MAGNETIC RESONANCE IMAGING DETECTION OF VASCULAROCCLUSION OF A PEDICLED MUSCLE FLAP

KENNETH HUI, M.D.,1 GORDON K. LEE, B.S., 1

FENG ZHANG, M.D., Ph.D.,1 KING LI, M.D.,2

LUKE CHEUNG, M.D.,2 and WILLIAM C. LINEAWEAVER, M.D. 1*

Contrast-enhanced magnetic resonance imaging (MRI) canbe a highly sensitive monitor of tissue blood perfusion. Thistechnique has been used to assess blood flow through liver,kidney, and certain tumors, but has not been widely appliedto the study of skeletal muscle circulation. In our study, weused a novel scanning software to obtain contrast-enhancedT2*-weighted gradient echo MRI images of pedicled quadri-ceps muscle flaps in rabbits in order to study images of ar-terial, venous, and arterio-venous occlusion. We adminis-tered an intravenous bolus of gadoteridol contrast agent atthe initiation of scanning, which produces a decrease in T2*-signal and improves the sensitivity of measuring blood per-fusion. Within 30 seconds of MRI scanning, control flaps withintact pedicles exhibited a rapid decrease in T2*-signal in-tensity, indicating adequate perfusion of blood throughmuscle tissue; however, occluded pedicled flaps showed no

significant change in signal intensity, indicating lack of bloodperfusion. Differences in signal intensities as measured byMRI between occluded and control flaps were statisticallysignificant (P < 0.05). Selective vascular occlusion of eitherartery alone, or both artery and vein were detected within 15minutes, whereas selective venous occlusion could be de-tected after 2 hours. We conclude that MRI has the ability toassess skeletal muscle perfusion, and is capable of nonin-vasively evaluating a cross-section of tissue in both superfi-cial and buried flaps. MRI, therefore, may have the potentialfor evaluating perfusion in muscle flaps (including buriedflaps), and other disorders of muscle circulation such as com-partment syndrome.

© 1997 Wiley-Liss, Inc.MICROSURGERY 17:306–312 1996

Microsurgical flap transfers are increasingly applied tocomplex defects as more surgeons become trained in suchprocedures.1 Flaps composed of skin, muscle, bone, viscera,and combinations of several tissues are used for reconstruc-tion of defects that previously required multiple proce-dures.2 However, even with current techniques, 10–15% offlaps develop complications requiring revision of the vas-cular anastomosis and 5–10% of flaps fail.3,4 Monitoringcirculation of microsurgical flaps has been repeatedly ex-amined for efficacy in early detection of vascular compli-cations in the hope that prompt detection and treatment ofsuch complications can avoid the tissue degeneration due toprolonged secondary ischemia and reperfusion cycles lead-ing to flap death.5–7

Thus far, no single method of flap circulation monitor-

ing has been described as reliable and ideal. The most com-monly used method of monitoring has been mere clinicalexamination, which employs observation of tissue color,capillary refill, and bleeding tests.8 Such methods are highlysubjective and dependent upon the experience of the ob-server. Furthermore, buried flaps, such as those used for softtissue augmentation, cannot be examined visually. Otherapproaches for monitoring include photoplethysmography,9

impedance plethysmography,10 pulse oximetry,11 pH mea-surement,12 dermofluorometry,7 oxygen tension,13 surfacethermometry,14–18 Laser Doppler,16,19–25 transcutaneouspO2/pCO2,20 and proton nuclear magnetic resonance(NMR) spectroscopy.26 Implantable Doppler probes havebeen devised to monitor buried flaps.27,28 Most of thesemethods have technical or validity problems restricting theirwide adaptation, and few of them (the buried Doppler probenotably excepted) can examine buried flaps.5,6,27–31

In our study, we examined the potential use of magneticresonance imaging (MRI) to assess skeletal muscle circula-tion. MRI is a relatively new, noninvasive diagnostic tool inwhich a tomographic cross-sectional image of tissue can bevisualized and analyzed. The ability of MRI to image deeporgans and tissues makes it ideal to assess buried flaps. MRIcan be combined with intravenous administration of con-trast agent, which further increases the resolution of tissue

From the 1Division of Plastic and Reconstructive Surgery, Stanford UniversitySchool of Medicine, Stanford, California, and 2Department of Radiology, Vet-erans Administration Hospital, Stanford University School of Medicine, PaloAlto, California.

