Three-Dimensional Modelling and Printing in the Complex Knee
There is a rapidly growing interest in three-dimensional (3D) modelling and printing in orthopaedic surgery. Advancements in imaging technology with the advent of 3D modelling and 3D printing have revolutionised the medical field and provided physicians with powerful new tools to treat complex injuries and deformities. The knee, being regarded as one of the most complex joints in the body, has benefitted from this advancement with patient-specific models and custom instruments providing the treating clinician with detailed anatomical roadmaps specific to their patient.
Computed tomography (CT) is routinely used to produce 3D images by combining or stacking a series of 2D slices one on top of the other. Early techniques using this method produced images that estimated the bone edge, resulting in rough outlines that did not accurately depict the bone surface. Novel work has advanced 3D surface reconstruction, with imaging software being capable of applying algorithms with edge sharpening and smoothening capabilities. The result is high-quality images that more precisely represent the external surface geometry of bone. 3D shape mapping has provided the clinician with improved spatial orientation of conditions such as bone defects or loss, bony dysplasia and complex fractures.
3D printing describes a variety of techniques used for making physical objects from CT or magnetic resonance imaging (MRI) obtained graphical data. Through an additive process, successive layers of material such as metal or plastic are laid down until the object is complete. This revolutionary technology has now become far more accessible and more cost effective such that it has taken on a mainstream role in many areas of medicine. As technology advances with high-end 3D printing applications becoming available, direct 3D printing of functioning biocompatible tissue will be performed, a process known as bioprinting . These techniques are still in the early stages of development and at the time of this writing are not yet available. 3D printing is especially useful in complex cases where surgeons benefit from so-called virtual operations simulating the planned surgery. Fig. 4.1 highlights a case where 3D CT modelling was performed for a femoral condylar defect after a failed osteochondritis dissecans fixation. A virtual operation was performed using 3D CT images for planning bone cuts for osteochondral allograft implantation and fixation. Based on the 3D CT modelling, a set of surgery and patient specific tools were printed and used in the operation.
Computed Tomography in Trauma
Advances in computerised tomography have reduced data acquisition and 3D reconstruction times, allowing for faster and more accurate diagnosis and surgical planning in complex knee trauma. Multidetector CT (MDCT) is increasingly used as a first-line diagnostic modality instead of a plain film trauma series in advanced trauma life support in many centres. MDCT is able to image acutely unwell patients quickly and in nonanatomical positions where necessary to provide excellent bony and soft tissue detail. Plain film radiography has been reported to underestimate the extent of the fracture lines and subsequently risks missing fractures. Studies of tibial plateau fractures have shown surgical plans based on plain film radiography are modified in 6% to 60% of cases after CT and 21% of cases after MRI. Wicky et al. compared the diagnostic efficiency of plain film radiography and spiral CT with 3D reconstructions of 42 tibial plateau fractures. They evaluated the proposed surgical plan devised from different imaging methods for each fracture and found 3D CT was more precise and more closely matched the surgical report and postoperative radiograph. Other authors reported similar results, such as Manjula and Venkataratnam, who reported 3D CT was superior to both plain film and multiplanar 2D CT in demonstrating the spatial relationships of fracture fragments. Current evidence supports the use of fine MDCT with 3D reconstruction capabilities in complex knee trauma to facilitate accurate assessment of fracture patterns, areas of depression and displacement to assist surgical planning ( Fig. 4.2 ).
Computed Tomography Assessment of Bone Healing and Union
CT provides superior assessment of nonunion and visualisation of fracture lines compared with plain film radiography. The rapid data acquisition and high spatial resolution has seen CT used as the preferred investigation over radiography and MRI. Petfield et al. reported that the sensitivity of CT in detecting tibia fracture nonunions approached 100% and the specificity was 62%. A comparative study by Warwick et al. found MRI correlated well in the identification of spinal column fracture nonunions but that CT should be the preferred investigation for problematic or inconclusive cases. Although newer techniques in MRI sequencing have seen its use widen to include identification of occult fractures, most centres will still reserve CT for the identification of nonunion in difficult cases.
One of the difficulties in assessing bone healing is the accurate visualisation of the bony interface after either fracture fixation or osteotomy. This is especially so when healing is occurring by absolute stability where limited callus is produced. Other difficulties include the presence of metal artefact from metallic implants, impeding the value and diagnostic accuracy of CT by distorting the visualisation of bone, bone–metal interfaces and soft tissue structures. The artefacts may be present in different degrees of severity because of a variety of metals, shapes and sizes being used. Historically, metal artefact distorted the quality of CT images, rendering them difficult to interpret and of poor diagnostic value. Today, modern scanners with metal artefact reduction (MARS) capability provide images to clinicians with better resolution and less distortion.
MARS sequences are based on reduction of all its primary causes, such as beam hardening, scatter, photon starvation, noise, edge effects and the combined effect. Strategies include modifying standard acquisition and reconstruction, modifying projection data and/or image data, and applying dual-energy CT (DECT). Using modern software, postprocessing techniques such as MARS allow images previously uninterpretable to be reformatted and interpreted with high diagnostic accuracy ( Fig. 4.3 ).
