementioned limitations and has been an active MR research area in the past decade.[13] Another NCE renal MRA technique - a nonsubtractive version of arterial spin labeling (ASL), slab-selective (SS) inversion recovery (IR) imaging - has become popular.[1415161718]

In this study, we compared CT angiography (CTA) performed routinely in prospective renal donors after the administration of iodinated contrast material with newer NCE-MRA, ASL with steady-state free precession (SSFP).

METHODS

Prospective renal donors referred to our Radiology department for CT renal angiogram protocol were included in the study after written informed consent. The exclusion criteria was age lower than 18 years or higher than 80 years, renal insufficiency, refusal or inability to sign the informed consent form, pregnancy, lactation, history of adverse reaction caused by iodine or gadolinium containing contrast agents, history of prior allergic reaction (not related to a contrast agent) associated with cardiorespiratory arrest, larynx edema, bronchospasm, angioedema or neurologic complications, contraindications to MRI such as cardiac pacemaker, metallic implants, and severe claustrophobia.

Each prospective renal donor underwent CTA protocol followed by NCE MRA protocol. Details of CTA and NCE MRA protocol are given in Tables 1 and 2, respectively. MR images were obtained with a Hi Speed Signa 1.5-T magnet (GE Medical Systems; Software version 5.6 Optima 450w Milwaukee, United States,). T2 coronal and axial images were taken to locate the renal area. Unenhanced-MRA was performed using a respiratory-triggered 3D fat saturated fast imaging employing steady-state acquisition with IR pulses (Inhance 3D Inflow IR; GE). Inhance 3D inflow IR is an unenhanced-MRA sequence based on the inherent inflow effects of blood. The SS-IR pulse sequence consists of an SS inversion pulse that excites the imaging volume as well as inferior veins, a subsequent delay time, and 3DFT balanced SSFP (b-SSFP) readout.[1920] A simple workflow example illustrating the technique for renal MRA is shown below in Figure 1. The area of interest is first saturated with one or more 180°-RF pulses. These saturation pulses invert signal from background tissue as well as venous blood. “Fresh” (unsaturated/fully magnetized) arterial blood next flows into the slab. After an appropriate inversion time (TI) that nulls signal from background tissue, MRA signal generation is initiated using a rapid 3D b-SSFP sequence.

Table 1

Scan parameters for data acquisition and postprocessing in renal angiography study for pretransplant living renal donor evaluation; using 64-slice multidetector computed tomography scanners

Table 2

Inhance three-dimensional inflow inversion recovery sequence parameters

Figure 1

The technique of renal magnetic resonance angiography. (a-c) Inversion pulse is applied in the field of view (blue box in a), followed by inflow of fresh blood from upstream the field of view (red arrows in b). (c) region of interest (yellow box)

The images from each CT examination and each MR examination were reviewed at a picture archiving and communication system workstation with cine capability. Two experienced radiologist evaluated the CTA and noncontrast MRA images. Both reviewers were blinded to the other imaging modality while analyzing either CTA or MRA. A study worksheet was completed independently for each examination; the worksheet detailed the numbers and sizes of renal arteries found on each side. For each artery, the presence of any stenosis or of a proximal branch was noted. For statistical analysis, any branch within 2.0 cm from the aorta was classified as a proximal branch.

A transplant surgeon completed a separate worksheet for each subject who underwent nephrectomy. The surgical worksheet identified the number of arteries and size of each artery and the presence of any proximal branch. Any unexpected surgical findings were described.

The findings of CTA and MRA were compared for intermodality agreement using Cohen's Kappa coefficient. The frequency of multiple renal arteries and proximal arterial branches were tabulated on the basis of the readings for each CT and MRA. Each frequency value was computed “by kidney”, and 95% confidence intervals (CIs) were computed.

A total of 43 patients were included in the study. To test 90% intermodality agreement between CTA and MRA with 95% CI, the sample size required is 43 and which obtained by using the equation n = k (1 − k)/1 − Pc × (Z-alpha/delta)2. We fixed k as 0.90 (90% intermodality agreement - agreement between CT and MRI) and Pc (precision) as 0.20. A level of significance “alpha” as 0.05, Delta = 0.10 so that determined k must be in ±10%.

The, SPSS 17.0, (Chicago, United States) statistical software was used for the analysis of the data and Microsoft word and Excel have been used to generate graphs, tables, etc.

