■ Feature Article

Acetabular Center Axis: Is It the Future of Hip Navigation?

SAM HAKKI, MD; VICTOR BILOTTA, MD; J. DANIEL OLIVEIRA, MD; LUIS DORDELLY, BS abstract

There are 2 distinct methods of cup navigation in total hip arthroplasty. One resorted to piercing the skin to improve predicts orientation of the acetabulum through bony landmarks outside the ac- their registration efforts; others resorted etabulum (eg, the anterior pelvic plane); its unreliability is well published. The to using “adjusted” APP registration other identifi es acetabular center axis (ACA) and is patient-specifi c method that in which the APP registration is calcu- is independent of pelvic tilt, making it more reliable. Data from readily pal- lated according to the change of APP pable acetabular registration points were compared with postoperative pelvic with changes of pelvic tilt. Some results computed tomography images in 137 cases. Findings show that ACA software showed fairly accurate registration and is accurate in determining acetabular/cup version and inclination. Cup center good prediction of the inclination angle axis should coincide within 4 mm of ACA to minimize impingement and maxi- and version of the acetabulum.16-18 How- mize stability without altering preoperative femoral version. ever, the new hip center could be cranial, caudal, anterior, or posterior in relation to the acetabular center, or it could be he femoral neck axis normally device (eg, a Tilt-Meter) or by imaging, medialized by implant design or to gain coincides with the acetabu- such as computed tomography (CT).1-6 A more stability. These changes made for T lar center axis (ACA), rotating classic example of the indirect method the new hip center and axis of rotation around the center of the acetabulum dur- of registering bony landmarks is palpa- must be addressed so that the preopera- ing range of motion of the hip, produc- tion of the anterosuperior iliac spines tive acetabulum/femur relationship is ing maximum stability and minimum and symphysis pubis, called the anterior maintained to avoid impingement and impingement. The ideal navigation sys- pelvic plane (APP).7-9 Other examples of instability. tem should be based on that simple fact indirect registration are (1) palpation of The direct method of registration has of reproducing the initial kinematics be- the transverse pelvic plane (TPP), which evolved in response to this need.19,20 The tween femur and acetabulum even in the relies on the posterosuperior iliac spines direct method of registration relies on severly arthritic hips because they usu- and the sacral slope adjustment (SSA), the ACA, which is patient-specifi c and ally maintain that relationship. which correlates with functional spinal Currently, total hip arthroplasty parameters in identifying the APP, and (THA) navigation relies on 2 different (2) palpation of the transverse acetabu- Drs Hakki, Bilotta, and Oliveira, and Mr 10-12 Dordelly are from the Bay Pines Medical Center, principles of registration to achieve the lar ligament (TAL). However, varia- Bay Pines, Florida. desired cup/stem relationship. The fi rst tion in thickness of subcutaneous tissue, Dr Hakki is a consultant for B. Braun Aesculap. method is the indirect method, which re- ossifi cation, movement during the regis- Drs Bilotta and Oliveira, and Mr Dordelly have no lies on digitalization of bony landmarks tration process, and anatomic variations relevant fi nancial relationships to disclose. Correspondence should be addressed to: outside of the acetabulum that defi ne the of acetabular version among healthy in- Sam Hakki, MD, PO Box 22429, St Petersburg, frontal pelvic plane by simple palpation dividuals have resulted in major errors FL 33742. or by ultrasound, or aided by a mechanical in cup orientation.13-15 Some clinicians doi: 10.3928/01477447-20100510-52

