Vascular Imaging
Original Research
MR Angiography of Lower Extremities at 3 T: Presurgical Planning of Fibular Free Flap Transfer for Facial Reconstruction
Audio Available | 
Share
OBJECTIVE. The purpose of our study was to evaluate the use of preoperative MR angiography of the lower extremities at 3 T in candidates for fibular free flap harvesting, identifying atherosclerotic occlusive disease and congenital anomalies in this population. Our intention was to document the influence of the imaging findings on the surgical approach used.
MATERIALS AND METHODS. Twenty-nine consecutive adult patients with facial abnormalities necessitating mandibular resection with subsequent osteocutaneous mandibular reconstruction who underwent preoperative MR angiography at 3 T were retrospectively reviewed. Images were evaluated by two observers with regard to image quality and visualization of arterial segments; severity of stenosis; and presence of noise, artifact, or venous contamination. The popliteal artery branching pattern present was also classified. The facial and reconstructive surgeon involved indicated whether the MR angiographic appearances influenced the decision regarding the side or location from which the flap was harvested or the flap design.
RESULTS. Arterial segments were visualized with good or excellent image quality in 722 of 725 segments for observer 1 and 721 segments for observer 2. The kappa coefficient indicated good interobserver agreement (κ = 0.78) with regard to quality of arterial segment depiction and scoring of stenoocclusive disease (κ = 0.64). No segments had venous contamination, noise, or artifact of a degree sufficient to compromise diagnostic interpretation. Imaging influenced the surgical approach in 16 (55.2%) of 29 patients.
CONCLUSION. Trifurcation vessel imaging should be a prerequisite to fibular free flap harvesting. High-spatial-resolution MR angiography at 3 T represents a desirable alternative to other invasive or cross-sectional imaging techniques in this regard.
Keywords: fibula, free flap, mandible, oral cancer, peroneal artery, reconstruction
| Introduction |   |
|---|
Congenital, traumatic, or oncologic afflictions of the head and neck region may be difficult to treat and are associated with considerable psychologic stress on the patient affected. In the region of the mandible, even small defects may have considerable effects on respiration, articulation, deglutition, mastication, and esthetic appearance [1]. Recent advances in microvascular surgical techniques have resulted in the emergence of free flap reconstruction as the standard of care for immediate, single-stage restoration of mandibular esthetic and functional integrity. This approach has been shown to provide durable results that are superior to those of pedicled flaps and nonvascularized bone grafts, with long-term viability exceeding a decade [2, 3]. Although a number of potential donor sites for such flaps are available, the fibular free flap has been shown to be particularly versatile and has been adopted as the dominant free flap for all mandibular reconstructions [4].
The use of preoperative lower extremity imaging before fibular free flap transfer has been advocated [5] to identify significant atherosclerotic occlusive disease or congenital anomalies of the trifurcation vessels. In a survey of 206 vascular surgeons in the United Kingdom, 88% of respondents indicated that not performing presurgical vascular imaging bordered on clinical negligence [6]. The purpose of imaging is to identify patients in whom the peroneal artery represents a significant contributory vessel to the foot, either congenitally or as a collateral vessel because of tibial arterial stenosis or occlusion, to avoid postsurgical ischemia. The viability of such flaps is entirely dependent on the integrity of their vascular supply, making the detection of peroneal disease of paramount importance [7].
Recent advances in MR angiography (MRA) techniques, using rapid three-axis gradient coils at high field strengths, have pushed the boundaries of spatial resolution and coverage, while avoiding ionizing radiation exposure and femoral arterial puncture. Furthermore, the promise of 3-T peripheral MRA in providing improved imaging quality when compared with imaging at 1.5 T has recently been shown [8]. The aim of this study was to evaluate the utility of preoperative MRA of the lower extremities at 3 T in candidates for fibular free flap harvesting. We sought to identify atherosclerotic occlusive disease and congenital anomalies in this patient population and to document the influence of the imaging findings on the surgical approach. To the best of our knowledge, such preoperative assessment has not previously been reported at 3 T.
