PRRACDa

O

M

R

C

Araobrnrabmst

t

FLC

3l

©A

Ambulatory and Office Urology

rospective Comparison of Magneticesonance Angiography with Selectiveenal Angiography for Living Kidney Donorssessment

hristopher Neville, Andrew A. House, Christopher Y. Nguan, Kenneth A. Beasley,avid Peck, Lisa M. F. Thain, Richard Rankin, Vivian C. McAlister, Alison R. Spouge,nd Patrick P. W. Luke

BJECTIVES For years, the reference standard in the evaluation of living donor vascular anatomy has beenselective renal angiography (SRA). Because of the potential morbidity associated with SRA, weprospectively evaluated magnetic resonance angiography (MRA) in the assessment of renaldonors.

ETHODS All patients had SRA and 53 renal units were prospectively evaluated by MRA. We used SRAsupplemented by findings at donor nephrectomy (DN) as our standard. We defined a positive testas the detection of any abnormality in the number of renal arteries.

ESULTS Selective renal angiography yielded a sensitivity of 86%, specificity of 95%, positive predictivevalue (PPV) of 75%, and negative predictive value (NPV) of 97% compared with findings atDN. MRA had a sensitivity of 64%, 88% specificity, 58% PPV, and 90% NPV. MRA correctlyidentified only 7 of 11 renal units with accessory arteries. MRA also incorrectly identified 5accessory arteries not present on SRA or DN. Two patients diagnosed with fibromusculardysplasia by SRA were missed using MRA.

ONCLUSIONS We have shown that MRA is not capable of replacing SRA as the reference standard in renal

donor imaging. UROLOGY 71: 385–389, 2008. © 2008 Elsevier Inc.

iuvctatwrd

M

BtMo

MWaps

ccurate preoperative evaluation of donor renalanatomy before nephrectomy is crucial to ensureoptimal outcomes for both kidney donor and

ecipient in transplantation. In the past, selective renalngiography (SRA) has been performed in the evaluationf renal donor vascular anatomy to determine the num-er, size, and location of renal vessels, and to excludeenal vascular and parenchymal disease.1– 4 However, re-al angiography is invasive and has been associated withare but significant complications including renal failure,neurysm formation, and anaphylaxis. Consequently,oth computed tomography angiography (CTA) andagnetic resonance angiography (MRA) have been con-

idered minimally invasive alternatives to replace SRA inhe evaluation of renal donor anatomy.

Both MRA and CTA have independently been showno be accurate in the detection of accessory renal arter-

rom the Division of Urology and Departments of Radiology, Nephrology, and Surgery,ondon Health Sciences Centre, The University of Western Ontario, London, Ontario,anadaReprint requests: Patrick Luke, London Health Sciences Centre–University Campus,

39 Windermere Road, London, Ontario, Canada, N6A 5A5. E-mail: patrick.luke@

whsc.on.ca

Submitted: December 22, 2006, accepted (with revisions): October 19, 2007

2008 Elsevier Inc.ll Rights Reserved

es.5–12 Both can also assess renal parenchymal disease,reteral anomaly, circumaortic renal veins, and otherariants, as well as screen for intraabdominal disease. Theontinuing evolution of these modalities, combined withheir minimally invasive nature, makes them attractivelternatives to SRA in the evaluation of potential renalransplant donors. To justify a transition away from SRA,e prospectively assessed the ability of MRA to identify

enal vascular anomalies in the evaluation of living renalonors.

ATERIAL AND METHODS

etween June 2000 and Oct 2002, we evaluated potential renalransplant donor renal units preoperatively using SRA and

RA. Independent of the SRA findings, cross-sectional radi-logists performed prospective interpretation of MRA.

RA Techniquee carried out MR imaging with a GE 1.5-T CVMR unit using

torso coil. The renal parenchyma was assessed with pre- andost- gadolinium coronal and axial three-dimensional, fastpoiled gradient echo with spectral inversion at lipid and T2-

eighted single shot fast spin echo. The renal vessels were

0090-4295/08/$34.00 385doi:10.1016/j.urology.2007.10.030

ipmteantedtMi(da

SWfoncisfDn

SArvtnameotsdi

mwCt0

RTdtrpvsrsfi

faitntmvsTds0

wtbccteaFalwaFbbrwitsr

3

maged during a 20- to 30-second breath-hold using the coronallane following injection of 0.1 mm/kg gadopentetate dimeglu-ine (Magnevist; Berlex Laboratories, Wayne, NJ) with a fast