OCTOBER 2010 | Volume 33 • Number 10/SUPPLEMENT 43 ■ Feature Article

The ACA registration was done by 3 3 Points Acetabular Rim REGISTRATION points of palpation of the superior rim, 3 points of the most anterior rim, and 3 points of the most posterior rim of the acetabulum (Figure 1). The superior Superior POINT point must be chosen consistently. We Anterior determined the superior point of the ac- POINT etabulum as the point of transection of a Posterior line from the prominent iliac tubercle to POINT the center of transverse acetabular liga- One Point Registration ment (acetabular notch) (Figure 2). The Base of the fovea definition 1 2 software averages the points (Figure 1) and maps out the orientation of the ace- Figure 1: Three points are needed for ACA software Figure 2: The superior point is the most important to calculate the acetabular center and center axis. point to obtain to ensure consistency. This point is not tabulum and its center, which determines One more point is needed at the base of the fovea at the 12-o’clock position, but rather is more poste- the ACA. This axis guides the surgeon to alert the surgeon during reaming that the reamer rior. The line from iliac tuberosity to mid-TAL touches where to position the reamer for place- is at the medial wall when the ACA is zero. the superior point of the acetabular rim. ment at the center of the acetabulum. The surgeon has the option of choosing the independent of variations in anatomy or comparable with CT in its accuracy in desired acetabular version or inclination pelvic position, and relies on readily ac- determining the inclination and ver- angle to accommodate individual patient cessible anatomic landmarks of the ace- sion angles of the acetabulum and cup anatomy while maintaining the center of tabulum. This method takes into account implant. The new hip center was Ͼ4 the acetabulum if possible when ream- the center of rotation of the acetabulum mm outside the desired ACA in only 4 ing. The center of the acetabulum is cal- and the new hip (cup) center. (The soft- of 135 patients; all 4 patients reported culated by the computer program as zero. ware for this method of acetabular/cup impingement symptoms. None of the Any deviation is calculated in millime- navigation is called ACA software.) remaining patients had dislocation or ters, whether directed medially, anteri- This system relies on reproducing the impingement. orly, posteriorly, caudally, or cranially preoperative hip center axis and adjust- (Figures 3, 4). We planned to be within ing the hip center, especially in dys- MATERIALS AND METHODS 2 mm from the center of acetabular axis plastic or protrusio hips. Therefore, the This prospective study compares the (Figure 5). In this way the preoperative version of the stem need not be changed ACA registration data with postoperative version of the neck of the femur could be because the axis of the femoral neck pelvis CT scan in anterolateral intermus- maintained without fear of impingement (version) will automatically coincide cular minimally-invasive computer-as- or excessive combined version of cup and with the ACA because the acetabular/fe- sisted THA. Since March 2007, 137 con- stem (Figures 4, 6). Stem navigation fol- mur axis relationship was not altered by secutive patients underwent OrthoPilot lowed ACA, and a femoral/patella/foot ACA software. Furthermore, changing navigated press-fi t Excia (B. Braun Aes- plane was established to ensure proper the version of the stem would throw the culap, Tuttlingen, Germany) or Metha leg navigation (Figure 7). Here, 3 points foot into internal or external deformity THA system (B. Braun Aesculap). Two are required for leg length and offset: 1 compared with its preoperative position. were excluded for lack of complete data. at any marked point (by Bovie or methy- Therefore, the navigation of the stem Patient age ranged from 34 to 86 years lene blue) of the greater trochanter, 1 at is limited to achieving the desired leg (mean, 65 years). One hundred sixteen a marked point at the lateral epicondyle length and offset. (The software of navi- patients had primary osteoarthritis; 14 (marker pen or methylene blue), and 1 at gation of the cup and the stem is called had avascular necrosis; 5 had rheuma- a marked point at the ankle. The surgeon ACA-PLUS.) toid arthritis. Mean body mass index was may use a small temporary K-wire or just Since March 2007, we collected data 32.2 kg/m2. Dysplastic acetabulum was drill a hole in the lateral epicondyle as from ACA software registration of mini- present in 12% of patients, and protru- a point of navigation for leg length with invasive computer-navigated THAs and sio was present in 4%. Patients were fol- the hip in neutral position. compared those data with postoperative lowed up for a minimum of 1 year with Preoperative plain pelvis radiographs CT images of the pelvis in 137 cases. postoperative radiographs, CT scan, and with anteroposterior and cross-table lat- Findings showed ACA software to be Harris score assessment. eral views were obtained as routine. They