| Materials and Methods | |
|---|
All examinations were HIPAA-compliant. Institutional review board approval was obtained and all study participants provided written informed consent. This retrospective review consisted of 29 consecutive adult patients (18 men and 11 women; mean age, 61.6 years; age range, 22–88 years) with abnormalities that necessitated segmental mandibular resection (orofacial squamous cell carcinoma in 22 patients and adenoid cystic carcinoma, mucoepidermoid carcinoma, submandibular adenocarcinoma, mandibular osteomyelitis, osteosarcoma, brown tumor, and radiation-induced osteo necrosis in one patient each), with subsequent osteo cutaneous mandibular reconstruction, as determined by a board-certified microvascular surgeon. Exclusion criteria were contraindications to MRI including implanted cardiac pacemaker, claustrophobia, or known allergy to gadolinium-based contrast agents. Severe renal impairment was also considered a relative contraindication given recent concerns regarding its potential link to nephrogenic systemic fibrosis and dermopathy. To this end, cutoff creatinine and glomerular filtration rates (GFRs) of 2.5 mg/dL and < 30%, respectively, were used.
All examinations were performed on a 3-T whole-body MR scanner (Magnetom Trio, Siemens Medical Solutions) equipped with 32 independent receiver channels and a multiarray coil design facilitating large field-of-view parallel imaging. This system has a three-axis gradient system providing a peak gradient amplitude of 45 mT/m and maximum slew rate of 200 mT/m/ms. The MRI table has a movement range of 185 cm and a maximum speed of 20 cm/s.
Patient positioning in a supine, feet-first orientation on the imaging table was followed by attachment of multiple phased-array surface coils covering the lower chest, abdomen, pelvis, and bilateral lower extremities. Six coil elements were positioned for abdominal and pelvic signal reception, six elements for the thighs, and two sets of phased-array coils for calf and foot coverage (total of 12 elements). The MRI system also contained a 24-element spine coil embedded within the patient table. A combination of surface phased-array coils and spine coil elements was used for data acquisition at each anatomic station; the number of activated coil elements for each station was 24 for the abdominal and pelvic station (station 2), 24 for the thigh station (station 3), and 21 for calf and foot data acquisition (station 1). Each station was imaged with a 500-mm field of view and 50-mm overlap, resulting in overall craniocaudal coverage of 1,400 mm.
Initial localizer images were acquired with multiplanar steady-state free precession (SSFP). A divided injection protocol was used consisting of initial high-resolution contrast-enhanced MRA of the calves (station 1) [9] and subsequent two-station abdominopelvic and thigh imaging (stations 2 and 3). Separate timing measurements to the calves and to the abdominal aorta were performed, each using 1.5 mL of gadopentetate dimeglumine (Magnevist, Bayer HealthCare) injected at 1.2 mL/s and flushed with 30 mL of saline at the same flow rate. The timing sequence was an ultrafast 3D spoiled gradient echo (SGE) with a temporal resolution of 1 second. All contrast material and saline injections were administered using a single automated power injector (Spectris, Medrad).
High-spatial-resolution contrast-enhanced MRA was performed using a 3D spoiled gradient-recalled echo sequence with an asymmetric k-space sampling scheme (partial Fourier factor, 80%) and zero interpolation in all three (x, y, and z) axes. Parallel imaging used a generalized autocalibrating partially parallel acquisition (GRAPPA) algorithm [10, 11]. The GRAPPA acceleration factors used were three for stations 2 and 3 and four for the calf region (station 1), with 24 reference k-space lines for calibration in the left-to-right phase-encoding direction. As illustrated in Table 1, these settings enabled acquisition of near-isotropic high-spatial-resolution 3D voxel imaging for each station.
For MRA, a total gadopentetate dose of 0.15 mmol per kilogram of body weight was infused during the two injections, each at a rate of 1.2 mL/s and flushed with 30 mL of saline at the same injection rate. Average table time per patient was 30 minutes.
