JOURNAL OF MAGNETIC RESONANCE IMAGING 41:550–554 (2015)
Technical Development
Regional Analysis of Femoral Head Perfusion Following Displaced Fractures of the Femoral Neck Jonathan P. Dyke, PhD,1* Lionel E. Lazaro, MD,2,3 Carolyn M. Hettrich, MD,4 Keith D. Hentel, MD,1 David L. Helfet, MD,2,3 and Dean G. Lorich, MD2,3 Purpose: To assess regional variations in the arterial and venous blood supply to the femoral head following displaced fracture of the femoral neck using dynamic contrast enhanced (DCE)-MRI quadrant analysis. Materials and Methods: A total of 27 subjects with displaced femoral neck fractures were enrolled in the study. Quadrant specific DCE-MRI perfusion analysis was performed on a 1.5 Tesla MRI scanner. Simultaneous imaging of control and displaced fractured hips was done for comparison. Results: Quadrant specific decreases were found in the arterial (A (0.52 versus 0.27; P ¼ 5.7E-13), Akep (1.0/ min1 versus 0.41/min1; P ¼ 1.3E-9) and venous (kel (0.05/min1 versus 0.02/min1; P ¼ 5.1E-5) supply to the femoral head between control and injured sides using a two-factor analysis of variance test. The fractional perfusion (initial area under the curve) in the supero/inferolateral quadrants was 49% min/54% min, in the supero/ inferomedial quadrants was 43% min/46% min and for the total femoral head was 39% min on the fracture versus control sides. Conclusion: Quadrant specific decreases in arterial and venous perfusion on the fracture side were observed when compared with control. Key Words: MRI; femoral neck fracture; bone perfusion; dynamic contrast enhanced MRI; DCE-MRI J. Magn. Reson. Imaging 2015;41:550–554. C 2013 Wiley Periodicals, Inc. V
AN ESTIMATED 352,000 hip fractures occur each year in the United States of which 90% are the result of a fall (1,2). The reported incidence of the two 1 Department of Radiology, Weill Cornell Medical College, New York, New York, USA. 2 Department of Orthopaedic Surgery, Weill Cornell Medical College, New York, New York, USA. 3 Department of Orthopaedic Surgery, Hospital for Special Surgery, New York, New York, USA. 4 Department of Orthopaedic Surgery, University of Iowa, Iowa City, Iowa, USA. *Address reprint requests to: J.P.D., Weill Cornell Medical College, 1300 York Avenue – Box 234, New York, NY 10021. E-mail: [email protected] Received September 30, 2013; Accepted November 9, 2013. DOI 10.1002/jmri.24524 View this article online at wileyonlinelibrary.com. C 2013 Wiley Periodicals, Inc. V
primary hip fracture types include intertrochanteric fractures at a rate of 54% and femoral neck fractures at a rate of 46% (3). Surgical fixation of a femoral neck fracture is often associated with a high complication rate (nonunion, osteonecrosis [ON] of the femoral head and late segmental collapse) and an increased reoperation rate (4,5). Traumatic injury and associated compromise of the tenuous blood supply to the femoral head is of major concern and is presumed to be one of the primary etiologic factors for these complications. Previous studies have shown that hemarthrosis in the hip joint resulted in increased intracapsular pressure even in nondisplaced fractures of the femoral neck (6). An increased intra-articular pressure is thought to impinge on blood flow in the retinacular system; the major blood supply to the femoral head and further increase the risk for ON (7). A noninvasive technique to assess femoral head perfusion in fractures of the femoral neck would allow for tailored surgical treatment, prognostic indicators of successful perfusion, and increased potential for healing (8–10). Dynamic contrast enhanced (DCE)MRI has proven to be a promising tool to monitor perfusion to the femoral head following femoral neck fractures in a noninvasive manner (11,12). Studies have grouped DCE-MRI time intensity curves (TIC) of the femoral head to obtain accurate predictions of successful osteosynthesis (13). These studies defined a small fixed size region of interest (ROI) in a single area of the femoral head (14,15). ON was noted only in curves having greater than a 70% decrease in perfusion from the control side. Additional studies have quantified in a cadaveric model the contributions of specific arteries to each quadrant of the femoral head using static contrast enhanced MRI. It was found the inferior retinacular (vincular) artery played a role in continued perfusion of the entire femoral head (more in the inferior quadrants) when perfusion through the superior retinacular artery was interrupted by tying-off the deep branch of the MFCA (16). Quadrant specific variances in static perfusion were found due to the differing entrance points of the various feeding vessels in the arterial architecture. Our purpose was to determine whether quantitative DCE-MRI quadrant analysis of the femoral head may identify regional perfusion
550
Perfusion of Femoral Head Postfracture
551
Normative renal function was confirmed as a prerequisite for administration of Gadolinium based contrast agents. A total of 27 subjects (9 males and 18 females) with an average age of 60.0 years (613.2) were enrolled (range, 28 to 85 years). All fractures were classified as Garden Stage III–IV in this study (17). The Pauwels angle was measured on AP X-ray images to assess the obliquity of the fracture line and assign a fracture classification. The Pauwels fracture classifications were made as follows: Type I (angle < 30 ), Type II (30 angle 50 ), Type III (angle > 50 ) (18). MRI Methods Figure 1. Dynamic contrast enhanced MRI shows the average time intensity curves (n ¼ 27) for the entire femoral head on both the fracture and contralateral hips.
