Modern total hip replacement is typically effective and durable, but early failures do occur. Component position influences functional outcome, durability, and risk of complications. Surgical robotics provides the detail-oriented surgeon with a robust tool to optimize the accuracy and precision of total hip arthroplasty, with the potential to minimize risk of mechanical failure. This article describes efficient workflows for using surgical robotics to optimize surgical precision without increasing surgical complexity.
Key points
- •
Component position influences functional outcome, durability, and risk of complications after total hip arthroplasty.
- •
Optimal component position requires meticulous planning based on reliable information, followed by accurate and precise execution of the plan.
- •
Surgical robotics provides the detail-oriented surgeon with a robust tool to optimize the accuracy and precision of total hip arthroplasty.
Introduction
Total hip replacement (THR) was described as the “operation of the [twentieth] century.” Although highly cross-linked polyethylene has reduced wear at the bearing surface and cementless fixation has reduced mechanical failure at the fixation interface, several attempts to improve on John Charnley’s innovation in the early twenty-first century have proved less rewarding. A trend toward cementless tapered femoral implants, cementless hemispherical acetabular implants, and cobalt-chromium or ceramic on highly cross-linked polyethylene articulations is emerging. When proven implants are used, high rates of patient satisfaction are experienced and durability longer than 20 years is anticipated.
Is there room for improvement in modern THR? Patients and payers are no longer willing to tolerate early failure after THR, but infection, dislocation, leg length discrepancy, and periprosthetic fracture continue to occur. Furthermore, edge loading, impingement, and other mechanical consequences of imprecise implant positioning continue to adversely affect implant durability for many patients. Recent Medicare data show that 10% of patients age 65 to 74 years at the time of hip replacement undergo revision surgery within the first 10 years, and recent European registry data show a 17% revision rate at 10 years for patients less than 50 years old at the time of surgery.
As surgeons try to meet ever-higher expectations, they must endeavor to embrace improvements without subjecting patients to the safety concerns that come with unproven technologies. The last 10 years have provided ample opportunity to reflect with humility on the consequences of supposedly improving THR. Nevertheless, avoidance of mechanical failure requires improvements in surgical implants and/or technique.
With surgeons, regulators, and patients now more skeptical of new THR implants, the greatest opportunities for improvement may be at the level of surgical technique. Imprecision of acetabular component position, a major source of variability in THR outcomes, presents such an opportunity.
Surgeons embarking on THR, whether simple or complex, must first establish targets for component position and then endeavor to reproduce the plan within accepted tolerances. Several investigators have proposed ranges of acceptable component acetabular position. Femoral offset and length must be informed by both femoral anatomy and the selected acetabular position so as to optimize limb length reconstruction, abductor function, joint stability, and impingement-free range of motion. Ideal implant position for a given patient may also be affected by surgical approach, soft tissue constraints, functional requirements, and extra-articular deformities such as fixed pelvic obliquity or tilt.
Planning based on plain radiographs remains limited by inability to (1) consistently control or assess magnification and (2) obtain simultaneous perfect anteroposterior (AP) images of the pelvis and proximal femur in patients with joint contractures. The imprecision of manual component positioning is well documented, and rigorous assessment of radiographic outcomes reveals that a large percentage of acetabular prostheses are implanted outside accepted parameters for optimal position.
Computer navigation was developed to improve on manual techniques and can be image guided or imageless. Although imageless navigation can improve intraoperative assessment of component position and limb length change, only image-based systems can improve surgical planning. Preoperative planning based on three-dimensional (3D) patient anatomy facilitates restoration of acetabular center of rotation and allows the ideal acetabular abduction and anteversion angle to be informed by relationships with bone anatomy. For example, patients with anteverted dysplastic acetabulae are at risk for psoas tendon impingement if the acetabular implant is not positioned within the anterior lip of the acetabular bone. Preoperative 3D planning allows the surgeon to select a position that avoids implant prominence but also avoids excessive anteversion or reaming through the medial wall of the acetabulum, technical errors that can easily occur in the service of a well-covered implant in a dysplastic acetabulum. However, navigation alone inadequately facilitates this precision because depth and location of acetabular reaming are not precisely controlled.