Presented at the Surgical Forum, American College of Surgeons, New Or-leans, Louisiana, October 24, 1995.

*Correspondence to: William C. Lineaweaver, Division of Plastic & Recon-structive Surgery, NC104, Stanford University School of Medicine, Stanford,CA 94305.

Received for publication 4 February 1997; Accepted 11 February 1997

© 1997 Wiley-Liss, Inc.

detail and sensitivity to tissue circulation. MRI has beenused to measure blood flow through organs, such as the liverand kidney where circulation is relatively large per tissueweight. Other investigators have used MRI to detect vascu-lar perfusion in myocardium and renal parenchyma,3 andintramuscular sarcomas.32 However, obtaining quantifiableresults in these studies has relied on complex calculationsand mathematical models that may not accurately reflect thephysiologic state.33–35

We used MRI to examine the circulation of a pedicledquadriceps muscle flap in a rabbit model, and determinedwhether or not we could identify selective occlusionsof either artery, vein, or both artery and vein of the vas-cular pedicle. We elevated bilateral flaps so that, in a sin-gle operative setting, one flap underwent selective vas-cular occlusion and the contralateral flap acted as anunoccluded control. We utilized a novel scanning se-quence developed in our institution to obtain MRI images.36

Simultaneous administration of an intravenous con-trast agent, gadoteridol, allowed for intravascular perfu-sion signal enhancement on T2*-weighted images with-in muscle tissue itself. Using this procedure for obtainingenhanced MRI images, we investigated our ability to de-tect vascular occlusions in these experimental muscleflaps.

MATERIALS AND METHODS

Pedicled Flap Model

New Zealand white rabbits (3–5 kg in weight) wererandomly divided into three separate groups for arterial (n4 3), venous (n4 4), and arterio-venous (n4 5) occlu-sion. Animals were intubated and maintained under generalanesthesia with 2% halothane. Axial incisions were madeanteriorly to expose the quadriceps muscle layer. The neu-rovascular supply for this muscle flap, namely the femoralartery, vein, and nerve were identified and preserved. Theinsertions and origins of the entire quadriceps muscles, aswell as collateral vessels, were completely divided (Fig. 1).

One of the bilateral muscle flaps was chosen at randomto undergo selective ligation with 3-O silk suture of eitherthe femoral artery, vein, or both artery and vein of thevascular pedicle. The contralateral flap, in which the vas-cular pedicle was intact, was used as the unoccluded con-trol. Arterial and arterio-venous occlusion groups werescanned within 15 minutes of vascular occlusion. Based onpreliminary trials, acute occlusion of the vein did not dem-onstrate marked signal differences that could be reliablydetected by MRI. Therefore, the venous occlusion groupunderwent selective occlusion of their vein for 2 hours priorto MRI scanning.

Figure 1. Pedicled quadriceps muscle flap in the rabbit. The vascular pedicle was identified and preserved. Selective ligations of either theartery, vein, or both artery and vein were carried out using 3-O silk ties in occluded flaps. Contralateral flaps with intact vascular pedicle wereunoccluded controls.

MRI Monitoring of Muscle Flap Perfusion 307

308 Hui et al.

MRI

Animals were placed supine in a 1.5T GE Signa wholebody MRI scanner with a 5-inch round surface coil placedon the animal’s ventral surface. At the start of the scanningsequence gadoteridol (ProHance) contrast agent at 0.3mmol/kg weight was administered as an intravenous bolus,and then followed by an 8-ml saline flush. The scanningsoftware was developed in the Stanford Department of Ra-diology, and utilizes a T2*-weighted gradient echo se-quence with spiral readout, employing the following param-eters: repetition time (TR)4 40 ms, echo time (TE)4 75ms, flip angle4 15 degrees, 10 interleaves, NEX4 2.36

Axial tomographic images of muscle flaps were obtainedevery 2 seconds for a total of 2 minutes. At the completionof the experiment all animals were sacrificed. Experimen-tation on animals in this study were in accordance with theUnited States National Research Council’s guide for thecare and use of laboratory animals.