All subjects were scanned on a 1.5 Tesla (T) MRI Scanner (General Electric, Milwaukee, WI) using an eight-channel phased array coil. Gadolinium–diethylenetriamine penta-acetic acid (Gd-DTPA) (MagnevistV ; Bayer HealthCare Pharmaceuticals Inc., Wayne, NJ) was administered at 0.1 mmol/kg at a rate of 2 cc/s followed by a saline flush using a power injector. The DCE-MRI sequence used a fast T1-weighted fat suppressed three-dimensional (3D) spoiled gradient echo pulse sequence (Liver Acquisition with Volume Acquisition, LAVA). Images were acquired in the coronal plane to simultaneously visualize both injured and contralateral hips. A temporal resolution of between 7 and 8 s/image for a total of 45 time points in a scan time of 6 min was maintained. Thirty slices of 4 mm thickness with a 40 cm field of view and a 256 128 matrix with a repetition time of 3.6 ms, echo time of 1.7 ms and 12 flip angle were used. R
characteristics that could potentially be averaged out by analysis of the entire femoral head or from a small centrally placed ROI. MATERIALS AND METHODS Clinical This study was conducted under a prospective research protocol approved by our Institutional Review Board with all subjects giving full informed consent. All subjects were initially diagnosed with displaced femoral neck fractures using digital radiography and clinical examination in the emergency room setting within 1.0 days (61.7) of the initial injury.
Figure 2. Dynamic contrast enhanced MRI of the average time curves (n ¼ 27) were analyzed for the four femoral head quadrants on both the fractured and contralateral matched control hips.
552
Dyke et al.
Table 1 Post-Hoc Comparison of the DCE Fit Parameters Between the Control and Fracture Side Was Made for the Entire Femoral Head and Each of the Associated Quadrants Using the Nonparametric Mann-Whitney test (n¼27) Average of Model Fits (n¼27) Whole control head Whole fracture head p-value Control superolateral Fracture superolateral p-value Control superomedial Fracture superomedial p-value Control inferomedial Fracture inferomedial p-value Control inferolateral Fracture inferolateral p-value
A
kep
kel
Akep
IAUC
Peak
0.526 0.293 2.10E-4 0.576 0.322 1.04E-3 0.470 0.191 1.90E-5 0.453 0.228 1.70E-4 0.578 0.343 3.66E-4
1.760 1.185 2.62E-02 2.122 1.723 1.11E-01 1.807 1.803 5.98E-01 2.064 1.777 3.92E-01 2.148 1.635 3.00E-2
0.067 0.038 1.97E-01 0.056 0.049 9.23E-04 0.038 0.018 5.39E-01 0.051 0.028 1.62E-02 0.055 0.031 1.33E-03
0.908 0.283 3.69E-05 1.196 0.488 2.90E-04 0.664 0.234 4.30E-04 0.874 0.337 8.00E-04 1.225 0.569 3.30E-04
0.426 0.168 7.44E-06 0.518 0.255 3.09E-04 0.317 0.136 6.67E-05 0.393 0.181 1.62E-04 0.533 0.288 3.30E-04
0.388 0.169 7.44E-06 0.459 0.250 5.60E-04 0.300 0.133 2.53E-05 0.359 0.180 2.20E-04 0.475 0.281 5.75E-04
*Significant (p < .05) differences were found for parameters highlighted in bold.
Image Analysis Techniques The Brix 2-compartment pharmacokinetic model was used to analyze the uptake curves in the injured and contralateral femoral heads. The model contains the parameters: A (signal amplitude), kep (exchange rate between plasma and extravascular extracellular space (EES), min1), and kel (elimination rate, min1) (19). The product of Akep was calculated to estimate the initial arterial wash-in or uptake slope in the region. The initial area under the curve (IAUC) and peak value were analytically computed at 90 s following injection from the Brix model equation. ROIs were placed on the entire femoral head and subdivided into quadrants using the same method as previously
published cadaveric studies (18). Analysis software was written in-house using IDL 8.1 (Exelis VIS, Boulder, CO) to fit the time intensity curves. Variance in angulation of the quadrants in the femoral head was accounted for by defining the angle between the head and neck in each subject. Regions of hyperemia that crossed into the femoral head were excluded.
