Computer Navigation for Pediatric Femoral ACL Tunnel Placement

Abstract

Background:

To compare accuracy, time and radiation exposure of pediatric femoral tunnel placement using computer navigation with a traditional freehand technique.

Methods:

A single all-epiphyseal femoral tunnel was placed in the distal femur of 20 Sawbones™ adolescent knee models. Ten tunnels were drilled using standard fluoroscopic guidance (FG). An additional 10 tunnels were drilled using 3D fluoroscopic computer navigation (CN). Both techniques aimed to match an exact point described by the quadrant system of Bernard. Time to perform the procedure was recorded as were number of single shot fluoroscopic images and approximate effective radiation doses.

Results:

The deviation from ideal femoral tunnel position was on average 6.4 ± 4.2 mm for FG tunnels and 2.7 ± 3.1 mm for CN tunnels (p<0.05) . There was no violation of the femoral growth plate using either technique. The surgeon was exposed to 17 ± 5.3 and 3 ± 0.66 single fluoroscopy exposures for FG and CN guidance, respectively (p<0.05). However, the effective dose for the CN because of the acquisition of 3D images was 0.52±.003 mSv and for FG was only 0.09mSv ± .027 (p <0.001). CN however required on average 12.5 ± 3.4 min compared to 4.6 ± 1.7 for FG (p<0.05) to complete drilling of the tunnel.

Concluson:

CN achieves a more accurate epiphyseal femoral ACL tunnel position but requires more time to complete and has a higher effective radiation dose than FG. Whether the CN ACL tunnels can translate to improved clinical outcomes is still unknown.

Level of Evidence: V

Introduction

The decision to treat conservatively or surgically reconstruct a torn anterior cruciate ligament in a skeletally immature patient remains controversial.^1-4 Fear of injury to the growth plate^2,3,5-7 has prevented many surgeons from proceeding with routine ACL reconstruction in this population. Furthermore, a prospective cohort study from Norway followed a non-operative treatment algorithm and found that ninety percent of the children were able to participate in sports at the two year follow up, with a small number of surgical operation for new meniscal injuries (13%).⁸ While nearly 40% of the children followed in this study had to decrease activity level, this study suggests that ACL deficient children can be physically active and may be an adequate treatment option in some patients.⁸ However, other studies have demonstrated that activity modification and bracing is associated with a poor prognosis and these children have a high rate of irreparable meniscal tears, chondral injury, pain and accelerated joint arthrosis.^{1-4, 9-12}

Many advocate early ACL reconstruction with specialized techniques to minimize growth plate violation.^{1-3, 6, 13-15} Treatment of an ACL tear in young children requires recognizing the patient with symptomatic instability and tailoring the surgical approach to the skeletal maturity of the patient.^2,3,16 Two main options exist for skeletally immature athletes with more then 3 years of growth remaining: all-epiphyseal reconstruction¹³ and the MacIntosh iliotibial band reconstruction.⁶ The MacIntosh surgical technique is advocated to reconstruct and stabilize the knee in the very immature athlete, but this is a non-anatomic procedure.⁶ An all-epiphyseal reconstruction technique follows many of the adult principles for ACL reconstruction with regards to tunnel placement and attempting to re-create normal anatomy.¹³ This surgical technique is technically challenging and carries a risk of injury to the physis.^13,17,18 With the growing trend to put in larger grafts, there is simply a small amount of space for accurate placement of the tunnels in small, skeletally immature patients.