44 ORTHOPEDICS | www.ORTHOSuperSite.com ACETABULAR CENTER AXIS FOR HIP NAVIGATION | HAKKI ET AL

3 4 Figure 3: The circled −4∘ refers to reaming at 4∘ less than the shaded 46∘ ac- Figure 4: The reamer should be within 2 mm of the ACA. Here, it is 1 mm etabular inclination. The circled +8∘ refers to reaming at 8∘ more anteversion caudal and 1 mm posterior. We medialized the reamer by 4 mm, but we did not than the shaded 14∘ acetabular version. The ACA diameter at the lower corner breach the base of the fovea. Note the red column reading of zero. accurately measures the size of the acetabulum. Reaming need not be more than 4 mm larger than 51 mm.

The large head choice will give a zero lengthening and 4-2= +2 offset

5 6 7 Figure 5: ACA software has identifi ed the Figure 6: The actual cup version will be Figure 7: Three points are required for leg length and offset: new hip center and will compare it with the 46∘+1∘=47∘, whereas actual cup version will 1 at any marked point (by Bovie or methylene blue) of the original center of acetabulum (see Figures be 14∘−2∘=12∘ anteverted in an anterolateral greater trochanter, 1 at a marked point at the lateral epicon- 3 and 4). It is from this point that the true approach. dyle (marker pen or methylene blue), and 1 at a marked point leg length will be calculated, as will ver- at the ankle. sion. Cup center is identifi ed for measure- ment of offset and leg length (within 2 mm of acetabular center axis). were used to template the expected cup anatomic acetabular version (ie, con- the reliability and statistical accuracy size, depth, and orientation to compare trol) was 21.2Њ (SDϮ4.3Њ) by CT com- of ACA software in identifying the ver- with acetabular anatomy. Postopera- pared with 20.01Њ (SDϮ4.8Њ) by ACA sion of the cup implant (Pϭ.98). We tive CT measurements of acetabular and software. Mean anatomic acetabular then divided the patients into 3 groups cup inclination and version angles were inclination was 47.56Њ (SDϮ7.0Њ) by according to anatomic variations of observed independently using special CT compared with 47.4Њ (SDϮ8.7Њ) the acetabulum: normal, protrusio, and software. Data were compared using the by ACA software. These results reflect dysplastic. In the first group, the size Fisher test and Student t test. The level the validity and reliability of ACA soft- of the cup closely matched the size of of signifi cance between CT and the vari- ware in identifying the version and in- the acetabulum (normal acetabulum). able should approach 1.0, which means clination of the acetabulum. The center The anatomic cup version was 21.7Њ the variable is as good as CT, and when of acetabulum was maintained within 2 (SDϮ10.2Њ) by CT compared with PϽ.05, the variable is signifi cantly infe- mm in 130 cases. Six cases were Ͼ2 21.4Њ (SDϮ8.6Њ) by ACA software rior to CT imaging. mm off center of the acetabulum; of (Pϭ.97). these, 4 were off center by Ͼ4 mm. Cup In the second group, the size of the RESULTS implant version was 22.97Њ (SDϮ5.4Њ) cup implant was smaller than the ace- Data for 135 of 137 consecutive pa- by CT compared with 23.0Њ (SDϮ4.4Њ) tabulum (as in protrusio hips). The ana- tients were eligible for analysis. Mean by ACA software. This result reflects tomic cup version was 22.0Њ (SDϮ9.9Њ)