Data Analysis

Cross-sectional images of muscle flaps were encircledby the aid of computer software, and T2*-signal intensitieswere obtained and plotted over time. Peak penetration ofgadoteridol into perfused muscle tissue produces a maximaldecrease in T2*-weighted signal denoted as the first-passrate of change, which was calculated as percent T2*-weighted signal change per second. The first-pass rates ofT2*-signal change for control and occluded flaps withineach group were averaged, and 95% confidence intervalswere computed.

RESULTS

Representative contrast-enhanced T2*-weighted imagesof typical arterial occlusion findings are shown in Figure 2.An axial image obtained prior to scanning shows the bilat-eral pedicle quadriceps muscle flaps (Fig. 2, arrows). Theleft muscle flap underwent selective arterial occlusion. Theright muscle flap with the vascular pedicle completely intactwas the unoccluded internal control. The control flapshowed an observable decrease in signal intensity within 24seconds of scanning. The arterial occluded flap showed noapparent change in signal intensity. MRI images for venousand arterio-venous occlusion were comparable to arterialocclusion, showing a decrease in signal intensity of controlflaps and no significant change in occluded flaps (data not

shown). Detecting gross changes in signal intensities of themuscle flaps was not always obvious with mere visual in-spection of MRI images. However, by using computer soft-ware to specifically encircle cross-sections of muscle flapson MRI images, we obtained more accurate, quantitativesignal intensity measurements for control and occludedflaps over the entire period of scanning.

We normalized percent T2*-weighted signal intensitiesof pedicled flaps, and plotted them for representative ani-mals from arterial, venous, and arterio-venous occludedgroups (Fig. 3A–C). Within 30 seconds of scanning, controlflaps showed a decrease in signal intensity, which corre-sponded to the first-pass of gadoteridol penetration into ad-equately perfused muscle tissue. There were no differencesbetween control flaps among the three groups. Arterial, ve-nous, and arterio-venous occluded flaps showed no signifi-cant change in signal intensity from the baseline, indicatingthe lack of adequate circulation. Therefore, occluded muscleflaps could be identified from control flaps. In our prelimi-nary trials, flaps with venous occlusion showed no signifi-cant differences from control flaps when occlusion had oc-curred within 15 minutes (data not shown). Only when oc-clusion had occurred for more than 2 hours did we detectmeasurable differences in T2*-signal intensities betweenvenous occluded and control flaps.

Flaps with arterial occlusion had a mean first-pass rateof change in T2*-weighted signal intensity of 1.14 ± 0.57%/sec (n4 3), venous occluded 1.02 ± 0.47%/sec (n4 4),and arterio-venous occluded 0.61 ± 0.24%/sec (n4 5).Control flaps had an average first-pass rate of change inT2*-weighted signal intensity of 2.10 ± 0.25%/sec (n412). Statistical significance (P < 0.05) was achieved in de-tecting arterial (P 4 0.02), venous (P 4 0.01), and arterio-venous (P 4 0.04) occluded flaps from control flaps. How-ever, based on computed 95% confidence intervals for first-pass rates of change, we were unable to identify specificallyand differentiate between arterial, venous, and arterio-venous occlusions (Fig. 4).

DISCUSSION

With the use of rapid-scanning software, as well as con-trast agent enhancement, we have demonstrated in our studythe utility of MRI to evaluate skeletal muscle circulation.Our data show that statistical significance was achieved indetecting occluded flaps from control flaps by contrast-

Figure 2. Contrast-enhanced T2*-weighted MRI images of typicalarterial occlusion findings: (A) axial image obtained prior to scanshowingbilateral pedicle muscle flaps. The left muscle flap (shortarrow) has its arterial supply selectively occluded, whereas the rightmuscle flap (long arrow) with an intact vascular pedicle was theunoccluded control; (B) 6 seconds after initiation of scan and intra-venous injection of gadoteridol; and (C) 24 seconds after initiation of

scan showing decrease in T2*-weighted signal in unoccluded controlflap. No observable difference could be seen in the occluded left flap.Gross signal intensity changes were not consistently obvious to de-tect by mere visual inspection. Computer-aided analysis of the cross-section of muscle flap was required to quantitatively demonstratesignal intensity differences between flaps.