Images acquired were independently and blindly evaluated by two experienced radiologists with 10 and 5 years of experience. In all, 25 individual arterial segments were evaluated and scored by each observer for each patient (abdominal aorta; bilateral common iliac arteries; external iliac arteries; common femoral arteries; superficial femoral arteries; profunda femoris arteries; popliteal arteries; tibioperoneal trunks; peroneal arteries; anterior, posterior tibial, and dorsalis pedis arteries; and plantar arches), yielding a total of 725 segments scored.
Image quality in terms of visualization of each arterial segment was evaluated using a 4-point scale: 1 (poor), poor anatomic visualization with substantial blurring or artifacts compromising the quality of arterial assessment; 2 (fair), satisfactory anatomic visualization though with blurring or artifact sufficient to prevent confident diagnostic assessment; 3 (good), good arterial delineation, image quality allowing confident arterial assessment; and 4 (excellent), highly diagnostic image quality with sharply defined arterial borders.
 View larger version (49K) | Fig. 1A —Classification of popliteal artery branching. PT = posterior tibial artery, PR = peroneal artery, AT = anterior tibial artery. (Reprinted with permission from Kim D, Orron DE, Skillman JJ. Surgical significance of popliteal artery variants: a unified angiographic classification. Ann Surg 1989; 210:776–781 [12]) Type I-A: Usual pattern of popliteal arterial branching and arterial supply to foot. Type I-B: Trifurcations: anterior, peroneal, and posterior tibial arteries arise at same point without intervening tibioperoneal trunk. Type I-C: Posterior tibial artery is first branch. Anterior tibial and peroneal arteries arise from common trunk. |
 View larger version (45K) | Fig. 1B —Classification of popliteal artery branching. PT = posterior tibial artery, PR = peroneal artery, AT = anterior tibial artery. (Reprinted with permission from Kim D, Orron DE, Skillman JJ. Surgical significance of popliteal artery variants: a unified angiographic classification. Ann Surg 1989; 210:776–781 [12]) Type II-A1: Anterior tibial artery arises above knee joint and has straight course in its proximal segment. Type II-A2: Anterior tibial artery arises above knee joint but takes medial swing, presumably resulting from its passage anterior to popliteus muscle. Type II-B: Posterior tibial artery arises at level of knee joint. Type II-C: Peroneal artery arises above knee joint. |
 View larger version (61K) | Fig. 1C —Classification of popliteal artery branching. PT = posterior tibial artery, PR = peroneal artery, AT = anterior tibial artery. (Reprinted with permission from Kim D, Orron DE, Skillman JJ. Surgical significance of popliteal artery variants: a unified angiographic classification. Ann Surg 1989; 210:776–781 [12]) Type III-A: Posterior tibial artery is hypoplastic and peroneal artery is large. At ankle, distal posterior tibial artery is replaced to peroneal artery. Type III-B: Anterior tibial artery is hypoplastic and peroneal artery is large. At ankle, dorsalis pedis artery is replaced to peroneal artery. Type III-C: Both anterior tibial artery and posterior tibial artery are hypoplastic. At ankle, dorsalis pedis and posterior tibial arteries are replaced to peroneal artery. |
The severity of arteriopathy was defined using a 5-point scale: 0, none; 1, irregularity with < 10% luminal compromise; 2, mild stenosis with 10–50% luminal compromise; 3, significant stenosis with 51–99% luminal compromise; and 4, occlusion. When two or more stenotic segments were identified within a single vascular segment, the more severe was used for grading and analysis.
The presence of noise or artifacts was graded using a 3-point scale: a score of 1 indicating absent or minimal noise or artifact; 2, mild to moderate noise or artifact insufficient to impair image quality; and 3, noise resulting in compromise of image interpretation. A similar 3-point scale was used to score venous contamination: a score of 1 indicating no or minimal venous contamination; 2, mild to moderate venous contamination not interfering with radiologic diagnosis; and 3, severe contamination impacting vascular segmental evaluation.