ime of flight spoiled gradient echo using the following param-ters: repetition time 4.9 milliseconds, minimum echo time, flipngle 34°, bandwidth � 62.5, 40-cm field of view, slice thick-ess 2.8 mm with zero interpolation resulting in an effective slicehickness of 1.4 mm, a matrix of 512 frequency � 256 phasencodings, and 1 excitation. Three-phase scanning was performeduring arterial, venous, and delayed phases postinjection of con-rast. We used automated triggering software (Smart Prep; GE

edical Systems) to determine the optimal time for arterial phasemaging after contrast injection. Two experienced MR radiologistsAS and LT) evaluated images primarily by reviewing the sourceata set supplemented by maximum intensity projection reformatsfter digital subtraction.

RA Techniquee performed angiographic evaluation by puncturing the right

emoral artery in a retrograde fashion after infiltration of theverlying skin with local anesthetic. Using the Seldinger tech-ique, a 5-Fr sheath was established and a 5-Fr straight flushatheter was advanced and positioned in the suprarenal abdom-nal aorta. We obtained an AP abdominal aortogram. Thetraight catheter was then exchanged for a Simmons 1 catheteror selective renal arterial cannulation and injection of contrast.elayed films were obtained to assess renal parenchymal, ve-ous and collecting system anatomy.

tatistical Analysisfter donor imaging, we collected data detailing the number of

enal vessels, parenchymal abnormalities, and presence of renalascular disease for each renal unit. In addition, we recordedhe number of renal arteries identified at the time of donorephrectomy in a separate table. We compared the number ofrteries identified by MRA and that identified on SRA supple-ented by findings at donor nephrectomy, which was consid-

red the reference standard. To evaluate the accuracy of SRA inur study population, we first compared the results of SRA withhe findings at the time of donor nephrectomy. Sensitivity,pecificity, positive predictive value (PPV), and negative pre-ictive value (NPV) were calculated for each noninvasivemaging modality versus the standard.

As a measure of diagnostic accuracy or agreement, we esti-ated Kappa statistics for each imaging modality comparedith the reference standard (SRA plus surgical findings) usinglinStat (St. George’s Hospital Medical School). Kappa less

han 0.1 indicated no correlation, whereas Kappa greater than.8 indicated excellent correlation.6

ESULTSo establish a baseline assessment of our current stan-ard, we compared SRA with findings at donor nephrec-omy for 45 renal units. SRA correctly identified 6 of 7enal units with supernumerary vessels but falsely re-orted multiple vessels in 2 of 38 kidneys with normalasculature. Because the renal vasculature was recon-tructed in the case that involved the missed accessoryenal artery, this vessel was believed to be clinicallyignificant. Sensitivity was calculated to be 86%, speci-

city was 95%, PPV � 75%, and NPV � 97%. t

86

With respect to MRA, we evaluated 54 renal unitsrom 27 potential donors. One donor had a pelvic kidneynd the contralateral normal kidney was not adequatelymaged, so it was not included in the analysis, decreasinghe renal units to 53. Of these, 11 renal units had super-umerary vessels and 42 units had normal renal vascula-ure. MRA failed to recognize 4 of 11 cases of supernu-erary vessels, and incorrectly suggested supernumerary

essels in 5 of 42 units. This yielded a sensitivity of 64%,pecificity of 88%, PPV of 58%, and NPV of 90% (Table 1).he Kappa statistic was 0.48, consistent with a moderateegree of agreement between MRA and the referencetandard (95% confidence interval 0.19 to 0.77; P �.0041).In addition to numeric assessment of accessory vessels,

e also assessed renal vascular anomalies. We diagnosedwo patients with bilateral fibromuscular dysplasia (FMD)y SRA. However, using MRA, the diagnosis of FMDould not be established in these same two cases. Recal-ulation of the Kappa statistics of MRA using the detec-ion of accessory renal arteries and/or renal arterial dis-ase as the end point would have led to a Kappa of 0.76nd 0.32, respectively. In one of these cases (patient 1),MD affected both the left and right main renal arteries,s well as segmental branches (Fig. 1). Because we be-ieved that this 48-year-old woman with bilateral FMDas at risk for progression of disease, she was excluded asrenal donor. In the second case, a solitary segment of