MRI Monitoring of Muscle Flap Perfusion 309

enhanced MRI in all three groups of arterial, venous, andarterio-venous occlusion. Occluded flaps demonstrated nochange in signal intensities after intravenous injection ofcontrast agent, which was consistent with lack of flap cir-culation. Control flaps with intact vascular pedicles allowedgadoteridol to penetrate into muscle tissue, and hence, de-crease in T2*-weighted signal intensities, which indicatedadequate flap circulation.

We found that while arterial and arterio-venous occlu-sion could be detected within 15 minutes, venous occlusionshowed statistically significant differences only when oc-clusion had exceeded 2 hours. We believe that persistentarterial inflow after acute venous occlusion did not allow usto observe differences between venous occluded and controlflaps. Occlusion of the vein for more than 2 hours, however,caused venous congestion to become great enough to deterarterial inflow, reduce contrast agent penetration, and at thatpoint, allowed us to distinguish it from the control flap.Furthermore, based on our data, we did not measure statis-tically significant differences between first-pass rates ofchange of arterial occlusion, venous occlusion, and arterio-venous occlusion, and therefore, did not give data differen-tiating these types of occlusion.

In the setting of free flap monitoring, thrombosis of themicroanastomosis occurs more frequently in the vein than inthe artery.5,6,30 It is believed that the low pressure venoussystem makes it prone to thrombus formation, although nosolid data have been produced to substantiate this claim. Anideal method for postoperative monitoring of tissue flapsshould be highly sensitive to venous occlusion. In our study,we were unable to detect venous occlusion in the acuteperiod, although with prolonged occlusion (greater than 2hours), MRI successfully measured signal intensity differ-ences to identify occluded from unoccluded control flaps. Itshould be noted that during acute venous compromise thereis a persistent amount of arterial inflow that may be suffi-cient to sustain the muscle tissue. Furthermore, to whatextent compromise of venous outflow warrants surgical re-exploration is not clear. We have shown that MRI detectsvenous occlusion only after arterial inflow becomes signifi-cantly reduced, and consequently, the resultant ischemia bydefinition would be a clear indication to perform reex-ploration and vascular anastomosis revision. With furtheradvancements in MRI technology we expect greater reso-lution, increased sensitivity, and earlier detection of vascu-lar occlusion.

There have been two studies using MRI to evaluateischemic muscle. These studies utilized conventional soft-

Figure 3. Normalized T2*-weighted signal intensities of control andoccluded muscle flaps from representative animals having under-went: (A) arterial occlusion; (B) venous occlusion; and (C) arterio-venous occlusion. In each case, control flaps show a rapid decreasein T2*-signal. Occluded flaps show no significant change in signalcompared to the baseline.

310 Hui et al.

ware to obtain nonenhanced MRI images. Both of thesereports claimed to have observed an increase in T2-signalsin occluded flaps compared to undisturbed muscle tissueafter approximately 35 minutes.37,38 However, a more re-cent study by Varnell et al.39 showed that in nonenhancedMRI there is nonspecific T2-signal increase of both oc-cluded and control flaps. They argued that an increase inT2-signal can be caused by postoperative edema and hem-orrhage, and is not specific for vascular occlusion. In addi-tion, Varnell et al.39 used standard MRI contrast agent,gadopentetate (Magnevist), to study flaps with occlusion ofthe entire vascular pedicle. They were able to observe maxi-mal differences between occluded and control flaps 30 min-utes after administration of gadopentetate.39 We used gado-pentetate in our preliminary trials, but found it to be inad-equate in our protocol, which combines MRI withsimultaneous injection of contrast agent. By using gadoteri-dol we achieved greater resolution and a more rapid assess-ment of muscle tissue perfusion.