Observers were also required to indicate the popliteal artery branching pattern present in each lower extremity, using the classification system proposed by Kim et al. [12] (Fig. 1A, 1B, 1C). In the case of discrepancy, agreement was reached by consensus.
A retrospective analysis of patient charts and operative notes was then performed documenting in particular the preoperative peripheral pulse examination findings; surgical approach used; and the side, location, and type of osteocutaneous flap harvested. A detailed analysis of patient continuation notes was also performed to identify those in whom postoperative complications relating to flap viability occurred. In each case, the surgeon involved indicated whether the MRA appearances influenced the decision regarding the side or location from which the flap was harvested or the flap design.
Statistical differences between the ratings of image quality and artifact or noise for the arterial segments between the two observers were assessed using the Wilcoxon's signed rank test. The degree of interobserver agreement for image quality assigned by the two observers as well as for the detection of arterial stenoses was calculated using the kappa coefficient (κ = 0, poor agreement; κ = 0.01–0.20, slight agreement; κ = 0.21–0.40, fair agreement; κ = 0.41–0.60, moderate agreement; κ = 0.61–0.80, good agreement; and κ = 0.81–1.00, excellent agreement). All statistical tests were two-tailed, and differences with p < 0.05 were regarded as statistically significant.
| Results | |
|---|
All contrast-enhanced MRA studies were performed successfully without adverse effect. None of the study participants were determined to have renal impairment according to the results of laboratory tests performed before referral for MRA.
Arterial segments were visualized at MRA with good or excellent image quality (scores 3 or 4) in 722 of 725 segments for observer 1 and 721 segments for observer 2 (Table 2). Evaluation of the scores assigned to the individual arterial segments by both observers using the Wilcoxon's rank sum test revealed no significant difference in scores between the two readers (p = 0.26). The kappa coefficient indicated good interobserver agreement (κ = 0.78) with regard to quality of arterial segmental depiction.
Overall, observer 1 identified atherosclerotic occlusive disease in 131 of 725 segments: irregularity (grade 1) in 84, mild (grade 2) in 18, significant (grade 3) in 16, and occlusion (grade 4) in 13. Observer 2 identified disease in 134 of 725 segments: irregularity (grade 1) in 93, mild (grade 2) in 16, significant (grade 3) in 16, and occlusion (grade 4) in nine. There was good interobserver agreement with regard to scoring of stenoocclusive disease (κ = 0.64; 95% CI, 0.60–0.70).
A number of incidental atherosclerotic findings were also detected, including small infrarenal abdominal aortic aneurysms in two patients (3.1 and 3.0 cm) and a 4.2-cm infrarenal abdominal aortic aneurysm with an associated 3.8-cm partially thrombosed splenic artery aneurysm in another patient (Fig. 2).
The degree of venous contamination was grade 1 (none or minimal) in 706 of 725 segments (97.4%) by observer 1 and 710 of 725 segments (97.9%) by observer 2; the remainder of segments were grade 2 (mild to moderate, not interfering with diagnosis) contamination. No segments had venous contamination of a degree sufficient to compromise diagnostic interpretation.
Scoring of noise or artifact was grade 1 (none or minimal) in 712 segments according to both observers. A mild to moderate degree of noise, insufficient to impair diagnostic interpretation, was observed in the remaining segments, due to motion artifact in a single patient in five segments, hemostatic clip metallic susceptibility artifact secondary to previous renal transplantation in one segment, and B1 inhomogeneity artifact in the right common femoral region in seven segments. This latter artifact, when present, was identified in the thigh station (station 2); however, confident evaluation of this region was possible in all cases because of the presence of field-of-view overlap from the adjacent abdominal–pelvic station.