Statistical Methods A two-way analysis of variance (ANOVA) test with replication was used to assess significance between the quadrants and between control and fracture groups. A nonparametric Mann-Whitney test was used to
Figure 3. The DCE-MRI perfusion parameters are compared versus the Pauwels fracture grade. A trend can be seen showing decreased perfusion with an increase in Pauwels angle and grade.
Perfusion of Femoral Head Postfracture
compare each of the DCE-MRI fit parameters between control and fractured sides. This test does not assume normality in the sample distribution and is more robust with respect to outliers. Significance for all tests was achieved when the two-tailed P-value fell below 0.05. In addition, the model curves were averaged over all subjects to obtain representative TIC’s for the entire femoral head (Fig. 1) and the individual quadrants (Fig. 2). The standard error of the mean was computed for each time point in these curves.
RESULTS A two-factor ANOVA test with replication showed differences between control and injured sides and between quadrants respectively as follows: A (0.52 versus 0.27; P ¼ 5.7E-13/0.004), kep (2.0/min1 versus 1.7/min1; P ¼ 0.12/0.97), kel (0.05/min1 versus 0.02/min1;5.1E-5/0.79), Akep (1.0/min1 versus 0.41/min1; P ¼ 1.3E-9/0.002), IAUC (0.44 %min versus 0.22 %min; P ¼ 3.9E-11/0.0001), peak (0.40 versus 0.21; P ¼ 3.8E-11/5.8E-5). A comparison of the DCE fit parameters between the control and fracture side for the entire femoral head and each of the associated quadrants was performed (Table 1). Significant (P < 0.05) differences were found on the injured side for the initial amplitude (A), slope (Akep), IAUC and peak values for the entire femoral head and all subsequent quadrants. Decreases were also seen in the wash-out elimination rate (kel) in all but the superomedial quadrant. Perfusion of the femoral head (IAUC, peak) was decreased in the injured versus contralateral sides. The fractional IAUC for the total femoral head on the fracture versus control side was 39%min. The IAUC of the superomedial quadrant was lowest for both control and fracture sides and was significantly decreased (P < 0.01) from the superolateral and inferolateral on both. The change in DCE-MRI perfusion parameters was compared with the Pauwels fracture classification (Fig. 3). A distinct trend was seen for the Akep (P ¼ 0.07), IAUC (P ¼ 0.18) and peak enhancement (P ¼ 0.18) versus fracture grade.
DISCUSSION Two principal perfusion deficits were seen within the femoral head of the fractured hip compared with that of the contralateral hip. Decreased arterial inflow to the region was seen by means of a reduction in the parameters A and Akep. This deficit in inflow supply may be related to both direct traumatic injuries to the retinacular system and fracture induced intracapsular tamponade secondary to hemarthrosis. A decrease in elimination or wash-out of the contrast from the region was seen as a more negative value in the parameter kel suggesting venous outflow obstruction or stasis as compared to that of the control side. This was visually apparent as an increasing time intensity curve at the end of the scan and may be due to pressure secondary to the hematoma tamponade effect at the venous structure of the retinacular system. Likewise, a trend indicated decreased perfusion and vascular disruption with increasing Pauwels angle and
553
shear forces at the fracture line. In total, the IAUC and peak enhancement were both significantly reduced on the fracture side indicating decreased perfusion. The peak enhancement on both control and injured sides was seen in the superolateral and inferolateral quadrants. This is a comparable finding to a previously published cadaveric study that reported that the superior and inferior retinacular arteries are the main blood supply to the femoral head (18). Maintenance of perfusion to the entire femoral head can potentially be accounted for from the inferior retinacular artery (17,20). This arterial system lays within fibrous extension of the capsule (inferior retinaculum of Weitbrecht) (21). This structure is elevated from the femoral neck possibly sparing the inferior retinacular arterial system from traumatic injury by the fracture. It therefore may remain a nutrient artery to the femoral head even in the event of a femoral neck fracture. Capsular decompression in displaced intracapsular hip fractures was found to significantly reduce the incidence and delay onset of ON (22). No significant differences in either union rates or time to union was found between the screw fixation with capsular decompression (N ¼ 99) and screw fixation alone (N ¼ 102) groups. However, the decrease in intra-articular pressure in the capsular decompression group was thought to mildly improve blood flow in the region thereby sufficiently reducing ON incidence and also delaying ON onset by minimizing the tamponade effect. This decreased perfusion may potentially be secondary to increased intra-articular pressure and the resultant tamponade effect of the fracture hemarthrosis. Limitations in this study were that significance was difficult to achieve between Pauwels angle and DCEMRI parameters with only (n ¼ 3) of Type I compared with (n ¼ 12) in Type II and (n ¼ 12) in Type 3 classifications. In addition, regions of hyperemia must be carefully avoided such that false increases in quadrant perfusion adjacent to the fracture line are not measured. In conclusion, the primary finding of this study was that quadrant specific decreases in arterial inflow and venous outflow were significant and could be quantified in the femoral head on the injured side compared with the contralateral hip. This study presents the potential that quantitative DCE-MRI may be used as a tool in the clinic to investigate quadrant specific perfusion disruption of the femoral head following femoral neck fractures. These data yielded objective information on regional arterial and venous perfusion deficits that may aid in providing a more accurate prognostic indicator of femoral head viability and successful surgical outcomes as well as aid in prospectively choosing patient specific treatment options. REFERENCES 1. http://orthoinfo.aaos.org/topic.cfm?topic¼A00309. 2012, American Academy of Orthopaedic Surgeons. 2. Baumgaertner MR, Higgins TF. Femoral neck fractures. In: Bucholz RW, Heckman JD, Rockwood CA, Green DP, editors. Rockwood and Green’s fractures in adults. Philadelphia: Lippincott Williams and Wilkins; 2002. p 1579.