The accuracy of drilling with computer navigation has been shown to be very high in the laboratory setting.¹⁹ Clinical studies also exist highlighting the precision possible with these systems, which are used for total knee arthroplasty, total hip arthroplasty, high tibial osteotomies and anterior cruciate ligament surgery.^20-23 To date computer navigation has shown ability to produce more accurate tunnel placement, but no significant functional benefit or improvement in knee stability in adult patients in many studies.^7,23-25 However, computer navigation to date has not been used or described to reconstruct ACL in the skeletally immature population. It has been described in standard ACL reconstruction as a way to enhance accuracy with placement of the femoral tunnel.²⁴ Proper tunnel position is thought to be an essential factor to a successful ACL reconstruction.²⁶ Unfortunately, there are studies in the literature that place the incidence of misplaced tunnels as high as 40%.²⁵

For the skeletally immature patient, tunnel accuracy is extremely important to avoid iatrogenic injury to the growth plate and for proper ACL function.^2,3 The purpose of this study was to compare placement of an all-epiphyseal femoral tunnel using standard fluoroscopic guided and with computer navigation. In this study, we compare the accuracy, safety, time, and radiation exposure of Computer Navigation (CN) versus fluoroscopy guided (FG) ACL femoral tunnel placement. We hypothesize that computer navigation will improve accurate placement of the epiphyseal tunnels and minimize injury to the growth plate.

Materials and Methods

20 pediatric left knee sawbones (Sawbones™, Vashon, WA) covered with foam soft tissue envelopes were used for this study. 10 knees underwent the procedure using FG for tunnel drilling with the remaining 10 using CN. All 20 tunnel drilling procedures were done by the same sports medicine fellowship trained surgeon. The study was performed in two sessions of 10 knees each. The order for the technique used for tunnel drilling was decided by the surgeon. In the first session the order was 2 FG, 6 CN and lastly 2 FN. In the second session 6 FG were performed followed by the remaining 4 CN. No set amount of time was required between tunneling attempts. The surgeon performing the procedure also obtained all the measurements.

The technique for the FG group followed the protocol recommended by Anderson.¹³ The image intensifier was brought in from the opposite side of the table. AP and lateral fluoroscopic images were obtained. The location used to define the optimal placement of the all-epiphyseal femoral tunnel was determined to be a point that is one-fourth of the distance from posterior to anterior along Blumensaat’s line and one-fourth of the distance distal from Blumensaat’s line (Figure 1).^13,27 The guide pin was then percutaneously inserted and directed toward the predefined optimal starting point with fluoroscopy in the lateral position. The C-arm was then brought into an AP view and the guide pin position verified to be below the distal femoral growth plate. The C-arm was then taken back to a perfect lateral and the pin advanced 2-3 mm with care taken to avoid anterior and posterior angulation. Additional AP and lateral views were taken as required to place the pin in the desired location. Once the position was verified, a 7mm femoral tunnel was drilled with a 7mm reamer (Arthrex™, Naples, FL) using an outside-in technique. A 7mm diameter tunnel was selected based on work of Anderson¹³ that identified the range of the quadrupled hamstring diameter in this age population between 6 and 8mm.

Figure 1.

Quadrant system as described by Bernard et al for determining location of the femoral tunnel position.²⁷ This point is approximately one-fourth of the distance from posterior to anterior along Blumensaat’s line and one-fourth of the distance down from Blumensaat’s line.¹³

Drawing by Evan P. Trupia, MD (New York, New York)

BrainLab® Curve Spine and Trauma 3D version 2.0 with 3D C-arm integration (Computer Assisted Surgery, Feldkirchen, Germany) was the computer assisted navigation platform and software used in this study. Most of the previous surgery and techniques utilizing this company’s CAS technology was for total knee arthroplasty or spine.²⁸ Two pins from the BrainLab set were drilled into the distal femur and the reflective marker balls were attached (Figure 2). The reflective markers were registered and visualized by the Brain Lab Curve platform system. The foam covered pediatric knee model complete with growth plates (Sawbones™ Vashon, WA) then underwent a multi-exposure 3D fluoroscopy spin to image the knee. During the spin, the C-arm rotated 190° around the model knee to acquire 100 single 2D images, which are processed by the attached computer platform to a 3D image set.²⁹ The navigated straight pointer was then used to confirm proper registration of structural landmarks within the knee including the medial and lateral part of the notch wall (Figure 3). The desired tunnel size of 7mm was en-A tered into the BrainLab software. The drill pin was then calibrated using the BrainLab instrument calibrator. With the use of the computer assisted surgery BrainLab system the optimal location of the femoral tunnel was determined according to the quadrant system described by Bernard²⁷ using real-time information about the anatomy of the femur and alignment of the drill (Figure 4). The tunnel was then drilled with a 7mm reamer (Arthrex™ Naples, FL). Deviation from ideal tunnel placement, distance from the growth plate, violation of the physis or blow out of the backwall, number of fluoroscopy images taken, effective radiation dose and time for completion of tunnel were recorded for the FG and CN groups.