 View larger version (20K) | Fig. 2 —72-year-old man with oropharyngeal squamous cell carcinoma. Composite volume-rendered abdominopelvic, thigh, and calf station image from 3-T MR angiography shows diffuse arteriomegaly with small infrarenal abdominal aortic aneurysm and left common iliac ectasia. Note presence of atherosclerotic occlusion of left anterior tibial artery, with focal high-grade stenosis of distal right anterior tibial artery. Conventional type I-A popliteal branching is present bilaterally. Arrow indicates incidental splenic artery aneurysm. |
Overall, 47 of the 58 lower extremities evaluated in this study had conventional trifurcation branching, correlating with type I-A in the classification proposed by Kim et al. [12]. The prevalence of each branching pattern encountered during the course of this study is provided in Table 3.
Surgical resection of the underlying oro-mandibular abnormality was subsequently performed in all patients. As a rule, the reconstructive surgeon involved prefers to harvest the fibular free flap from the lower extremity ipsilateral to the point where the recipient cervical vessels are located to facilitate geometric planning during flap insetting. This was considered the default surgical situation whereby the flap was harvested from the ipsilateral fibula, deviation from which on the basis of the MRA findings was considered to be an influence on the surgical approach.
The resultant mandibular defect was corrected using a fibular free flap in 27 of these patients (93.1%). Two patients underwent latissimus dorsi–serratus anterior vascular rib flap mandibular reconstructions when the surgeon judged that MRA findings precluded the use of fibular free flaps.
High-resolution 3D contrast-enhanced MRA of the lower extremities influenced the surgical approach in 16 (55.2%) of 29 patients in this study. In 13 (81.25%) of these 16 patients, the MRA findings caused a change in the lower extremity used for flap harvest (Fig. 3). The modified approach was due to atherosclerotic stenoocclusive disease in seven patients (53.8%) and due to the contribution of the peroneal artery to pedal perfusion in six (46.2%). In two patients, findings at MRA resulted in preferential harvest of latissimus dorsi–serratus anterior vascular rib flaps because of bilateral pedal perfusion by the peroneal artery in one patient as a result of type III-B trifurcation branching (Fig. 4) and bilateral severe atherosclerosis in the second patient. Finally, in a single patient in whom bilateral dorsalis pedis arterial occlusion was identified at MRA, this examination confirmed the posterior tibial artery as the dominant artery to both feet, suggesting that peroneal sacrifice would not compromise pedal perfusion.
Mean duration of patient follow-up was 7 months (range, 2–14 months). No patient mortalities occurred during the follow-up period. Only a single osteocutaneous flap complication was encountered during this time. That complication was in a 43-year-old man in whom type III-B trifurcation branching precluded the use of the right peroneal artery for flap harvesting. The contralateral peroneal artery (type I-A branching pattern) was free from atherosclerotic disease and did not contribute to pedal perfusion and thus was used as the donor vessel. This patient experienced skin-paddle necrosis, salivary fistula formation, and necrotizing infection and was successfully managed with débridement and IV antibiotics. No patients developed symptoms of leg ischemia after fibula flap harvest.
| Discussion | |
|---|
Recent developments in parallel imaging techniques have greatly improved the performance of contrast-enhanced MRA [10]. The penalties imposed on parallel imaging, including loss of signal-to-noise ratio, may be mitigated by imaging at higher field strengths such as 3 T [11, 13]. Indeed, initial experience with application of such techniques to peripheral MRA at 3 T have been encouraging, providing improvements in both signal-to-noise and contrast-to-noise ratios for the popliteal and trifurcation vessels and providing objective and subjective image quality similar to that obtained at 1.5 T for the iliac and thigh stations [8].