554 3. Fox KM, Magaziner J, Hebel JR, Kenzora JE, Kashner TM. Intertrochanteric versus femoral neck hip fractures: differential characteristics, treatment, and sequelae. J Gerontol 1999;54: M635–M640. 4. Tidermark J, Zethraeus N, Svensson O, Tornkvist H, Ponzer S. Quality of life related to fracture displacement among elderly patients with femoral neck fractures treated with internal fixation. 2002. J Orthop Trauma 2003;17:S17–S21. 5. Miyamoto RG, Kaplan KM, Levine BR, Egol KA, Zuckerman JD. Surgical management of hip fractures: an evidence-based review of the literature. I: femoral neck fractures. J Am Acad Orthop Surg 2008;16:596–607. 6. Beck M, Siebenrock KA, Affolter B, Notzli H, Parvizi J, Ganz R. Increased intraarticular pressure reduces blood flow to the femoral head. Clin Orthop Relat Res 2004;424:149–152. 7. Ogden JA. Changing patterns of proximal femoral vascularity. J Bone Joint Surg Am 1974;56:941–950. 8. Ehlinger M, Moser T, Adam P, et al. Early prediction of femoral head avascular necrosis following neck fracture. Orthop Traumatol Surg Res 2011;97:79–88. 9. Gautier E, Ganz K, Krugel N, Gill T, Ganz R. Anatomy of the medial femoral circumflex artery and its surgical implications. J Bone Joint Surg Br 2000;82:679–683. 10. Dyke JP, Aaron RK. Noninvasive methods of measuring bone blood perfusion. Ann N Y Acad Sci 2010;1192:95–102. 11. Lang P, Mauz M, Schorner W, et al. Acute fracture of the femoral neck: assessment of femoral head perfusion with gadopentetate dimeglumine-enhanced MR imaging. AJR Am J Roentgenol 1993; 160:335–341.
Dyke et al. 12. Kamano M, Narita S, Honda Y, Fukushima K, Yamano Y. Contrast enhanced magnetic resonance imaging for femoral neck fracture. Clin Orthop Relat Res 1998;350:179–186. 13. Hirata T, Konishiike T, Kawai A, Sato T, Inoue H. Dynamic magnetic resonance imaging of femoral head perfusion in femoral neck fracture. Clin Orthop Relat Res 2001;393:294–301. 14. Konishiike T, Makihata E, Tago H, Sato T, Inoue H. Acute fracture of the neck of the femur: an assessment of perfusion of the head by dynamic MRI. J Bone Joint Surg Br 1999;81:596–599. 15. Kaushik A, Sankaran B, Varghese M. Prognostic value of dynamic MRI in assessing post-traumatic femoral head vascularity. Skeletal Radiol 2009;38:565–569. 16. Boraiah S, Dyke JP, Hettich C, et al. Assessment of vascularity of the femoral head using gadolinium (Gd-DTPA)-enhanced magnetic resonance imaging. J Bone Joint Surg Br 2009;91B:131–137. 17. Garden RS. Low-angle fixation in fractures of the femoral neck. J Bone Joint Surg Br 1961;43:647–661. 18. Pauwels F. Der Schenkelhalsbruch. Ein mechanisches Problem, F. Enke, Stuttgart: 1935. 19. Tofts P. Modeling tracer kinetics in dynamic Gd-DTPA MR imaging. J Magn Reson Imaging 1997;7:91–101. 20. Papadakis SA, Segos D, Kouvaras I, Dagas S, Malakasis M, Grivas TB. Integrity of posterior retinaculum after displaced femoral neck fractures. Injury 2009;40:277–279. 21. Gojda J, Bartonicek J. The retinacula of Weitbrecht in the adult hip. Surg Radiol Anat 2012;34:31–38. 22. Wong TC, Yeung SH, Ip FK. The effectiveness of capsular decompression for internal fixation of intracapsular hip fractures. J Orthop Surg (Hong Kong) 2007;15:282–285.