OCTOBER 2010 | Volume 33 • Number 10/SUPPLEMENT 45 ■ Feature Article

by CT compared with 22.8Њ (SDϮ9.5Њ) When using ACA software methods, that can be used for all patients.21 Be- by ACA software. In the third group, the safety zone of Lewinnek was main- cause of the wide range of inclination the size of the cup was larger than the tained and no serious complications and anteversion figures, half of cases in acetabulum (as in dysplastic hips or the were related to implant or software use. the study were outside the safety zone cup implant was larger by choice). The recommended by Lewinnek.22 This is anatomic version was 26.3Њ (SDϮ7.1Њ) DISCUSSION another reason why we believe that pa- by CT compared with 26.5Њ (SDϮ6.4Њ) The APP has been the cornerstone of tient-specific direct software is superior by ACA software. The ACA software image-based hip navigation technolo- to other indirect methods of registra- was as accurate as CT (Pϭ0.96), where- gies. Ultrasonography, skin piercing tion, because it determines the orienta- as APP software in a previous study20 technique, registration with the patient tion of the normal acetabulum and the was less accurate (Pϭ.04). As for the in supine position, and adjusted APP ACA that is patient-specific.19 inclination angle of the cup implant, the software calculations are examples of Is there a limitation to the useful- mean anatomic cup inclination angle attempts to improve the accuracy of ness of the ACA software if there is for all groups was 42.5Њ (SDϮ4.2Њ) by APP registration to identify the orien- a pathologic variation of the acetabu- CT compared with 40.5Њ (SDϮ3.5Њ) by tation of the acetabulum (indirectly). lar anatomy (protrusio, dysplastic, or ACA software (Pϭ.93). Planes other than APP have been in- large osteophytes)? Will placing the Similarly, we divided the patients troduced, for example, TPP, SSA, and cup implant in the ACA of the acetabu- into 3 groups to compare cup inclina- TAL, and they all predict the inclina- lum recreate the original deformity and tion. In the fi rst group, the cup matched tion and version of the acetabulum in- thus misplace the cup implant? That the acetabulum: the anatomic cup incli- directly and share similar problems in question was resolved before the study nation angle was 42.7Њ (SDϮ3.6Њ) by registration through skin and subcuta- began by obtaining standard pelvis an- CT compared with 43.1Њ (SDϮ4.7Њ) by neous tissue or ossified ligaments. Ad- teroposterior and cross-table lateral ACA software. Again ACA software was ditionally, they do not identify the ACA radiographs preoperatively. If the ac- accurate (Pϭ.73). In the second group, nor do they attempt to recreate the rela- etabulum is determined to be dysplas- the cup size was smaller than the acetab- tionship of soft tissue envelope tension tic or the desired cup is larger than the ulum (representing protrusio). The in- and direction as the ACA software does. acetabulum, then templating gives an clination angle was 42.6Њ (SDϮ4.0Њ) by Acetabular center axis identification is approximate inclination angle of the CT compared with 46.8Њ (SDϮ6.6Њ) by a direct method of measurement that is acetabulum and the cup. In protrusio, ACA software, which again was accurate easy to register, with no subcutaneous the acetabular inclination angle (Ͻ40Њ) (Pϭ.92). In the third group (represent- fat to consider, and it is independent of is expected to be smaller than cup in- ing dysplastic), the cup size was larger pelvic and spine positioning. clination angle, whereas in dysplastic than the acetabulum, the anatomic cup The reliability of the OrthoPilot cases (Ͼ50Њ), the acetabular inclination inclination angle was 46.0Њ (SDϮ4.8Њ) navigation system we used has been angle is expected to be larger than cup by CT and 42.70 (SDϮ12.2Њ) by ACA tested, and special ACA software reg- inclination angle. The difference in de- software (Pϭ.45). istration points were developed and the grees is calculated and used intraopera- data compared with postoperative CT tively to adjust the computer-navigated Clinical Results of the pelvis as a control. Computed ACA measurement to the desired angle There were no dislocations, and leg tomography of the pelvis revealed sig- while maintaining the proper acetabu- length was maintained within Ͻ8 mm in nificant variation of normal acetabulum lar/cup center axis. In the extreme rare all cases. Offset was maintained within version anatomy ranging from 5Њ to 30Њ case of preoperative subluxating hips, 5 mm less than established preopera- of anteversion, whereas the anatomic we recommend preoperative CT imag- tively. In 2 cases, offset was allowed to CT inclination angle range was less ing to determine the degree of instabil- be Ͻ5 mm larger. variable (47ЊϮ8Њ). This puts the safety ity and whether bone augmentation is Impingement pain was noted in range zone of Lewinnek in question. For ex- needed to restore the acetabular center. of motion (mainly adduction/fl exion) in ample, a cup version of 29Њ is outside Adjustment of version of the cup and 4 patients. Interestingly, ACA was off by the safety zone of Lewinnek when one offset should resolve the instability Ͼ4 mm in all 4 of these patients. Mean presumes the cup should be 18Њ of ver- as long as the preoperative acetabular Hospital for Special Surgery scores were sion. One study demonstrated that there center is considered. However, when 92.5Ϯ4.6 with preoperative scores of is perhaps no ideal position for the cup the ACA software reads the final cup 36.7Ϯ11.2. (45Њ inclination and 20Њ anteversion) position as ϩ1Њ (Figure 6), then the cup