Figure 2.

Shows the first part of the set up to use the BrainLab system. Two pins from the set were drilled into the saw bone femur and the reflective marker balls were attached. These reflective markers need to registered and recognized by the BrainLab platform system.

Figure 3.

(A) is an image using the navigated straight pointer to confirm landmarks and then in subsequent sequences (B) Probe’s eye view(C) Sagittal and (D) Axial images the pointer can be used like a drill to gauge trajectory of the femoral tunnel.

Figure 4.

Images taken from a real surgical case using computer navigation from BrainLab to place the femoral all-epiphyseal tunnel in a 10 year old male. (A) is an image from the CN used to calibrate the drill to the desired location. (B) is a view from the Probe or drill called the Probe’s eye view to localize your tunnel placement (C) and (D) are views from coronal CT slices with the projected tunnel location. The 7mm flipcutter has been calibrated by the BrainLab instrument calibrator and appears yellow in images (C) and (D).

Each sawbone was then removed of its overlying foam soft issue envelope. The ideal femoral center was determined from a perfect lateral fluoroscopic image with the superimposed quadrant system of Bernard.²⁷ The distance between the center of the ideal femoral tunnel and the center of the drilled tunnel either by FG or CN was then measured using a digital caliper (Mitutoyo 6” Absolute; Kangawa, Japan). The direction of deviation and violation of the simulated distal femoral physis were recorded. This was done by direct visualization of the knee model and then verified by fluoroscopy. Finally the distance from the outer femoral tunnel and intra-articular femoral tunnel to the nearest part of the distal femoral growth plate were also measured with the digital calipers.

Effective dose of radiation assumed each 2D fluoroscopic image of the knee to have 0.005 millisievert (mSv).³⁰ The effective radiation dose for the 3D scan was based on work of Kraus et al using the Brainlab 3D platform and software acquiring 100 2D images (0.5mSv).²⁹ Student’s t-test was used for statistical analysis and the Statistical Analysis Software v9.3 (SAS Institute Cary, North Carolina) was used to perform the analysis.

Results

The deviation from the defined optimal all-epiphyseal femoral tunnel position was 6.4 ± 4.2 mm for FG tunnels and 2.7 ± 3.1 mm for CN tunnels (p<0.05). No violation of the distal femoral growth plate occurred using either the FG or CN technique for placement of the all-epiphyseal femoral tunnel. There was no back wall blowout for either technique verified by direct visualization and fluoroscopy of the knee models. The distance from the femoral tunnel to the growth plate with the CN group was 14.9 ± 3.95 mm from the outer diameter of the tunnel compared to 13.1 ± 5.72 mm in the FG (p < 0.24.) .The distances to the growth plate from the intra-articular part of the tunnel was 17.8± 4.47 mm in the CN group compared to 17.1 ± 4.9 mm (p <0.74).

The CN group received a higher effective dose of radiation because of all the images required for the 3D reconstruction. The effective dose for the CN because of the acquisition of 3D images was 0.52±.003 mSv and for FG was only 0.09mSv ± .027 (p <0.001). However, after the 3D scan it took the surgeon only 3 ± 0.66 extra fluoroscopy exposures to drill the guide pin for the epiphyseal femoral tunnel. The surgeon required 17.4 ± 5.3 single fluoroscopy exposures using the freehand technique (p<0.05). The computer navigation group also required on average close to eight more minutes to obtain the femoral tunnel position 12.5 ± 3.4 min compared to 4.6 ± 1.7 for FG (p<0.05).