Technical Development
Regional Analysis of Femoral Head Perfusion Following Displaced Fractures of the Femoral Neck Jonathan P. Dyke, PhD,1* Lionel E. Lazaro, MD,2,3 Carolyn M. Hettrich, MD,4 Keith D. Hentel, MD,1 David L. Helfet, MD,2,3 and Dean G. Lorich, MD2,3 Purpose: To assess regional variations in the arterial and venous blood supply to the femoral head following displaced fracture of the femoral neck using dynamic contrast enhanced (DCE)-MRI quadrant analysis. Materials and Methods: A total of 27 subjects with displaced femoral neck fractures were enrolled in the study. Quadrant specific DCE-MRI perfusion analysis was performed on a 1.5 Tesla MRI scanner. Simultaneous imaging of control and displaced fractured hips was done for comparison. Results: Quadrant specific decreases were found in the arterial (A (0.52 versus 0.27; P ¼ 5.7E-13), Akep (1.0/ min1 versus 0.41/min1; P ¼ 1.3E-9) and venous (kel (0.05/min1 versus 0.02/min1; P ¼ 5.1E-5) supply to the femoral head between control and injured sides using a two-factor analysis of variance test. The fractional perfusion (initial area under the curve) in the supero/inferolateral quadrants was 49% min/54% min, in the supero/ inferomedial quadrants was 43% min/46% min and for the total femoral head was 39% min on the fracture versus control sides. Conclusion: Quadrant specific decreases in arterial and venous perfusion on the fracture side were observed when compared with control. Key Words: MRI; femoral neck fracture; bone perfusion; dynamic contrast enhanced MRI; DCE-MRI J. Magn. Reson. Imaging 2015;41:550–554. C 2013 Wiley Periodicals, Inc. V
AN ESTIMATED 352,000 hip fractures occur each year in the United States of which 90% are the result of a fall (1,2). The reported incidence of the two 1 Department of Radiology, Weill Cornell Medical College, New York, New York, USA. 2 Department of Orthopaedic Surgery, Weill Cornell Medical College, New York, New York, USA. 3 Department of Orthopaedic Surgery, Hospital for Special Surgery, New York, New York, USA. 4 Department of Orthopaedic Surgery, University of Iowa, Iowa City, Iowa, USA. *Address reprint requests to: J.P.D., Weill Cornell Medical College, 1300 York Avenue – Box 234, New York, NY 10021. E-mail: [email protected] Received September 30, 2013; Accepted November 9, 2013. DOI 10.1002/jmri.24524 View this article online at wileyonlinelibrary.com. C 2013 Wiley Periodicals, Inc. V
primary hip fracture types include intertrochanteric fractures at a rate of 54% and femoral neck fractures at a rate of 46% (3). Surgical fixation of a femoral neck fracture is often associated with a high complication rate (nonunion, osteonecrosis [ON] of the femoral head and late segmental collapse) and an increased reoperation rate (4,5). Traumatic injury and associated compromise of the tenuous blood supply to the femoral head is of major concern and is presumed to be one of the primary etiologic factors for these complications. Previous studies have shown that hemarthrosis in the hip joint resulted in increased intracapsular pressure even in nondisplaced fractures of the femoral neck (6). An increased intra-articular pressure is thought to impinge on blood flow in the retinacular system; the major blood supply to the femoral head and further increase the risk for ON (7). A noninvasive technique to assess femoral head perfusion in fractures of the femoral neck would allow for tailored surgical treatment, prognostic indicators of successful perfusion, and increased potential for healing (8–10). Dynamic contrast enhanced (DCE)MRI has proven to be a promising tool to monitor perfusion to the femoral head following femoral neck fractures in a noninvasive manner (11,12). Studies have grouped DCE-MRI time intensity curves (TIC) of the femoral head to obtain accurate predictions of successful osteosynthesis (13). These studies defined a small fixed size region of interest (ROI) in a single area of the femoral head (14,15). ON was noted only in curves having greater than a 70% decrease in perfusion from the control side. Additional studies have quantified in a cadaveric model the contributions of specific arteries to each quadrant of the femoral head using static contrast enhanced MRI. It was found the inferior retinacular (vincular) artery played a role in continued perfusion of the entire femoral head (more in the inferior quadrants) when perfusion through the superior retinacular artery was interrupted by tying-off the deep branch of the MFCA (16). Quadrant specific variances in static perfusion were found due to the differing entrance points of the various feeding vessels in the arterial architecture. Our purpose was to determine whether quantitative DCE-MRI quadrant analysis of the femoral head may identify regional perfusion
550
Perfusion of Femoral Head Postfracture
551
Normative renal function was confirmed as a prerequisite for administration of Gadolinium based contrast agents. A total of 27 subjects (9 males and 18 females) with an average age of 60.0 years (613.2) were enrolled (range, 28 to 85 years). All fractures were classified as Garden Stage III–IV in this study (17). The Pauwels angle was measured on AP X-ray images to assess the obliquity of the fracture line and assign a fracture classification. The Pauwels fracture classifications were made as follows: Type I (angle < 30 ), Type II (30 angle 50 ), Type III (angle > 50 ) (18). MRI Methods Figure 1. Dynamic contrast enhanced MRI shows the average time intensity curves (n ¼ 27) for the entire femoral head on both the fracture and contralateral hips.