46 ORTHOPEDICS | www.ORTHOSuperSite.com ACETABULAR CENTER AXIS FOR HIP NAVIGATION | HAKKI ET AL

is 1Њ inside the acetabulum. If the ace- the femoral head/neck when performing 11. Rousseau MA, Lazennec JY, Boyer P, Mora N, Gorin M, Catonné Y. Optimiza- tabulum shows protrusio, then the com- THA. That includes the inclination/ver- tion of total hip arthroplasty implantation: puter readings of inclination should be sion angles of the cup, the offset, and is the anterior pelvic plane concept valid? in the “ϩve” (positive) range, which soft tissue tension (leg length), without J Arthroplasty. 2009; 24(1):22-26. means that the cup will be smaller altering the version of the stem. Calcu- 12. Archbold HA, Mockford B, Molloy D, McConway J, Ogonda L, Beverland D. than the acetabulum. If the acetabulum lated adjustment is also readily made in The transverse acetabular ligament: an aid is dysplastic, then the computer read- protrusio and dysplastic hips in ACA to orientation of the acetabular component during primary total hip replacement. J ing of the cup inclination should be software. Bone Joint Surg Br. 2006; 88(7):883-886. in the “Ϫve” (negative) range (Figure 13. Pinoit Y, May O, Girard J, Laffargue P, 3), which means that the cup is larger REFERENCES Ala Eddine T, Migaud H. Low accuracy than the acetabulum or that some of the 1. Jaramaz B, DiGioia AM, Blackwell M, of anterior pelvic plane to guide the posi- tion of the cup with imageless computer rim of the cup is outside the acetabu- Nikou C. Computer assisted measurement of cup placement in total hip replacement. assistance: variation of position in 106 lar superior rim. This enables the sur- Clin Orthop Relat Res. 1998; (354):70-81. patients. Rev Chir Orthop Reparatrice Ap- par Mot. 2007; 93(5):455-460. geon to make the proper adjustment in 2. Tannast M, Langlotz U, Siebenrock KA, grossly abnormal acetabula to achieve Wiese M, Bernsmann K, Langlotz F. Ana- 14. Wolf A, Digioia AM, Mor AB, Jaramaz B. Cup alignment error model for total hip the ideal cup position specific to each tomic referencing of cup orientation in total hip arthroplasty. Clin Orthop Relat arthroplasty. Clin Orthop Relat Res. 2005; patient. However, no such adjustments Res. 2005; (436):144-150. (437):132-137. are needed for the stem version. 3. Spencer JM, Day RE, Sloan KE, Beaver RJ. 15. Parratte S, Pagnano MW, Coleman-Wood Basically, there are 2 techniques Computer navigation of the acetabular K, Kaufman KR, Berry DJ. The 2008 component: a cadaver reliability study. J Frank Stinchfield award: variation in (direct and indirect) for referencing of Bone Joint Surg Br. 2006; 88(7):972-975. postoperative pelvic tilt may confound the navigated THA in the lateral position. accuracy of hip navigation systems. Clin 4. Kieffer H, Othman A. Ultrasound vs pointer Orthop Relat Res. 2009; 467(1):43-49. Our new reference axis (ACA) has the palpation based method in THA naviga- 16. Dorr LD, Hishiki Y, Wan Z, Newton D, advantage of being patient-specific and tion: a comparative study. Orthopedics. 