The results of our study suggest that high-spatial-resolution MRA had a major impact on the surgical planning of fibular free flap procedures. Preoperative imaging using the technique described influenced the surgical approach in 55.2% of cases in this study, precluding fibular flap harvesting in two patients (6.9%). This influence on surgical approach significantly exceeds that reported in the study of Lorenz and Esclamado [14]. However, atherosclerotic stenoocclusive disease fuelled this alteration in surgical approach in eight of the 16 patients in the current study, suggesting that the higher spatial resolution achieved during the present study relative to that of the time-of-flight approach used by Lorenz and Esclamado may facilitate improved detection of potentially problematic arterial stenoses, thus lowering the threshold for modification of the surgical approach used. In addition, we encountered a significant proportion of anomalous popliteal branching patterns, particularly those of Types II-B and III-B (Fig. 5), compared with the figures quoted by Kim et al. [12], which were based on an analysis of 1,000 lower extremities. It is likely therefore that our figures are reflective of a much smaller patient population, with resultant exaggeration of the prevalence of these anomalies.
 View larger version (56K) | Fig. 3 —58-year-old woman with squamous cell carcinoma of the right neck. Composite thigh and calf MRA stations show type II-B right lower extremity trifurcation branch pattern. |
 View larger version (120K) | Fig. 4 —72-year-old woman with salivary mucoepidermoid carcinoma. Full thickness calf contrast-enhanced MRA maximum intensity projection shows type III-A right lower extremity trifurcation branch pattern. As a result, left peroneal artery was used for fibular flap harvesting. |
Conventionally, infrageniculate arterial anatomy is such that the peroneal artery terminates above the ankle joint, the anterior and posterior tibial arteries continuing across this joint to supply the dorsum and plantar surfaces of the foot, respectively. However, congenital anomalies or acquired disease of the trifurcation vessels is estimated to result in a dominant peroneal artery in 7–12% of the population [15]. An uncommon but important anatomic variant is that of arteria peronea magna, a congenital anomaly in which the peroneal artery is the sole vessel to the foot, with patients having normal distal pulses and the absence of associated symptomatology. This condition has been described as occurring in 0.2–0.9% of the population [16, 17], although we did not encounter any such cases in our study. Furthermore, the peroneal artery may be congenitally hypoplastic, precluding consideration of the ipsilateral fibula for flap harvesting.
 View larger version (106K) | Fig. 5 —60-year-old woman with buccal squamous cell carcinoma. Calf contrast-enhanced MRA shows bilateral type III-B lower extremity branch patterns. Physical examination in this patient revealed presence of bilaterally strong dorsalis pedis and posterior tibial pulses. Patient subsequently underwent latissimus dorsi–serratus anterior vascular rib flap mandibular reconstruction. |
Controversy continues regarding the need for, and optimal means of, preoperative lower extremity vascular assessment before fibular free flap transfer. Given that the most common indication for flap transfer in the mandibular region is squamous cell carcinoma, which is strongly linked to cigarette smoking, the prevalence of atherosclerotic occlusive disease in this patient group is often higher than in the population in general. As a result, several authors have suggested that a thorough history and physical examination should act as primary screening tools for the presence of diminished or absent pedal pulses, further imaging being reserved for those in whom abnormal pulses, a history of trauma, or questionable claudication is elicited [16]. However, the incidence of congenital anomalies in which the peroneal artery provides a significant contribution to the foot circulation that is not detectable on physical examination has been estimated at approximately 5.6% [5, 16]. Furthermore, of particular concern using this approach is that arteria peronea magna may not be detected by clinical examination alone. It also does not result in claudication. Therefore, it potentially may be undiagnosed until the development of foot ischemia in the postoperative period.
Conventional catheter angiography has long been the gold standard in the evaluation of lower extremity vascular patency, allowing high-resolution assessment of the trifurcation vessels, including assessment for the presence of significant stenoses or congenital branch anomalies. Numerous investigators have reported angiographically detected abnormalities that altered the operative plan in up to 25% of cases [18, 19]. However, this technique is invasive and involves considerable ionizing radiation exposure to both the patient and the interventional suite staff. Similarly, the role of CT angiography (CTA) has been evaluated in such patients, yielding favorable results [20, 21]. Nonetheless, CTA involves considerable ionizing radiation exposure, iodine-based contrast medium administration, and potential for nonvisualization of the trifurcation vessels due to the presence of blooming artifact from adjacent calcified plaque.