Computed Tomography and Soft Tissue Injuries
Soft tissue injuries are commonly found in intraarticular fractures around the knee. The true incidence is unknown and highly variable but has been quoted to be between 52% to 99%. , Because CT, along with radiography, is a first-line investigative tool in the assessment of the severely injured knee, it may be used to identify soft tissue injuries acutely. Although most centres will rely on MRI, some authors have demonstrated the value of CT in the diagnosis of intraarticular ligament tears and avulsions in tibial plateau fractures. , They found that a smooth ligament contour without obscuration by the adjacent soft tissues was a reliable CT indicator for excluding ligament injury. Previously the thickness of the CT slices limited its accuracy, but with the advent of fine-slice CT the identification of soft tissue injuries has improved such that it may provide valuable information for the treating clinician. Mui et al. found only 2% of ligaments deemed intact on careful CT evaluation had partial or complete tears on MRI. The authors also attempted to determine whether fracture gap or articular depression on CT could be used to predict meniscal injury. Although more severe fracture gaps and depressed joint surfaces were associated with meniscal injury, no threshold could be judged and many were missed when no meniscal injury was predicted based on undisplaced fractures.
Most centres will favour MRI over CT arthrograms(CTas) in the assessment of intraarticular pathological conditions. Evidence suggests MRI provides higher interreader agreement and accuracy in assessing acute meniscal tears and chondral lesions ( Fig. 4.4 ). Traditionally, however, CTas were highly reliable and were commonly requested in patients presenting with injured knees to exclude suspected joint pathological conditions. During the 1970s and 1980s, CTa was the gold standard for the diagnosis of meniscal tears, bucket handle tears and meniscocapsular separation, with reliability reported to be 83% to 94%. With continuous rotation scanning, spiral acquisitions provide high-quality 2D multiplanar reconstructions with thin slices. Coronal, sagittal and even axial slices can detect tears that are not visible on MRI, and meniscocapsular separations based on contrast enhancement between meniscal wall and capsule. Today, some centres still prefer CTa to MRI in assessing meniscal healing after repair. Pujol et al. used CTa in assessing meniscal healing rates 6 months after repair, noting that CTa was the preferred modality because MRI was thought to be unsuitable and unreliable as a result of persistent nonspecific hypersignal within the tear ( Fig. 4.5 ). Nevertheless, contemporary approaches to suspected intraarticular pathological conditions generally favour MRI, CTa being used when MRI is contraindicated.
In assessing osteoarthritis, CTa remains useful in assessing cartilage thickness thanks to its spatial resolution and high contrast between low-attenuating cartilage and high-attenuating deep subchondral bone, with contrast material filling the gap in between. Omoumi et al. found CTa to be more accurate than MRI in evaluating cartilage thickness even on cartilage-sensitive sequences ( Fig. 4.6 ). CTa has been used in the predictive measurement of knee cartilage defects with printed 3D models, which are then used to build custom implants. Michalik et al. compared CTa with MRI in predicting the size of chondral defects using 3D modelling and a process of segmentation ( Fig. 4.7 ). CTa images were smoother and more vivid than the MRI processed images. They concluded that MRI underestimated the defect on average by 12% and CTa overestimated the defects by 3%. The popularity of CTa has nevertheless declined as MRI, which has the advantage of being noninvasive and without ionising radiation, has become more readily accessible.
Primary Anterior Cruciate Ligament Reconstruction
The ability to accurately place both femoral and tibial tunnels in anterior cruciate ligament (ACL) reconstruction is crucial to the success of the operation. Despite improvement in surgical techniques, up to 25% of patients still fail to regain satisfactory function after ACL reconstruction. Current approaches to ACL surgery favour anatomical placement of tunnels with each tunnel drilled separately for optimal positioning. The double bundle technique, requiring two tunnels in both the femur and the tibia, also aims to place the tunnels in the anatomically correct position of the native ACL. Much of this shift towards anatomical placement has relied on mapping of the true footprints of the native ACL using imaging such as 2D and 3D CT.
Purnell et al. used high-resolution CT with 3D reconstructions of eight cadavers to simulate arthroscopic, sagittal and axial perspectives of the knee. They were able to identify the relationships of bony landmarks to both the femoral and tibial origins of the native ACL. These landmarks were then identified on the simulated arthroscopic view for clinical relevance. On the femur the anterior-most aspect of the ACL origin was defined by a bony ridge on the medial wall of the lateral femoral condyle ( Fig. 4.8 ), known as resident’s ridge as described by Clancy, and the posterior and inferior borders were consistently judged to be 3 mm from the articular surface of the lateral femoral condyle. On the tibia distinct bony landmarks included the articular margin of the lateral edge of the medial tibial condyle and the base of a ridge of bone at the anterior aspect of the tibial eminence or spine for the medial- and posterior-most borders of the ACL footprint, respectively ( Fig. 4.9 ).