MD was discovered in the left superior renal arterialranch using SRA. Again, using MRA, FMD could note definitively diagnosed. Because we believed that theisk for disease progression in a 62-year-old female donorith unilateral segmental FMD was low, we obtained

nformed consent and performed successful living-relatedransplantation using the diseased kidney with recon-truction of the affected segment. Neither donor norecipient is currently receiving antihypertensive medica-

Table 1. Ability of MRA to detect accessory renal arteries

Parameter MRA vs. Reference Standard

No. 27 (53 renal units)Positive tests 11TN 37FN 4TP 7FP 5Sensitivity 0.64Specificity 0.88PPV 0.58NPV 0.90K 0.4895% confidence interval 0.19–0.77P-value 0.0041

FN � false negative; FP � false positive; NPV � negative predic-tive value; PPV � positive predictive value; TN � true negative; TP� true positive.

Magnetic resonance angiography (MRA) results are correlatedwith reference standard findings (renal angiography supple-mented by findings at donor nephrectomy).

ions and the serum creatinine of the donor and recipient

UROLOGY 71 (3), 2008

ip

DAt

nhbrnin

SWrsOoSpfidfnlirbpFtrua

piioip

primTcutwsgmbocos

Ffitpbd

U

s 90 and 120 �mol/L, respectively, with over 2.5 years’ostoperative follow-up.

ISCUSSIONlthough early MRA studies demonstrated poor sensi-

igure 1. Comparison of imaging modalities in detectingbromuscular dysplasia. (A) Selective renal angiography ofhe left renal artery in patient 1 reveals fibromuscular dys-lasia (FMD) in the main renal artery and renal arteryranches. (B) Magnetic resonance angiography is unable toetect FMD in the left renal artery in patient 1.

ivity in the detection of accessory renal arteries in kid- c

ROLOGY 71 (3), 2008

ey donors,13,14 studies using more advanced technologyave concluded that both MRA can determine the num-er of accessory renal arteries with reasonable accu-acy.9,15–17 Before replacing SRA evaluation of the pote-tial renal donor with cross-sectional imaging at our

nstitution, validation of the performance of MRA tech-ologies versus the reference standard was warranted.Previous studies have compared MRA with either

RA or findings at the time of donor nephrectomy.16–21

ith 18% of evaluated renal units having supernumeraryenal arteries in our series, we found that SRA had aensitivity and specificity of 86% and 95%, respectively.f note, these measures of diagnostic accuracy are likely

verestimates, because significant abnormal findings onRA would preclude donor nephrectomy (leading to aotential underestimate of false positives), and negativendings on SRA would never be confirmed surgically ifonor nephrectomy of the contralateral kidney were per-ormed (leading to a potential underestimate of falseegatives). Hence, SRA is itself an imperfect test with

imited ability to detect accessory renal vessels, and stud-es comparing imaging modalities without surgical cor-oboration may give misleading results. In addition,ecause anatomic discovery at the time of donor ne-hrectomy may miss intrinsic vascular disease such asMD, we supplemented SRA results with findings at theime of donor nephrectomy for the establishment of oureference standard. The definition of our standard isnique and has subsequently led to important findingsnd policy changes at our institution.

Our results however, indicate that MRA correlatesoorly with the reference standard and is significantlynferior to SRA in the detection of accessory renal arter-es. This finding is consistent with results from severalther studies12,15,19,20,22 that assessed MRA with SRA orntraoperative findings. Furthermore, MRA was not ca-able of demonstrating FMD with any reassurance.Fibromuscular dysplasia may result in renin-mediated hy-

ertension resulting from renal artery stenosis or, morearely, result in dissection or rupture of the renal artery. Thencidence of FMD in potential renal donors has been esti-ated at 2.0% to 6.6% in normotensive, healthy adults.23,24

he most common form, medial fibroplasia, occurs moreommonly in females and tends to progress in patientsnder 50 years of age.25 Furthermore, Cragg showed that upo one-quarter of asymptomatic patients with FMD whoere excluded from kidney donation developed hyperten-

ion requiring medical therapy or angioplasty.23 This pro-ression occurred at a significantly higher rate than age-atched normotensive controls. Therefore, FMD appears to

e a clinically significant entity that can affect the outcomef the asymptomatic, normotensive individual. Theoreti-ally, more serious consequences may arise in the recipientf a donor kidney with FMD, or the donor harboring aolitary kidney with FMD.