All subjects were scanned on a 1.5 Tesla (T) MRI Scanner (General Electric, Milwaukee, WI) using an eight-channel phased array coil. Gadolinium–diethylenetriamine penta-acetic acid (Gd-DTPA) (MagnevistV ; Bayer HealthCare Pharmaceuticals Inc., Wayne, NJ) was administered at 0.1 mmol/kg at a rate of 2 cc/s followed by a saline flush using a power injector. The DCE-MRI sequence used a fast T1-weighted fat suppressed three-dimensional (3D) spoiled gradient echo pulse sequence (Liver Acquisition with Volume Acquisition, LAVA). Images were acquired in the coronal plane to simultaneously visualize both injured and contralateral hips. A temporal resolution of between 7 and 8 s/image for a total of 45 time points in a scan time of 6 min was maintained. Thirty slices of 4 mm thickness with a 40 cm field of view and a 256 128 matrix with a repetition time of 3.6 ms, echo time of 1.7 ms and 12 flip angle were used. R
characteristics that could potentially be averaged out by analysis of the entire femoral head or from a small centrally placed ROI. MATERIALS AND METHODS Clinical This study was conducted under a prospective research protocol approved by our Institutional Review Board with all subjects giving full informed consent. All subjects were initially diagnosed with displaced femoral neck fractures using digital radiography and clinical examination in the emergency room setting within 1.0 days (61.7) of the initial injury.
Figure 2. Dynamic contrast enhanced MRI of the average time curves (n ¼ 27) were analyzed for the four femoral head quadrants on both the fractured and contralateral matched control hips.
552
Dyke et al.
Table 1 Post-Hoc Comparison of the DCE Fit Parameters Between the Control and Fracture Side Was Made for the Entire Femoral Head and Each of the Associated Quadrants Using the Nonparametric Mann-Whitney test (n¼27) Average of Model Fits (n¼27) Whole control head Whole fracture head p-value Control superolateral Fracture superolateral p-value Control superomedial Fracture superomedial p-value Control inferomedial Fracture inferomedial p-value Control inferolateral Fracture inferolateral p-value
A
kep
kel
Akep
IAUC
Peak
0.526 0.293 2.10E-4 0.576 0.322 1.04E-3 0.470 0.191 1.90E-5 0.453 0.228 1.70E-4 0.578 0.343 3.66E-4
1.760 1.185 2.62E-02 2.122 1.723 1.11E-01 1.807 1.803 5.98E-01 2.064 1.777 3.92E-01 2.148 1.635 3.00E-2
0.067 0.038 1.97E-01 0.056 0.049 9.23E-04 0.038 0.018 5.39E-01 0.051 0.028 1.62E-02 0.055 0.031 1.33E-03
0.908 0.283 3.69E-05 1.196 0.488 2.90E-04 0.664 0.234 4.30E-04 0.874 0.337 8.00E-04 1.225 0.569 3.30E-04
0.426 0.168 7.44E-06 0.518 0.255 3.09E-04 0.317 0.136 6.67E-05 0.393 0.181 1.62E-04 0.533 0.288 3.30E-04
0.388 0.169 7.44E-06 0.459 0.250 5.60E-04 0.300 0.133 2.53E-05 0.359 0.180 2.20E-04 0.475 0.281 5.75E-04
*Significant (p < .05) differences were found for parameters highlighted in bold.
Image Analysis Techniques The Brix 2-compartment pharmacokinetic model was used to analyze the uptake curves in the injured and contralateral femoral heads. The model contains the parameters: A (signal amplitude), kep (exchange rate between plasma and extravascular extracellular space (EES), min1), and kel (elimination rate, min1) (19). The product of Akep was calculated to estimate the initial arterial wash-in or uptake slope in the region. The initial area under the curve (IAUC) and peak value were analytically computed at 90 s following injection from the Brix model equation. ROIs were placed on the entire femoral head and subdivided into quadrants using the same method as previously
published cadaveric studies (18). Analysis software was written in-house using IDL 8.1 (Exelis VIS, Boulder, CO) to fit the time intensity curves. Variance in angulation of the quadrants in the femoral head was accounted for by defining the angle between the head and neck in each subject. Regions of hyperemia that crossed into the femoral head were excluded.