RESULTS

The study was started in June 2014. A total of 43 subjects were studied including 37 females. The donors were distributed within the ages of 27 and 65 with an average age of 43 years.

CTA readings recorded 23 supernumerary arteries and 63 single arteries in 43 subjects, whereas MRA readings recorded 22 supernumerary arteries and 64 single [Table 3]. “K” measure of agreement is 0.9707 (95% CI = 0.9137–1; standard error 0.029). There were no discordant readings between two radiologists with respect to supernumerary arteries on CTA as well as on NE-MRA.

Table 3

Comparision frequency of renal arteries

CTA readings recorded early branch in 17 renal arteries and no early branch in 69 renal arteries. There were no discordant readings between two radiologists with respect to early branching on CTA; whereas for NE-MRA, the first radiologist reported early branch in 15 renal arteries. The second radiologist reported early branch in 16 renal arteries. Both the radiologists agreed on 14 renal arteries as having early branch. Hence, total three renal arteries had discordant reporting on NE-MRA, which were considered as normal branches instead of early branches in the final reporting for NE-MRA [Table 4]. “K” measure of agreement is 0.8822 (95% CI = 0.7522–1; standard error 0.0664).

Table 4

Comparision of frequency of early renal branching

None of the kidneys showed focal lesions. Simple cortical cysts with average size of 2 cm were found in 8% of cases. None of the subject found to have nephrolithiasis.

The surgical correlation was available for all 43 cases. However, as the side with less complex vascular anatomy was selected for nephrectomy surgical correlation was usually not available for variant renal anatomy. Two small upper capsular arteries (<2 mm) were missed by CTA and NCE MRA.

DISCUSSION

At present CT renal angiography is the imaging modality of choice for morphological evaluation of kidneys and renal vascular pedicle.[4589] As CTA is associated with significant amount of radiation, and intravenous contrast administration, several studies evaluated the role of MR in preoperative renal donor evaluation.[678] The NC MRA technique of ASL with 3D SSFP sequence has been evaluated in some recent studies to assess renal vascular anatomy and has produced promising results.[2122] In the current study, we compared Inhance Inflow IR Renal MRA, basically ASL with 3D SSFP sequence and CTA in preoperative renal donors evaluation.

In this study, a novel SSFP MR sequence was used with an SS inversion prepulse for selective renal MRA [Figures 2 and 3]. The SSFP technique was used because the previous studies demonstrated that in addition to a high SNR, this sequence bears intrinsically high (T2-like) contrast between the blood pool and soft tissue.[141516171819202122] SSFP imaging is supposed to be flow compensated in all three spatial coordinates due to the symmetric shape of the gradient pulses.[1823] The application of an SS inversion prepulse as used in this study resulted in a good signal suppression of static tissue and veins, which allowed the selective visualization of the renal arteries without the use of any contrast medium application. Average time required for the completion of sequence and postprocessing for noncontrast MRA was approximately 10–15 min and for CT, it was approximately 10 min. As the study was done to evaluate only renal artery anatomy and its variations, there is not much time difference between both modalities. However, time limitation will be a factor if ureteric anatomy and venous anatomy are also considered in the MR imaging which will require additional sequences.

Figure 2

(a and b) Supernumerary renal arteries noncontrast-enhanced magnetic resonance angiography and computed tomography angiography three renal arteries on the right side and single renal artery on the left side are seen in coronal maximum intensity projection image of noncontrast enhanced magnetic resonance angiography (a) and volume-rendered image on computed tomography angiography (b)

Figure 3

(a) Volume rendered and (b) coronal maximum intensity projection reformations noncontrast enhanced magnetic resonance angiography of different patients showing single bilateral renal arteries with polar branch (curved arrow) and gonadal artery (middle arrow in b). (c and d) Coronal three-dimensional reformatted computed tomography angiography image (c) and coronal maximum intensity projection noncontrast-enhanced magnetic resonance angiography image (d) in same patient showing early branching (12.8 mm from ostium) on right side

Female predominance in donors (6:1) was observed and was in concordance with previous studies.[324] Although MRI is inferior to CT in documenting stones, none of the subjects in our study had significant stone disease as renal calcifications in young donors have very low prevalence.[25] Although this may be considered as drawback of MRA over CTA, the presence of significant stone disease can be easily assessed with another easily available, radiation free, less expensive modality such as ultrasonography.