Alternatively, several authors have concluded that

arefully selected patients with FMD can safely be used as

387

kamFetdFdwenburapedptMnt

aFtitAcmddfdtrkt

CAtav

R

1

1

1

1

1

1

1

1

1

1

2

2

2

2

2

2

3

idney donors.26 To optimize renal functional capacitynd reduce the risk of complications, Novick recom-ended that renal arteries or branches afflicted with

MD should be reconstructed to exclude or remove dis-ased vessels.25 Our series highlights the implications ofhe FMD diagnosis. Two cases of FMD (4.3%) wereiagnosed by SRA in our study. MRA failed to identifyMD in 2 of 2 cases. In one case with bilateral FMDiscovered in both right and left main renal arteries asell as the majority of branches on SRA, the patient wasxcluded from renal donation. Significantly, MRA couldot detect FMD in the main arteries and multipleranches. The other individual had isolated FMD in thepper pole branch of the left renal artery. Because FMDarely progresses in individuals greater than 50 years ofge, and the right renal vasculature was normal in thisatient, transplantation and reconstruction of the dis-ased left kidney were performed. Without SRA, livingonation would have occurred in patient 1, potentiallylacing both renal donor and recipient at risk for hyper-ension and/or renal failure. In patient 2, the inability of

RA to detect FMD may have led to selection of theondiseased kidney for donation and possible complica-ions in the donor.

In addition to our group, others have also raised concernsbout the ability of cross-sectional imaging to detectMD.13,27–30 The clinical relevance of our study depends onhe institutional policy on FMD. For many centers, allndividuals with FMD are excluded from donating a kidney;herefore, MRA could not adequately screen renal donors.t our institution, the finding of FMD is thought to be

linically relevant and is an important factor in the assess-ent of potential renal donors. However, we believe that a

iagnosis of FMD by itself should not exclude a patient fromonating a kidney. As we have shown, SRA provides in-ormation that can be used for both appropriate kidney andonor selection and exclusion. Most important, we thinkhat potential donors with FMD should be informed aboutisk of disease progression, and if they choose to donate aidney, intensive clinical and biochemical follow-up ofhese donors is mandatory.

ONCLUSIONSlthough evolution of emerging technologies may even-

ually support the use of MRA in evaluating renal donornatomy, MRA is inadequate in detecting supernumeraryessels and vascular anomalies.

eferences1. Adamis MK, Goldszer RC, Pulde MF, et al: Renal vasculature in

potential renal transplant donors: comparison of MR imaging anddigital subtraction angiography. Radiology 197: 467–472, 1995.

2. Chkhotua AB, Klein T, Shabtai E, et al: Kidney transplantationfrom living-unrelated donors: comparison of outcome with living-related and cadaveric transplants under current immunosuppressiveprotocols. Urology 62: 1002–1006, 2003.

3. Derauf B, and Goldberg ME: Angiographic assessment of potential

renal transplant donors. Radiol Clin North Am 25: 261–265, 1987.

88

4. Strauser GD, Stables DP, and Weil R III: Optimal technique ofrenal arteriography in living renal transplant donors. AJR Am JRoentgenol 131: 813–816, 1978.

5. Bakker J, Ligtenberg G, Beek FJ, et al: Preoperative evaluation ofliving renal donors with gadolinium-enhanced magnetic resonanceangiography. Transplantation 67: 1167–1172, 1995.

6. Giessing M, Kroencke TJ, Taupitz M, et al: Gadolinium-enhanced three-dimensional magnetic resonance angiography versus conventional dig-ital subtraction angiography: which modality is superior in evaluatingliving kidney donors? Transplantation 76: 1000–1002, 2003.

7. Kapoor A, Kapoor A, Mahajan G, et al: Multispiral computed tomo-graphic angiography of renal arteries of live potential renal donors: areview of 118 cases. Transplantation 77: 1535–1539, 2004.

8. Khamanarong K, Prachaney P, Utraravichien A, et al: Anatomy ofrenal arterial supply. Clin Anat 17: 334–336, 2004.

9. Low RN, Martinez AG, Steinberg SM, et al: Potential renal trans-plant donor: evaluation with gadolinium-enhanced MR angiogra-phy and MR urography. Radiology 207: 165–172, 1998.

0. Patil UD, Ragavan A, Nadaraj, et al: Helical CT angiography inevaluation of live kidney donors. Nephrol Dial Transplant 16:1900–1904, 2001.

1. Winterer JT, Strey C, Wolffram C, et al: Preoperative examination ofpotential kidney transplantation donors: value of gadolinium-enhanced3D MR angiography in comparison with DSA and urography. RofoFortschr Geb Rontgenstr Neuen Bildgeb Verfahr 172: 449–457, 2000.