Statistical Methods A two-way analysis of variance (ANOVA) test with replication was used to assess significance between the quadrants and between control and fracture groups. A nonparametric Mann-Whitney test was used to
Figure 3. The DCE-MRI perfusion parameters are compared versus the Pauwels fracture grade. A trend can be seen showing decreased perfusion with an increase in Pauwels angle and grade.
Perfusion of Femoral Head Postfracture
compare each of the DCE-MRI fit parameters between control and fractured sides. This test does not assume normality in the sample distribution and is more robust with respect to outliers. Significance for all tests was achieved when the two-tailed P-value fell below 0.05. In addition, the model curves were averaged over all subjects to obtain representative TIC’s for the entire femoral head (Fig. 1) and the individual quadrants (Fig. 2). The standard error of the mean was computed for each time point in these curves.
RESULTS A two-factor ANOVA test with replication showed differences between control and injured sides and between quadrants respectively as follows: A (0.52 versus 0.27; P ¼ 5.7E-13/0.004), kep (2.0/min1 versus 1.7/min1; P ¼ 0.12/0.97), kel (0.05/min1 versus 0.02/min1;5.1E-5/0.79), Akep (1.0/min1 versus 0.41/min1; P ¼ 1.3E-9/0.002), IAUC (0.44 %min versus 0.22 %min; P ¼ 3.9E-11/0.0001), peak (0.40 versus 0.21; P ¼ 3.8E-11/5.8E-5). A comparison of the DCE fit parameters between the control and fracture side for the entire femoral head and each of the associated quadrants was performed (Table 1). Significant (P < 0.05) differences were found on the injured side for the initial amplitude (A), slope (Akep), IAUC and peak values for the entire femoral head and all subsequent quadrants. Decreases were also seen in the wash-out elimination rate (kel) in all but the superomedial quadrant. Perfusion of the femoral head (IAUC, peak) was decreased in the injured versus contralateral sides. The fractional IAUC for the total femoral head on the fracture versus control side was 39%min. The IAUC of the superomedial quadrant was lowest for both control and fracture sides and was significantly decreased (P < 0.01) from the superolateral and inferolateral on both. The change in DCE-MRI perfusion parameters was compared with the Pauwels fracture classification (Fig. 3). A distinct trend was seen for the Akep (P ¼ 0.07), IAUC (P ¼ 0.18) and peak enhancement (P ¼ 0.18) versus fracture grade.
DISCUSSION Two principal perfusion deficits were seen within the femoral head of the fractured hip compared with that of the contralateral hip. Decreased arterial inflow to the region was seen by means of a reduction in the parameters A and Akep. This deficit in inflow supply may be related to both direct traumatic injuries to the retinacular system and fracture induced intracapsular tamponade secondary to hemarthrosis. A decrease in elimination or wash-out of the contrast from the region was seen as a more negative value in the parameter kel suggesting venous outflow obstruction or stasis as compared to that of the control side. This was visually apparent as an increasing time intensity curve at the end of the scan and may be due to pressure secondary to the hematoma tamponade effect at the venous structure of the retinacular system. Likewise, a trend indicated decreased perfusion and vascular disruption with increasing Pauwels angle and
553
shear forces at the fracture line. In total, the IAUC and peak enhancement were both significantly reduced on the fracture side indicating decreased perfusion. The peak enhancement on both control and injured sides was seen in the superolateral and inferolateral quadrants. This is a comparable finding to a previously published cadaveric study that reported that the superior and inferior retinacular arteries are the main blood supply to the femoral head (18). Maintenance of perfusion to the entire femoral head can potentially be accounted for from the inferior retinacular artery (17,20). This arterial system lays within fibrous extension of the capsule (inferior retinaculum of Weitbrecht) (21). This structure is elevated from the femoral neck possibly sparing the inferior retinacular arterial system from traumatic injury by the fracture. It therefore may remain a nutrient artery to the femoral head even in the event of a femoral neck fracture. Capsular decompression in displaced intracapsular hip fractures was found to significantly reduce the incidence and delay onset of ON (22). No significant differences in either union rates or time to union was found between the screw fixation with capsular decompression (N ¼ 99) and screw fixation alone (N ¼ 102) groups. However, the decrease in intra-articular pressure in the capsular decompression group was thought to mildly improve blood flow in the region thereby sufficiently reducing ON incidence and also delaying ON onset by minimizing the tamponade effect. This decreased perfusion may potentially be secondary to increased intra-articular pressure and the resultant tamponade effect of the fracture hemarthrosis. Limitations in this study were that significance was difficult to achieve between Pauwels angle and DCEMRI parameters with only (n ¼ 3) of Type I compared with (n ¼ 12) in Type II and (n ¼ 12) in Type 3 classifications. In addition, regions of hyperemia must be carefully avoided such that false increases in quadrant perfusion adjacent to the fracture line are not measured. In conclusion, the primary finding of this study was that quadrant specific decreases in arterial inflow and venous outflow were significant and could be quantified in the femoral head on the injured side compared with the contralateral hip. This study presents the potential that quantitative DCE-MRI may be used as a tool in the clinic to investigate quadrant specific perfusion disruption of the femoral head following femoral neck fractures. These data yielded objective information on regional arterial and venous perfusion deficits that may aid in providing a more accurate prognostic indicator of femoral head viability and successful surgical outcomes as well as aid in prospectively choosing patient specific treatment options. REFERENCES 1. http://orthoinfo.aaos.org/topic.cfm?topic¼A00309. 2012, American Academy of Orthopaedic Surgeons. 2. Baumgaertner MR, Higgins TF. Femoral neck fractures. In: Bucholz RW, Heckman JD, Rockwood CA, Green DP, editors. Rockwood and Green’s fractures in adults. Philadelphia: Lippincott Williams and Wilkins; 2002. p 1579.