Surgical robotics allows the coupling of 3D planning with precision bone preparation and implant insertion. Robots have been investigated for use in joint replacement since the 1980s and used clinically since 1992. ROBODOC (Curexo Technology Corporation, Fremont, CA) was the first surgical robot developed and commercialized for THR. Although initially developed domestically by IBM, the ROBODOC active robotic system has had limited popularity in the United States and much of the published experience is from Europe and Asia. The system is approved by the US Food and Drug Administration (FDA) for THR, but has not yet been widely accepted by the domestic orthopedic community.
Widespread interest in robotic joint replacement surgery began with the commercialization of the RIO Robotic Arm Interactive Orthopedic System (Stryker Mako Surgical Corporation, Fort Lauderdale, FL) for partial knee replacement. The device has been shown to improve the precision of limb and implant alignment compared with manual techniques, but does not remove the necessity of attention to details such as cement technique. Short-term clinical outcomes have been favorable, but long-term results are not available. Software and hardware to facilitate THR were recently introduced. The robot assists with reaming of the acetabular cavity and positioning the acetabular implant using haptics, and its software package allows navigation of the femoral neck cut, leg length, and offset. The remainder of this article describes techniques for leveraging surgical robotics to optimize implant positioning for hip reconstruction, with figures to clarify the technique and illustrate the capacity of the robot to simplify complex reconstructions. The details described are specific to the widely available Mako RIO robot, but the concepts are generally applicable to other robotic platforms using image-based navigation and haptic control.
Introduction
Total hip replacement (THR) was described as the “operation of the [twentieth] century.” Although highly cross-linked polyethylene has reduced wear at the bearing surface and cementless fixation has reduced mechanical failure at the fixation interface, several attempts to improve on John Charnley’s innovation in the early twenty-first century have proved less rewarding. A trend toward cementless tapered femoral implants, cementless hemispherical acetabular implants, and cobalt-chromium or ceramic on highly cross-linked polyethylene articulations is emerging. When proven implants are used, high rates of patient satisfaction are experienced and durability longer than 20 years is anticipated.
Is there room for improvement in modern THR? Patients and payers are no longer willing to tolerate early failure after THR, but infection, dislocation, leg length discrepancy, and periprosthetic fracture continue to occur. Furthermore, edge loading, impingement, and other mechanical consequences of imprecise implant positioning continue to adversely affect implant durability for many patients. Recent Medicare data show that 10% of patients age 65 to 74 years at the time of hip replacement undergo revision surgery within the first 10 years, and recent European registry data show a 17% revision rate at 10 years for patients less than 50 years old at the time of surgery.
As surgeons try to meet ever-higher expectations, they must endeavor to embrace improvements without subjecting patients to the safety concerns that come with unproven technologies. The last 10 years have provided ample opportunity to reflect with humility on the consequences of supposedly improving THR. Nevertheless, avoidance of mechanical failure requires improvements in surgical implants and/or technique.
With surgeons, regulators, and patients now more skeptical of new THR implants, the greatest opportunities for improvement may be at the level of surgical technique. Imprecision of acetabular component position, a major source of variability in THR outcomes, presents such an opportunity.
Surgeons embarking on THR, whether simple or complex, must first establish targets for component position and then endeavor to reproduce the plan within accepted tolerances. Several investigators have proposed ranges of acceptable component acetabular position. Femoral offset and length must be informed by both femoral anatomy and the selected acetabular position so as to optimize limb length reconstruction, abductor function, joint stability, and impingement-free range of motion. Ideal implant position for a given patient may also be affected by surgical approach, soft tissue constraints, functional requirements, and extra-articular deformities such as fixed pelvic obliquity or tilt.
Planning based on plain radiographs remains limited by inability to (1) consistently control or assess magnification and (2) obtain simultaneous perfect anteroposterior (AP) images of the pelvis and proximal femur in patients with joint contractures. The imprecision of manual component positioning is well documented, and rigorous assessment of radiographic outcomes reveals that a large percentage of acetabular prostheses are implanted outside accepted parameters for optimal position.