These limitations have prompted the evaluation of alternative, less invasive, and less expensive imaging techniques in the preoperative routine assessment of patients being considered for fibular free flap procedures. Aly et al. [22] reported a sensitivity of 92% and specificity of 99% for Doppler duplex sonography compared with conventional angiography in the evaluation of lower limb arterial occlusive disease. Futran et al. [23] applied this technique to the assessment of 38 patients in preparation for fibular free flap harvesting, with the vascular anatomic delineation obtained precluding flap transfer in four (10.5%). Doppler duplex sonography affords the additional opportunity to map the location and extent of cutaneous perforators from the peroneal artery, an important determinant of skin-paddle viability in the osteocutaneous fibular flap. However, Doppler duplex sonography is highly operator-dependent, allows only segmental views of vascular anatomy rather than longitudinal images, and may be difficult to interpret in the presence of dense mural atherosclerotic calcification and abundant subcutaneous fat.
MRA has also been assessed with promising results as a potential successor to conventional arteriography in the preoperative evaluation of these patients. This technique offers the capability of reliable noninvasive vascular imaging in the absence of associated ionizing radiation exposure and without the use of iodine-based contrast medium injection. Furthermore, MRA has the additional attraction of costing less than one half that of conventional angiography. Lorenz and Esclamado [14] reported their experience of a time-of-flight angiographic technique in 32 patients before fibular free flap harvesting. MRA resulted in a change of side harvested in 12.5% and exclusion of the fibula as a potential donor site in 9%. Furthermore, one patient required alteration in flap design because of the presence of an abnormally long tibioperoneal trunk. In all, 25% of patients in this study had alterations in their treatment plan as a result of preoperative MRA, and 70% showed pulse examinations that were discordant to vessel patency on MRA.
Notably, all 29 of the patients involved in the Lorenz and Esclamado [14] study subsequently underwent successful free flap surgery without ischemic complication. However, time-of-flight angiography is susceptible to a number of flow-related artifacts. It is also of limited sensitivity and specificity relative to its contrast-enhanced counterparts and has in recent years been replaced by these latter techniques [24]. Indeed, Kelly et al. [25] used their experience in more than 100 contrast-enhanced examinations performed on a 1.5-T system to exemplify the thought processes involved in preoperative MRA as a prelude to fibular flap harvesting. Similarly, Hölzle et al. [26] directly compared contrast-enhanced MRA at 1.5 T with conventional digital subtraction angiography in 15 patients before fibular flap transfer and found this noninvasive technique to be of at least equal benefit.
Peripheral MRA at 3 T allows high spatial resolution depiction of the arterial vasculature from the diaphragm to the toes, with delineation of the trifurcation branching pattern and resultant potential avoidance of foot ischemia. Our study strongly supports the hypothesis that trifurcation vessel imaging should be a prerequisite to fibular free flap harvesting and that this technique represents a feasible alternative to other invasive or cross-sectional imaging techniques in this regard. Indeed, had we performed MRA only in those patients in whom abnormal peripheral pulses were detected on physical examination in this group, a single unilateral type I-B, two unilateral type II-B, two unilateral types III-A and III-B, and a single case of bilateral type III-B branching patterns would have been over-looked, with the potential for ipsilateral foot ischemia in five (17.2%) of these 29 patients.
We acknowledge the presence of limitations within this study, including the lack of comparable imaging techniques and relatively small patient numbers. However, we think that the consistently high confidence and level of interobserver agreement validate this technique.
In conclusion, our study strongly supports the hypothesis that trifurcation vessel imaging should be a prerequisite to fibular free flap harvesting and that high-spatial-resolution MRA at 3 T represents a desirable alternative to other invasive or cross-sectional imaging techniques in this regard.
Address correspondence to D. G. Lohan ([email protected]).
| References | |
|---|