The incidence of variant renal arterial anatomy was 35%, which correlated with earlier studies. One accessory lower polar renal artery was missed by NCE MRA which was visualized originating from proximal right common iliac, renal artery, and supplying the lower pole of the left kidney on CT. The right kidney was selected for transplantation in this subject based on complex left side arterial and venous anatomy on CT readings [Figure 4]. Both the modalities showed a good agreement for supernumerary renal arteries. As the noncontrast MRA technique used in the study was dependent on inflow effect of the blood during systole, the craniocaudal coverage of the abdominal aorta was limited by flow velocity of blood. Moreover, the left lower polar artery was not in the field of view (FOV) through most of its course. However, retrospective evaluation of the above-mentioned subject showed distal portion of the accessory renal artery which was in the FOV albeit with significant low intensity. As the polar artery was originating from a distal circulation, it was expected.

Figure 4

(a and b) Coronal reformatted maximum intensity projection computed tomography angiography images (a) left lower polar accessory renal artery originating from the right common iliac artery (arrow in a). This accessory left lower polar artery was missed on noncontrast enhanced magnetic resonance angiography (b)

Some studies have shown further modifications of NCE MRA, with Shin et al. used NCE MRA method with velocity-selective inversion preparation for renal artery assessment.[23] This technique is said to be more robust with increased craniocaudal coverage as compared to SS-IR method. Atanasova et al. used an approach of a global inversion followed by a sagittally oriented SS inversion of the abdominal aorta to invert all tissues, but arterial blood in the aorta so that arterial inflow can start from the origin of the iliac arteries.[26] Another flow-dependent method employed flow-sensitive dephasing preparation to obtain a dark-artery image, which was subtracted from a reference image to generate an artery-only image.[23]

Both CTA and NCE MRA showed 14 early branches, whereas 69 renal arteries with no early branch in 43 subjects. CTA showed that additional three early arterial branches which were not shown as early branch on NCE MRA. The three early branches which were seen as early branch on CTA were detected on NCE MRA as normal distal branches. Hence, NCE MRA overestimated the length of the renal artery before branching as early branching is said to present within 2 cm from aorta. The “K” measure of agreement was 0.88 for early arterial branching between CTA and NCE MRA. This was marginally unacceptable as per our assumptions.

Based on the readings of CTA, the surgeon selected the side of nephrectomy. The surgical correlation was available for all 43 cases. However, as the side with less complex vascular anatomy was selected for nephrectomy, surgical correlation was usually not available for variant renal anatomy. Two small upper capsular arteries (<2 mm) were missed by CTA and NCE MRA. However, both of the misreading had no significant impact on outcome was noted.

In our study, there were no discordant readings in CTA by both observers in detection of supernumerary renal arteries and early branching. However, there were discordant readings in MRA in documenting early branching. Discordant readings might be largely due to interobserver variability. In addition, a definite learning curve has been shown for renal vasculature evaluation by the previous studies.

The major limitation of our study was the fact that no external standard was available for comparison. As the surgeon, based on CTA findings selected the kidney with less complex vascular anatomy, the surgical findings of variant renal anatomy were not available for comparison. Nonetheless, CTA has been established as the imaging modality for preoperative evaluation of renal donors by the previous studies and NCE MRA showed an acceptable level of agreement with CTA. As the operative field during laparoscopic surgery is very limited, the variant venous anatomy is as important as arterial anatomy to the laparoscopic surgeon. In addition, the impact of missed renal arteries/early branches on clinical outcome of the patient was not studied in the present study. The missed renal artery/early branch may increase procedure time and or additional back table procedure with more frequent reconstructed arteries rather than an unacceptable clinical outcome.[6]

Hence, findings of the present study suggest that newer NCE MRA using ASL with SSFP technique shows potential as an effective alternative for CTA renal protocol. However, further studies with regard to the use of NCE MRA for the evaluation of renal venous system may be required before clinical use, as well as studies with regard to impact of discordant NCE MRA readings and interobserver variability on clinical outcome of renal transplant procedures may further validate the use of NCE MRA for the purpose.

CONCLUSIONS

Newer methods of nonenhanced MRA, especially ASL with SSFP hold promise as substitute for established CTA for evaluation of renal donors. Studies on the impact of missed renal artery on clinical outcome may further validate the use of newer NCE MRA techniques The accuracy of NCE MRA in the evaluation of renal venous anatomy needs to be studied before clinical use of NCE MRA for preoperative evaluation of candidates for prospective renal donation.