The data for each individual tunnel by technique is listed in Table 1 and 2. Comparison of tunnel data for each technique is shown graphically in Figure 5.

Table 1.

Computer Navigation Technique Tunnel Results

Knee Number	Technique Number	Deviation from ideal placement (mm)	Fluoroscopy exposures	Effective doses (mSv)	Time to complete technique (min)
3	1	0.04	3	0.515	16.58
4	2	3.07	2	0.51	9.58
5	3	1.8	4	0.52	8
6	4	6.94	3	0.515	7.4
7	5	9.79	2	0.51	11.83
8	6	0.79	4	0.52	11.5
17	7	0.59	3	0.515	15
18	8	1.65	3	0.515	14
19	9	3.02	3	0.515	16
20	10	0.05	3	0.515	15

Computer Navigation Technique Tunnel Results. Knee number is the order completed in the full set of twenty knees.

Table 2.

Fluoroscopic Guidance Technique Tunnel Results

Knee Number	Technique Number	Deviation from ideal placement (mm)	Fluoroscopy exposures	Effective doses (mSv)	Time to complete technique (min)
1	1	2.21	14	0.07	4.9
2	2	9.08	20	0.1	5
9	3	5.78	15	0.075	3.5
10	4	2.12	28	0.14	9.08
11	5	9.23	14	0.07	4
12	6	1.97	17	0.085	3.45
13	7	10.4	16	0.08	5
14	8	14	18	0.09	3.83
15	9	6.78	23	0.115	4.17
16	10	2.67	9	0.045	2.83

Fluoroscopic Guidance Technique Tunnel Results. Knee number is the order completed in the full set of twenty knees.

Figure 5.

Direct comparison of CN vs FG by order performed in each subset. (A) is distance from optimal femoral tunnel position. (B) is number of fluoroscopy exposures. (C) is effective dose and (D) is time required to complete the technique. CN data is solid bars, FG data is striped bars.

Discussion

Placement of tunnels in the correct position is one of the most essential factors for success with anterior cruciate ligament reconstruction.^25,26,31 The need for accurate tunnel placement has driven interest in computer navigation to improve consistent accurate tunnel placement.²⁴ All of the computer navigation or Computer-Assisted Orthopedic Surgery (CAOS)³² to date have explored tunnel accuracy in an adult knee. This study is the first to assess accuracy of all-epiphyseal femoral tunnel placement with CAOS in an adolescent knee model. With the rising number of ACL injuries in the skeletally immature population, more surgical reconstructions are being performed in this age group.^2,3,16 The skeletally immature athlete brings additional concerns regarding iatrogenic growth plate injury causing deformity and arrest around the knee.^2,5 In a study by Kocher et al, they looked at the results of a questionnaire given out to the members of the Herodicus Society.³³ 11% of the responding members had seen a case of growth arrest resulting from an ACL reconstruction performed in a skeletally immature athlete. Additionally a recent review paper surveying PRISM members reported 29 new cases of growth disturbance due to ACL reconstruction.¹⁸ The most common disturbance seen was closure of the lateral distal femoral physis causing a valgus deformity. The true prevalence of growth disturbance with skeletally immature ACL surgery is not known at this time. While not common, the concern for injury to the growth plate is real and reinforces the need for careful attention to technical details during ACL reconstruction in the young athlete, especially with placement of the tunnels.