The utility of gadolinium-based MRI contrast agents haspreviously been limited to detecting abnormalities of theblood-brain barrier and malignancies.40,41 However, theseagents have been employed to obtain measurements of tis-sue perfusion and blood volume, but not in skeletal muscle.The standard contrast agent gadopentetate (Magnevist) is anionic compound composed of a gadolinium ion chelatedwith a linear molecule (diethylenetriaminepentaacetic acid),whereas gadoteridol (ProHance) is composed of a gadolin-ium ion chelated with a macrocyclic molecule (10-[2-hydroxypropyl]-1,4,7,10-tetraazacyclododecane-1,4,7 tri-acetic acid).42 Gadoteridol, like gadopentetate, is a para-

magnetic agent that enhances the relaxation rates of waterprotons in the vicinity of the agent and produces a decreasein the T2*-weighted signal. However, the nonionic, lowviscosity nature of gadoteridol allows it to be given rela-tively safely and rapidly at three times the concentration ofgadopentetate.43 Furthermore, gadoteridol’s macrocyclicstructure reduces the rate of dissociation and increases itsstability in the intravascular space.44 These physicochemi-cal properties of gadoteridol make it an ideal agent to beused to evaluate skeletal muscle circulation and detect vas-cular compromise.

The use of gadolinium-enhanced MRI in assessingmuscle circulation has other potential clinical applications,such as assisting in the diagnosis of compartment syndrome.Compartment syndrome is characterized by severe com-pression of nerve and muscle tissue within a compartmentalor osseofascial space, and commonly occurs after trauma tothe upper or lower extremities. Sequelae of compartmentsyndrome include permanent hypesthesia, contractures, sep-sis, limb loss, and death. Peripheral pulses are often present,and limb edema with pain may be the only physical find-ing.45 Clinical diagnosis may be difficult, and a timely de-cision must be made by the physician as to when surgicaldecompression and fasciotomy are necessary. Intracompart-mental pressure monitoring with needle technique, wickcatheter, and slit catheters have been used, with significantpressures considered to be greater than 30 mm Hg. How-ever, these methods may not be accurate and tend to bedifficult to apply.46 We foresee the utility of contrast-enhanced MRI, which would allow for a noninvasive, ac-curate assessment of skeletal muscle circulation, and assistin the early diagnosis and management of compartment syn-drome.

In summary, MRI may become a valuable tool in theassessment of skeletal muscle circulation. MRI allowed usto visualize and analyze a large cross-sectional area of themuscle flap tissue, which was accurate in detecting vascularcompromise. MRI monitoring of skeletal muscle tissue maybe used in postoperative pedicled flap and free-tissue trans-fer monitoring, as well as in the early diagnosis of compart-ment syndrome, and any setting in which muscle perfusionneeds to be evaluated. At this time, MRI is a relativelyexpensive imaging method, but as technology improves, theuse of MRI as a monitor may become more practical andless costly.

ACKNOWLEDGMENTS

The authors acknowledge Pei-Ran Ho, Curtis Song,Rosemary Broome, and Jenny Woodside for their technicalassistance in the preparation of this study.

Figure 4. Mean first-pass rates for contrast-enhanced images of ar-terial (1.14 ± 0.57%/sec), venous (1.02 ± 0.47%/sec), and arterio-venous (0.61 ± 0.24%/sec) occluded and control flaps (2.10 ±0.25%/sec). MRI achieved statistical significance (P < 0.05) in iden-tifying arterial (P = 0.021), venous (P = 0.008), and arterio-venous (P= 0.041) occluded flaps from control flaps.

MRI Monitoring of Muscle Flap Perfusion 311

REFERENCES

1. Khouri RK: Free flap surgery: The second decade.Clin Plast Surg19:757–761, 1992.

2. Hallock GG: ‘‘Microleaps’’ in the progression of flaps and grafts.ClinPlast Surg23:117–138, 1996.

3. Oliva A, Lineaweaver WC, Buncke HJ, Buncke GM, Siko P, JacksonRL, Samaha FJ, Alpert BS: Salvage of wounds following failed tissuetransplantation.J Reconstr Microsurg9:257–263, 1993.

4. Yajima H, Tamai S, Mizumoto S, Ono H, Fukui A: Vascular compli-cations of vascularized composite tissue transfer: Outcome and salvagetechniques.Microsurgery14:473–478, 1993.

5. Furnas H, Rosen JM: Monitoring in microvascular surgery.Ann PlastSurg26:265–272, 1991.

6. Machens HG, Mailaender W, Rieck B, Berger A: Techniques of post-operative blood flow monitoring after free tissue transfer An overview.Microsurgery15:778–786, 1994.