554 3. Fox KM, Magaziner J, Hebel JR, Kenzora JE, Kashner TM. Intertrochanteric versus femoral neck hip fractures: differential characteristics, treatment, and sequelae. J Gerontol 1999;54: M635–M640. 4. Tidermark J, Zethraeus N, Svensson O, Tornkvist H, Ponzer S. Quality of life related to fracture displacement among elderly patients with femoral neck fractures treated with internal fixation. 2002. J Orthop Trauma 2003;17:S17–S21. 5. Miyamoto RG, Kaplan KM, Levine BR, Egol KA, Zuckerman JD. Surgical management of hip fractures: an evidence-based review of the literature. I: femoral neck fractures. J Am Acad Orthop Surg 2008;16:596–607. 6. Beck M, Siebenrock KA, Affolter B, Notzli H, Parvizi J, Ganz R. Increased intraarticular pressure reduces blood flow to the femoral head. Clin Orthop Relat Res 2004;424:149–152. 7. Ogden JA. Changing patterns of proximal femoral vascularity. J Bone Joint Surg Am 1974;56:941–950. 8. Ehlinger M, Moser T, Adam P, et al. Early prediction of femoral head avascular necrosis following neck fracture. Orthop Traumatol Surg Res 2011;97:79–88. 9. Gautier E, Ganz K, Krugel N, Gill T, Ganz R. Anatomy of the medial femoral circumflex artery and its surgical implications. J Bone Joint Surg Br 2000;82:679–683. 10. Dyke JP, Aaron RK. Noninvasive methods of measuring bone blood perfusion. Ann N Y Acad Sci 2010;1192:95–102. 11. Lang P, Mauz M, Schorner W, et al. Acute fracture of the femoral neck: assessment of femoral head perfusion with gadopentetate dimeglumine-enhanced MR imaging. AJR Am J Roentgenol 1993; 160:335–341.
Dyke et al. 12. Kamano M, Narita S, Honda Y, Fukushima K, Yamano Y. Contrast enhanced magnetic resonance imaging for femoral neck fracture. Clin Orthop Relat Res 1998;350:179–186. 13. Hirata T, Konishiike T, Kawai A, Sato T, Inoue H. Dynamic magnetic resonance imaging of femoral head perfusion in femoral neck fracture. Clin Orthop Relat Res 2001;393:294–301. 14. Konishiike T, Makihata E, Tago H, Sato T, Inoue H. Acute fracture of the neck of the femur: an assessment of perfusion of the head by dynamic MRI. J Bone Joint Surg Br 1999;81:596–599. 15. Kaushik A, Sankaran B, Varghese M. Prognostic value of dynamic MRI in assessing post-traumatic femoral head vascularity. Skeletal Radiol 2009;38:565–569. 16. Boraiah S, Dyke JP, Hettich C, et al. Assessment of vascularity of the femoral head using gadolinium (Gd-DTPA)-enhanced magnetic resonance imaging. J Bone Joint Surg Br 2009;91B:131–137. 17. Garden RS. Low-angle fixation in fractures of the femoral neck. J Bone Joint Surg Br 1961;43:647–661. 18. Pauwels F. Der Schenkelhalsbruch. Ein mechanisches Problem, F. Enke, Stuttgart: 1935. 19. Tofts P. Modeling tracer kinetics in dynamic Gd-DTPA MR imaging. J Magn Reson Imaging 1997;7:91–101. 20. Papadakis SA, Segos D, Kouvaras I, Dagas S, Malakasis M, Grivas TB. Integrity of posterior retinaculum after displaced femoral neck fractures. Injury 2009;40:277–279. 21. Gojda J, Bartonicek J. The retinacula of Weitbrecht in the adult hip. Surg Radiol Anat 2012;34:31–38. 22. Wong TC, Yeung SH, Ip FK. The effectiveness of capsular decompression for internal fixation of intracapsular hip fractures. J Orthop Surg (Hong Kong) 2007;15:282–285.