This study supports computer navigation for guiding more precise placement of epiphyseal femoral ACL tunnels. From our study, CN ACL tunnels were more accurate- but also required more time. The distance from the ideal tunnel placement in this study (2.7+ 3.1 CN and 6.4+ 4.2mm FG) compared favorably to the findings in the literature of distance from the ideal position using CN as opposed to traditional ACL techniques. Picard et al, compared a randomized control study of CN and traditional technique for ACL reconstruction in 20 foam knees as well, and found the distance from ideal femoral tunnel position with CN was 2.7mm + 1.9 compared to 4.2 + 1.8 mm for the traditional method.²⁵ A group from the Netherlands looked at accuracy of computer navigation for drilling anteromedial (AM)and posterolateral bundles (PL) of the ACL and found their accuracy to be within 2.7 mm AM and 3.2mm for PL.³¹ Another advantage of this technique is the ability to have real-time guidance, allowing for correction of trajectory errors prior to complete placement of the guide pin (Figure 3). While CN ACL femoral tunnels can be more accurately placed, whether this can translate to improved clinical outcomes is still unknown.^24,32

Computer navigation brings improved accuracy and precision to the drilling of the tunnel, but what are the disadvantages of this technique? There are a few pertinent negatives to using computer navigation. First, the effective radiation dose is significantly higher with all the images required to make the 3D image for viewing. This is an almost 6 fold difference (0.09 vs 0.52 m Sv) in the effective radiation dose and is not inconsequential when dealing with a pediatric and adolescent patient population. There was also a statistically significant increase in time for the CN group, almost 8 minutes to complete the tunnel drilling. This number is consistent with previous publications on the CN set up for ACL surgery requiring a range from an extra 9 to 27 minutes.^23,34,35 While we did not measure the exact time for drilling after the 3D image reconstruction was complete, it took on average only 3 extra 2D fluoroscopic images to confirm the start point compared to 17 exposures necessary for the freehand technique. It can be inferred that the good majority of the extra time is spent with the set up, placement of pins with the reflective balls for the computer platform to recognize and the acquisition of the hundred 2D fluoroscopic images to create the 3D image. All these factors are responsible for the extra time required to use the computer navigation. Another significant weakness to computer navigation is the extra drilling into pediatric bone for the pins and reflective markers to register with the computer navigation platform.

There are several limitations worth mentioning with regard to this study. The first is that our model is not physiologic. While the soft tissue foam envelope encloses the sawbones pediatric knee model, it does not reproduce the feel of the soft tissues and the challenge to obtaining the start point with the freehand Anderson technique. Second, while this study may show improved accuracy of tunnel placement with CN, there remains no clear consensus that this accuracy translates to improved clinical outcomes following ACL reconstruction. There is a significant learning curve with regards to the set up and use of the CN so some of the increased time required to obtain the correct start point could be decreased with additional surgeon experience. In addition, newer techniques also use arthroscopic guides in combination with fluoroscopy to create and position the femoral tunnel. This was not studied. All the knee models in this study were left knees. This could have an effect on surgeon timing, skill and number of fluoroscopy exposures as this may be a dominant or non-dominant side. The computer navigation system we used in this study from BrainLab (Computer Assisted Surgery, Feldkirchen, Germany) has previously published work on total knee arthroplasty but has not been validated or published on previously with regards to anterior cruciate ligament surgery. Also, we chose to study the femur because it is more critical than the tibia but the tibia deserves future attention and study. We also did not plan for any additional tunnels to be drilled in the femur as part of a double bundle ACL reconstruction technique. We did not measure effective radiation dose with radiation badges or a thermoluminscence dosimeter (TLD), rather we used previously published numbers of effective radiation doses. Finally, computer navigation may not be available to those who wish to use it. Specifically the cost per case may be a limiting factor for using this technique. Previous studies in spine surgery and total knee arthroplasty highlighted the possible cost effectiveness of computer navigation.^36,37 However the cost of our technique in a clinical setting was not studied.

In conclusion, the use of computer navigation contributed to more accurate femoral tunnel placement for an all-epiphyseal technique. However, it did require more time to drill the tunnel then the free hand technique and was associated with a higher effective radiation dose. While the tunnel placed with CN may be more accurate then the freehand technique, whether this will translate into better clinical results is not known. Therefore it is difficult to state that CN should be used routinely for every case without postoperative follow up data.