7. Whitney TM, Lineaweaver WC, Billys JB, Siko PP, Buncke GM,Alpert BS, Oliva A, Buncke HJ: Improved salvage of complicatedmicrovascular transplants monitored with quantitative fluorometry.Plast Reconstr Surg90:105–111, 1992.

8. Buncke HJ, Lineaweaver WC, Valauri F, Bunche GH: Monitoring.Chap. 37. In Buncke H (ed.):Microsurgery: Transplantation-Replantation.Philadelphia, Lea & Febiger, 1991, pp. 715–721.

9. Jones JW, Wiebalck R: Continuous postoperative free-flap monitoringwith an EKG-interfaced photoplethysmograph.J Reconstr Microsurg8:61–62, 1992.

10. Concannon MJ, Stewart DH, Welsh CF, Puckett CL: Impedance pleth-ysmography: A new method for continuous muscle perfusion moni-toring. Plast Reconstr Surg88:292–298, 1991.

11. Lindsey LA, Watson JD, Quaba AA: Pulse oximetry in postoperativemonitoring of free muscle flaps.Br J Plast Surg44:27–29, 1991.

12. Dunn RM, Kaplan IB, Mancoll J, Terzis JK, Trengove-Jones G: Ex-perimental and clinical use of pH monitoring of free tissue transfers.Ann Plast Surg31:539–545, 1993.

13. Hofer SO, Timmenga EJ, Christiano R, Bos KE: An intravascularoxygen tension monitoring device used in myocutaneous transplants:A preliminary report.Microsurgery14:304–311, 1993.

14. Delhomme G, Newman WH, Roussel B, Jouvet M, Bowman HF,Dittmar A: Thermal diffusion probe and instrument system for tissueblood flow measurements: Validation in phantoms and in vivo organs.IEEE Trans Biomed Eng41:656–662, 1994.

15. Khouri RK, Shaw WW: Monitoring of free flaps with surface-temperature recordings: is it reliable?Plast Reconstr Surg89:495–502,1992.

16. Nitzan M: Simultaneous measurement of skin blood flow by the tran-sient thermal-clearance method and laser doppler flowmetry.Med BiolEng Comput26:407–410, 1988.

17. Shichino K, Yukimura T, Miura K, Yamamoto K, Okubo M, ShimazuA: Regional blood flow in skeletal muscle measured by the heatedthermocouple method during electrical nerve stimulation of the caninegracilis muscle.Tohoku J Exp Med164:145–156, 1991.

18. Sudarsky LA, Salomon J: Liquid crystal temperature monitoring formicrosurgery.J Reconstr Microsurg7:13–14, 1991.

19. Clinton MS, Sepka RS, Bristol D, Pederson WC, Barwick WJ, SerafinD, Klitzman B: Establishment of normal ranges of laser doppler bloodflow in autologous tissue transplants.Plast Reconstr Surg87:299–309,1991.

20. Franzeck UK, Stengele B, Panradi U, Wahl P, Tillmanns H: Cutaneousreactive hyperemia in short-term and long-term type I diabetes—continuous monitoring by a combined laser doppler and transcutane-ous oxygen probe.VASABand 19; Heft 1:8–15, 1990.

21. Hallock GG: Critical threshold for tissue viability as determined bylaser doppler flowmetry.Ann Plast Surg28:554–558, 1992.

22. Hallock GG, Altobelli JA: Assessment of TRAM flap perfusion usinglaser doppler flowmetry, an adjunct to microvascular augmentation.Ann Plast Surg29:122–127, 1992.

23. Hallock GG, Koch TJ: External monitoring of vascularized jejunum

transfers using laser doppler flowmetry.Ann Plast Surg24:213–215,1990.

24. Schliephake H, Schmelzeisen R, Neukam FW: Long-term results ofblood flow and cutaneous sensibility of flaps used for the reconstruc-tion of facial soft tissues.J Oral Maxillofac Surg52:1247–1252, 1994.

25. Svensson H, Holmberg J, Svedman P: Interpreting laser doppler re-cordings from free flaps.Scandinavian J Plast Reconstr Surg27:81–87, 1993.

26. Cuono CB, Armitage IM, Marquetand R, Chapo GA: Nuclear mag-netic resonance spectroscopy of skin. Predictive correlates for clinicalapplication.Plast Reconstr Surg81:1–11, 1988.