Computer navigation was developed to improve on manual techniques and can be image guided or imageless. Although imageless navigation can improve intraoperative assessment of component position and limb length change, only image-based systems can improve surgical planning. Preoperative planning based on three-dimensional (3D) patient anatomy facilitates restoration of acetabular center of rotation and allows the ideal acetabular abduction and anteversion angle to be informed by relationships with bone anatomy. For example, patients with anteverted dysplastic acetabulae are at risk for psoas tendon impingement if the acetabular implant is not positioned within the anterior lip of the acetabular bone. Preoperative 3D planning allows the surgeon to select a position that avoids implant prominence but also avoids excessive anteversion or reaming through the medial wall of the acetabulum, technical errors that can easily occur in the service of a well-covered implant in a dysplastic acetabulum. However, navigation alone inadequately facilitates this precision because depth and location of acetabular reaming are not precisely controlled.
Surgical robotics allows the coupling of 3D planning with precision bone preparation and implant insertion. Robots have been investigated for use in joint replacement since the 1980s and used clinically since 1992. ROBODOC (Curexo Technology Corporation, Fremont, CA) was the first surgical robot developed and commercialized for THR. Although initially developed domestically by IBM, the ROBODOC active robotic system has had limited popularity in the United States and much of the published experience is from Europe and Asia. The system is approved by the US Food and Drug Administration (FDA) for THR, but has not yet been widely accepted by the domestic orthopedic community.
Widespread interest in robotic joint replacement surgery began with the commercialization of the RIO Robotic Arm Interactive Orthopedic System (Stryker Mako Surgical Corporation, Fort Lauderdale, FL) for partial knee replacement. The device has been shown to improve the precision of limb and implant alignment compared with manual techniques, but does not remove the necessity of attention to details such as cement technique. Short-term clinical outcomes have been favorable, but long-term results are not available. Software and hardware to facilitate THR were recently introduced. The robot assists with reaming of the acetabular cavity and positioning the acetabular implant using haptics, and its software package allows navigation of the femoral neck cut, leg length, and offset. The remainder of this article describes techniques for leveraging surgical robotics to optimize implant positioning for hip reconstruction, with figures to clarify the technique and illustrate the capacity of the robot to simplify complex reconstructions. The details described are specific to the widely available Mako RIO robot, but the concepts are generally applicable to other robotic platforms using image-based navigation and haptic control.
Surgical technique
The design of a surgical robot could theoretically limit the surgeon’s choice of surgical approaches. The Mako RIO robot has software packages to facilitate THR through posterior, lateral, anterolateral, and direct anterior approaches. The posterior approach is emphasized in this review, followed by changes in the workflow for the direct anterior approach.
Preoperative Planning
Segmentation
Surgical planning software accompanies the Mako RIO surgical robot. A computed tomography (CT) scan of the pelvis and both femora is performed according to a specific protocol. CT images are segmented and 3D reconstruction is performed. Bony landmarks such as the anterior superior iliac spines and the medial tips of the lesser trochanters are identified. This process is performed by engineers but can be verified by the surgeon.
Establishing reference planes and quantifying preoperative deformity
In the current Mako software (THA version 2.1), the anterior pelvic plane is determined using the tilt of the pelvis while the patient is supine in the CT scanner, and the medial-lateral axis is defined as the line connecting the anterior superior iliac spines. Acetabular anteversion is thus calculated in a functional plane. The surgeon must decide how best to use this information, because pelvic tilt can change between sitting and standing positions and between preoperative and postoperative states, but use of a functional reference is important. When anatomic planes alone are used to guide acetabular positioning, excessive functional abduction and anteversion can occur in the patient with decreased pelvic tilt, whereas functional retroversion can occur in the patient with increased pelvic tilt.
The software transforms the 3D models of the pelvis and femora into a two-dimensional image that mimics an idealized AP pelvis radiograph in the functional plane. If there is no pelvic deformity, this should result in an image with symmetric obturator foramina and the midline of the sacrum should align with the pubic symphysis. The femora are aligned so that the long axis of each femur is oriented parallel to the frontal plane, as is the maximum femoral offset. The virtual AP radiograph thus eliminates any distortion from flexion contracture and shows true femoral offset.