2007; 30(10 suppl):153-156. Yun A. Development of imageless com- independent of variations in anatomy puter navigation for acetabular component 5. Nakamura S, Matsuda K, Arai N, Kobayashi or pelvic position. The system relies on position in total hip replacement. Iowa M, Wakimoto N, Matsushita T. Method to Orthop J. 2005; 251-9. readily accessible anatomic landmarks reduce variations of inclination angle of 17. Dorr LD. Hip Arthroplasty: Minimally Inva- of the acetabulum, making it signifi- the acetabular component during mini- incision hip arthroplasty. J Orthop Sci. sive Techniques and Computer Navigation. cantly attractive for surgeons who use 2006; 11(3):254-28. Philadelphia, PA: WB Saunders; 2006. CT-free planning and navigation. Im- 6. Ecker TM, Murphy SB. Application of 18. Dorr LD. Acetabular cup position: the im- pingement through ACA software is surgical navigation to total hip arthro- perative of getting it right. Orthopedics. 2008; 31(9):898-899. easily predicted by keeping the cup plasty. Proc Inst Mech Eng [H]. 2007; 221(7):699-712. 19. Murray DW. The definition and measure- center axis within 4 mm from zero (the 7. DiGioia AM, Jaramaz B, Blackwell M, et al. ment of acetabular orientation. J Bone normal preoperative ACA) (Figures 3, Image guided navigation system to mea- Joint Surg Br. 1993; 75(2):228-232. 4). The stem follows the same version sure intraoperatively acetabular implant 20. Hakki S, Dorrdelly L, Oliveira D. Com- alignment. Clin Orthop Relat Res. 1998; of the neck of the femur as preopera- parative study of acetabular center axis (355):8-22. vs anterior pelvic plane registration tech- tively; consequently, no change in foot 8. DiGioia AM, Jaramaz B, Plakseychuk AY, et nique in navigated hip arthroplasty. Or- rotation is made. The length and offset al. Comparison of a mechanical acetabu- thopedics. 2008; 31(10 Suppl 1):ii. is controlled by the modularity of the lar alignment guide with computer place- 21. Laffargue P, Pinoit Y, Tabutin J, Giraud F, ment of the socket. J Arthroplasty. 2002; Puget J, Migaud H. Periprosthetic frac- stem to ensure the femoral neck/im- 17:359-364. tures around total hip and knee arthroplas- plant neck distance is measured from 9. Jolles DM, Genoud P, Hoffmeyer P, et al. ty. Mortality, morbidity and prognostic the new hip center that the computer Computer-assisted cup placement tech- factors of periprosthetic femoral fractures following hip arthroplasty: multicentric has calculated. niques in total hip arthroplasty improve accuracy of placement. Clin Orthop Relat prospective assessment of 115 cases. Finally, navigation through identifi - Res. 2004; (426):174-179. Rev Chir Orthop Reparatrice Appar Mot. 2006; 92(4):316-325. cation of ACA may be the future focus 10. Dandachli W, Richards R, Sauret V, Cobb of THA navigation because it is capable JP. The transverse pelvic plane: a new 22. Lewinnek GE, Lewis JL, Tarr R, et al. and practical reference frame for hip ar- Incidence of dislocation following hip ar- of reproducing the preoperative rela- throplasty. Comput Aided Surg. 2006; throplasty for patients. J Bone Joint Surg tionship between the acetabulum and 11(6):322-326. Am. 1978; 60(2):217-220.

OCTOBER 2010 | Volume 33 • Number 10/SUPPLEMENT 47