2. Tsuda K, Murakami T, Kim T, et al: Helical CT angiography ofliving renal donors: comparison with 3D Fourier transformationphase contrast MRA. J Comput Assist Tomogr 22: 186–193, 1998.

3. Gourlay WA, Yucel EK, Hakaim AG, et al: Magnetic resonanceangiography in the evaluation of living-related renal donors. Trans-plantation 60: 1363–1366, 1995.

4. Meyers SP, Talagala SL, Totterman S, et al: Evaluation of the renalarteries in kidney donors: value of three-dimensional phase-con-trast MR angiography with maximum-intensity-projection or sur-face rendering. Am J Roentgenol 164: 117–121, 1995.

5. Buzzas GR, Shield CF, Pay NT, et al: Use of gadolinium-enhanced,ultrafast, three-dimensional, spoiled gradient-echo magnetic reso-nance angiography in the pre-operative evaluation of living renalallograft donors. Transplantation 64: 1734–1737, 1997.

6. Halpern EJ, Mitchell DG, Wechsler RJ, et al: Preoperative evalu-ation of living renal donors: comparison of CT angiography andMR angiography. Radiology 216: 434–439, 2000.

7. Rajab A, Khabiri H, Pelletier RP, et al: Magnetic resonance an-giography for preoperative evaluation of potential kidney donors.J Surg Res 120: 195–200, 2004.

8. Pozniak MA, Balison DJ, Lee FT Jr, et al: CT angiography of potentialrenal transplant donors. Radiographics 18: 565–587, 1998.

9. Steinborn M, Wintersperger BJ, Heck A, et al: Contrast enhancedMR angiography in the preoperative evaluation of living kidneydonors. Rofo Fortschr Geb Rontgenstr Neuen Bildgeb Verfahr 171:313–318, 1999.

0. Bhatti AA, Chugtai A, Haslam P, et al: Prospective study compar-ing three-dimensional computed tomography and magnetic reso-nance imaging for evaluating the renal vascular anatomy in poten-tial living renal donors. BJU Int 96: 1105–1108, 2005.

1. Zhang H, and Prince MR: Renal MR angiography. Magn ResonImaging Clin N Am 12: 487–503, 2004.

2. Rankin SC, Jan W, and Koffman CG: Noninvasive imaging of livingrelated kidney donors: evaluation with CT angiography and gadolinium-enhanced MR angiography. Am J Roentgenol 177: 349–355, 2001.

3. Cragg AH, Smith TP, Thompson BH, et al: Incidental fibromus-cular dysplasia in potential renal donors: long-term clinical follow-up. Radiology 172: 145–147, 1989.

4. Neymark E, LaBerge JM, Hirose R, et al: Arteriographic detectionof renovascular disease in potential renal donors: incidence andeffect on donor surgery. Radiology 214: 755–760, 2000.

5. Schreiber MJ, Pohl MA, and Novick AC: The natural history ofatherosclerotic and fibrous renal artery disease. Urol Clin North

Am 11: 383–392, 1999.

UROLOGY 71 (3), 2008

2

2

2

2

3

U

6. Indudhara R, Kenney, Bueschen AJ, et al: Live donor nephrectomyin patients with fibromuscular dysplasia of the renal arteries. J Urol162: 678–681, 1999.

7. Andreoni KA, Weeks SM, Gerber DA, et al: Incidence of donorrenal fibromuscular dysplasia: does it justify routine angiography?Transplantation 73: 1112–1116, 2002.

8. Bakker J, Ligtenberg G, Beek FJ, et al: Preoperative evaluation ofliving renal donors with gadolinium-enhanced magnetic resonance

angiography. Transplantation 67: 1167–1172, 1995.

ROLOGY 71 (3), 2008

9. Vasbinder GB, Nelemans PJ, Kessels AG, et al: Renal ArteryDiagnostic Imaging Study in Hypertension (RADISH) StudyGroup: accuracy of computed tomographic angiography and mag-netic resonance angiography for diagnosing renal artery stenosis.Ann Intern Med 141: 674–682, 2004.

0. Dellegrottaglie S, Sanz J, and Rajagopalan S: Technology insight:clinical role of magnetic resonance angiography in the diagnosisand management of renal artery stenosis. Nat Clin Pract Cardio-

vasc Med 3: 329–338, 2006.

389