The software measures preoperative differences in leg length and combined offset (defined as the distance between the midsagittal plane of the pelvis and the diaphyseal axis of the femur). Measurements assume no pelvic obliquity and no length discrepancy below the lesser trochanters. The surgeon retains responsibility for determining the clinical leg length difference. Planned deformity correction should compromise between radiographic and clinical leg length discrepancy, informed by patient perceptions and clinician judgment.
Misidentification of bone landmarks can distort the image and resultant measurements (leg length, offset and acetabular version), so the surgeon must visually confirm that the reconstructed preoperative AP pelvis radiograph appears optimized. If the obturator foramina are not symmetric, the interteardrop line is not parallel to the horizontal reference line, or the sacrum is not centered on the pubic symphysis, the surgeon can identify the bone deformity or error in segmentation that distorted the image and change the landmarks if appropriate. This process can be done in the operating room and does not require the surgeon to review the plan days or hours in advance.
Acetabular templating
The procedure starts with the virtual acetabular template positioned in 40° abduction and 20° of anteversion, with the medial aspect of the template at the lateral edge of the acetabular teardrop. Size is selected to remove cartilage and sclerotic bone but preserve as much bone as possible for fixation, and to restore the center of rotation at or medial to the anatomic location. AP position is selected to remove similar amounts of subchondral bone from the anterior and posterior walls of the acetabulum. Proximal-distal position is selected to match the location of the contralateral center of rotation (when normal), and to penetrate but not remove the subchondral bone of the acetabular sourcil. The surgeon can view the planned reaming of the acetabulum on the 3D model and the planned 3D appearance of the implant within the bone, in addition to cross sections of the planned position in the axial, coronal, and sagittal planes. This technique allows identification of problems such as exposed implant at the psoas recess. Acetabular version references the functional plane, so the initial plan does not always optimally match the local anatomy. Minor alterations of component size, orientation, and position can and should be made to compromise between competing goals ( Fig. 1 ).
Templating in 3D on CT images is helpful with complex anatomy ( Fig. 2 ) but initially unfamiliar. The planning software allows the surgeon to visualize how the planned cup position will appear on the postoperative AP pelvis radiograph. This perspective is familiar and allows additional adjustments to the plan as desired.
Femoral templating
A femoral implant is chosen that will wedge within the bony confines of the proximal femur in a location that allows reconstruction of limb length and offset with the available neck and head options ( Fig. 3 ). Three-dimensional femoral planning allows accurate assessment of femoral offset and length. When coupled with 3D acetabular planning, it also allows anticipation of the changes in femoral anatomy that are necessary to compensate for changes in the acetabular center of rotation. If the acetabular center of rotation is shifted medially in accordance with Charnley principles, increased femoral offset is required to restore combined offset, optimizing stability, abductor function, and impingement-free range of motion.
For any combination of selected implant sizes and positions, the software measures the change in limb length and combined offset relative to (1) the preoperative status and (2) the contralateral hip. The surgeon must decide which of these measurements to prioritize. If the contralateral hip is normal, clinical and radiographic leg length discrepancies are equal, and the virtual AP pelvis appears optimized, it is simplest to match the contralateral side. When these variables are not aligned, the author makes a clinical assessment of desired change in limb length and combined offset and then references off the preoperative position for the surgical hip.
The software measures native femoral anteversion. When mild to moderate version abnormalities are encountered, the surgeon can compensate with small adjustments in acetabular version and/or use of an elevated rim liner. When femoral version is markedly abnormal, the surgeon can plan to use a femoral stem that allows version to be decoupled from bone morphology (eg, Wagner-style, modular, or cemented stems).
Although not part of the manufacturer’s recommended workflow, the software can be used to measure (1) the lesser-to-center distance (LTC) from the proximal aspect of the lesser trochanter to the center of the femoral head, and (2) the distance of the neck cut from the top of lesser trochanter. The preoperative LTC will be different from the planned LTC after reconstruction when the selected acetabular position does not restore the prearthritic center of rotation. Both the preoperative and anticipated postreconstruction LTC values should be measured during preoperative planning.