27. Swartz WM, Izquierdo R, Miller MJ: Implantable venous dopplermicrovascular monitoring: Laboratory investigation and clinical re-sults.Plast Reconstr Surg93:152–163, 1994.

28. Fernando B, Young VL, Logan SE: Miniature implantable laser dopp-ler probe monitoring of free tissue transfer.Ann Plast Surg20:434–442, 1988.

29. Chang BW, Rumley TO: Monitoring a free flap.Plast Reconstr Surg87:581, 1991.

30. Jones NF: Intraoperative and postoperative monitoring of microsurgi-cal free tissue transfers.Clin Plast Surg19:783–797, 1992.

31. Neligan PC: Monitoring techniques for the detection of flow failure inthe postoperative period.Microsurgery14:162–164, 1993.

32. Belfi CA, Paul CR, Shan S, Ngo FQ: Comparison of the effects ofhydralazine on tumor and normal tissue blood perfusion by MRI.Int JRadiat Oncol Biol Phys29:473–479, 1994.

33. Chenevert TL, Pipe JG, Williams DM, Brunberg JA: Quantitativemeasurement of tissue perfusion and diffusion in vivo.Magn ResonMed 17:197–212, 1991.

34. Morvan D, Leroy-Willig A, Malgouyres A, Cuenod CA, Jehenson P,Syrota A: Simultaneous temperature and regional blood volume mea-surements in human muscle using an MRI fast diffusion technique.Magn Reson Med29:371–377, 1993.

35. Su MY, Jao JC, Nalcioglu O: Measurement of vascular volume frac-tion and blood-tissue permeability constants with a pharmacokineticmodel. Studies in rat muscle tumors with dynamic Gd-DTPA en-hanced MRI.Magn Reson Med32:714–724, 1994.

36. Spielman DM, Pauly JM, Meyer CH: Magnetic resonance fluoroscopyusing spirals with variable sampling densities.Magn Reson Med34:388–394, 1995.

37. Elias DL, Nelson RC, Herbst MD, Zubowicz VN: Magnetic resonanceimaging for detection of arterial and venous occlusion in caninemuscle flaps and bowel segments.Ann Surg206:624–627, 1987.

38. Greenberg B, Mezrich R, Prymak C, Kressel H, LaRossa D: Applica-tion of magnetic resonance imaging technique in determining caninemuscle and human free-flap viability.Plast Reconstr Surg79:959–965, 1987.

39. Varnell RM, Flint PW, Dalley RW, Maravilla KR, Cummings CW,Shuman WP: Myocutaneous flap failure: Early detection with Gd-DTPA-enhanced MR imaging.Radiology173:755–758, 1989.

40. Yuh WT, Fisher DJ, Mayr-Yuh NA, Tali ET, Nguyen HD, Gao F,Ehrhardt JC: Review of the use of high-dose gadoteridol in the mag-netic resonance evaluation of central nervous system tumors.InvestRadiol Suppl 1:S39–44, 1992.

41. Runge VM, Kirsch JE, Wells JW, Dunworth JN, Awh MH, BittnerDF: Magnetic resonance contrast agents in neuroimaging: New agentsand applications.Neuroimaging Clin N Am4:175–183, 1994.

42. Tweedle MF: Physicochemical properties of gadoteridol and othermagnetic resonance contrast agents.Invest Radiol27(Suppl 2):S2–6,1992.

43. Olukotun AY, Parker JR, Meeks MJ, Lucas MA, Fowler DR, LucasTR: Safety of gadoteridol injection: U.S. clinical trial experience.JMagn Reson Imaging5:17–25, 1995.

44. Tweedle MF, Wedeking P, Kumar K: Biodistribution of radiolabeled,formulated gadopentate, gadoteridol, gadoterate, and gadodiamide inmice and rats.Invest Radiol30:372–380, 1995.

45. Moore RE, Friedman RJ: Current concepts in pathophysiology anddiagnosis of compartment syndromes.J Emerg Med7:652–657, 1989.

46. Mabee JR: Compartment syndrome: A complication of acute extremitytrauma.J Emerg Med12:651–656, 1994.

312 Hui et al.