Preparation and Patient Positioning
For the posterior approach, the patient is positioned in the lateral decubitus position on a pegboard. The navigation tower and robotic arm are positioned anterior to the patient and opposite the surgeon, who stands behind the patient. Mobility of the robotic arm is optimal at its ideal working height. The surgical table can be lowered for the obese patient with a wide pelvis, and tilted away from the robot to facilitate acetabular reaming in adequate anteversion.
The limb is prepped and draped free, including the iliac crest and the proximal half of the femur in the sterile field. The robot is kept well away from the surgical field during exposure. This technique allows access to the patient and allows the robot to be draped and registered by the surgical technician during exposure. Standardized steps are performed, taking less than 3 minutes. This task can safely be delegated, because the software does not allow required steps to be bypassed and a 5-second final check by the surgeon is required before reaming.
Surgical Approach
For the posterior approach, a gently curved skin incision of 8 to 12 cm is made just anterior to the posterior margin of the femur. The length of the incision has little effect on patient outcome and should not be a higher priority than adequate femoral mobility to accurately prepare the acetabulum. The distal portion of the fascial incision should be anterior to the insertion of the gluteal sling at the linea aspera to avoid tethering the surgical window. In morbidly obese patients, muscular men, and patients with severe stiffness or shortening at the hip, partial or complete release of the gluteal sling is considered.
The quadratus femoris is released from the femur along with the overlying bursa, leaving a small cuff to facilitate anatomic repair. The piriformis, conjoint tendon, and obturator externus tendon are detached from their insertions on the greater trochanter and trochanteric fossa, along with the posterior capsule. A superior capsulotomy is performed along the posterior and inferior aspects of the gluteus minimus. A posterior-inferior capsulotomy is performed, with care to protect the obturator externus and associated branches of the medial femoral circumflex artery. If desired, a Steinman pin can be placed in the infracotyloid groove, perpendicular to the floor, and the location of the pin can be marked on the femur as a reference for leg length. The availability of new technology does not mean the surgeon must neglect established techniques.
Surgical Procedure
At this point, the steps of the procedure depend on the extent to which the surgeon wishes to navigate femoral preparation, leg length, and offset. The current surgical software suggests one of 2 workflows: an express workflow with limited femoral registration in which the surgeon navigates leg length and offset, and an enhanced workflow with comprehensive femoral registration that allows the surgeon to also navigate femoral version and neck cut. The express workflow is sufficient for most cases. The additional information gleaned from the enhanced workflow is helpful in the absence of easily identifiable femoral landmarks for planning the neck resection or measurement of the LTC distance.
Navigating the femur is helpful if the surgeon is unable to reliably measure the LTC or chooses a femoral implant not supported by the 3D templating software. It is preferred to navigate the femur in cases in which changes to the plan may be made during surgery, such as revisions, fused hips, and cases with acetabular bone loss. The standard express and enhanced workflows are described later, followed by an efficient cup-only workflow without femoral navigation, often used by the author for routine primary cases.
Express workflow
In the express workflow, no formal femoral registration is performed and no femoral array is required. An electrocardiogram (ECG) lead is placed at the superior aspect of the patella before skin preparation so that it is palpable through the drapes and a single checkpoint is placed on the greater trochanter. Percutaneous pins are used to secure the pelvic array to the iliac crest. Registration of limb length and offset is performed before dislocation by identifying the femoral checkpoint and ECG lead with the navigation probe. The pelvic array is used for reference, which requires that limb position be reproducible, and that the pelvic array be placed before dislocation.
The hip is dislocated posteriorly. The center of the femoral head is marked and a metal ruler is used to measure the LTC distance from the junction of the proximal aspect of the lesser trochanter to the center of the femoral head ( Fig. 4 ). There is typically a vascular foramen along the intertrochanteric ridge. Measuring the distance between this foramen and the head center provides a second point of reference and reflects the femoral offset more than the limb length. The ruler is used to mark the location of the femoral neck cut relative to the lesser trochanter as planned before surgery. The femoral neck is cut manually with a power saw and the femoral head is removed from the wound.