On the femur these bony landmarks have been shown to reliably place the tunnel close to the anatomical centre of the ACL as defined radiographically by the use of the grid method popularised by Bernard and Hertel. Bony landmarks for the tibia, however, have been less widely adopted. As opposed to the femoral tunnel, the techniques of applying, orienting and measuring the tibial tunnel from 3D CT have varied significantly in the literature. The true anatomical position that provides the best biomechanical stability is yet to be defined and agreed upon. Data suggest that a more anteriorly placed tibial tunnel relative to the centre of the native attachment site provides more stability and improves Lachman and pivot shift test results. Other authors have advised against this for fear of roof impingement and a resultant extensor lag. As a result, most surgeons will assess the position of the graft in extension and rely on soft tissue references such as the ACL remnant stump or the anterior horn of the lateral meniscus, in addition to bony landmarks, when placing the tibial tunnel.
Intraoperative 3D fluoroscopic navigation is another technique used to identify bony landmarks and optimise tunnel position. The technique uses a reference frame that is rigidly attached to the femur or tibia before an intraoperative image is taken and transferred to a navigation system. The navigation software then builds a 3D image allowing the surgeon to cross-reference bony landmarks both arthroscopically and on 3D CT in real time. The surgeon can then visualise the planned socket position and the trajectory of the tunnel on a 3D image and adjust the position until satisfied before drilling ( Fig. 4.10 ). Several authors have reported on the benefits of this technique in double bundle ACL reconstruction and in revision surgery; however, its use is limited because of the added time and cost involved. , Most centres, where available, reserve its use for technically demanding revision cases or for other surgical specialties such as spinal surgery or complex reconstructive trauma.
Revision Anterior Cruciate Ligament Reconstruction
Mapping and Tunnel Positions
Revision ACL reconstruction is a demanding procedure with inferior outcomes to primary ACL reconstruction. Meticulous preoperative planning using clinical and radiological assessment is necessary before undertaking the procedure or staged procedures. The most common cause of primary ACL failure is reported to be tunnel malposition. , Inaccurate placement of the femoral tunnel may result in a loss of flexion or an elongated graft caused by excessive traction; likewise positioning of the tibial tunnel too anteriorly may result in impingement and an extensor lag or graft rupture, whereas a too posteriorly placed tibial tunnel fails to control laxity. An essential part of the preoperative planning process includes an assessment of tunnel position, size and orientation. Plain film radiography remains the first-line investigative tool; however, CT has been shown to be more accurate in assessing tunnel position and should be considered for the surgeon planning a revision. Hoser et al. demonstrated identification of the femoral tunnel in particular was poor on plain film radiographs. They found 92% (46/50) and 84% (42/50) were invisible on x-ray images in the anteroposterior (AP) and lateral planes, respectively. In comparison CT was able to consistently identify the tunnel and allowed accurate assessment of tunnel length and width. CT has therefore been the mainstay of ACL tunnel assessment in clinical practice since it became widely available in the 1980s.
Parkar et al. compared plain film radiography and 2D MRI imaging with 3D CT for tunnel position after ACL reconstruction. Only moderate correlations were found with 2D MRI for femoral tunnel depth and AP kinematic tunnel measurements compared with CT. Modern 3-Tesla 3D MRI scanners using multiplanar reconstructible isotropic sequences now allow images to be displayed in the plane of the tunnel axis, previously impossible with the 2D MRI scanners. This has made assessment of tunnel position and size more accurate as was shown by Ducouret et al., who found a high degree of correlation between 3D CT and 3-Tesla 3D MRI. However, the authors noted that longer acquisition times with MRI increased the likelihood of motion artefact in all three planes in a multiplanar reconstruction. As a result, the authors recommended that 3D CT remain the gold standard until further advancements in MRI imaging are made.
Tunnel enlargement, or osteolysis, is a well-known phenomenon in ACL reconstruction surgery. Despite there being no reported correlation between tunnel enlargement and clinical outcomes, tunnel widening may have serious implications when undertaking revision ACL surgery. Assessing tunnel width is critical in preoperative planning for the revision ACL surgeon.
Traditionally, plain film radiography and CT and later MRI had been used to estimate the approximate width of tunnels. CT was favoured because it was fast, eliminated scaling issues present in radiographs and appeared to be less affected by geometrical factors that may influence tunnel measurement during image acquisition, such as the position of the knee or movement artefact. Although CT was widely used and accepted, the literature reported poor interobserver and intraobserver reliability. ,
With the advent of 3D CT technology, technicians use imaging software to produce 3D images and models that can accurately depict bone tunnels ( Fig. 4.11 ). One such technique was described by Getgood in 2013 where software was used to cut away segments of a formatted 3D image for more accurate visualisation of the femoral and tibial tunnels. Technology such as this has improved intraobserver reliability in assessing tunnel position and identifying potential tunnel conflict. In revision surgery, accurate assessment of the tunnel diameter is crucial and has several implications on decision making such as graft choice, bone grafting and staging of the procedure. Crespo et al. were the first to report on the accuracy of 3D CT in measuring tunnel diameter, comparing measurements with drill sizes used intraoperatively. They found the 3D CT presented a high correlation to the drill sizes used and was more accurate than 2D CT.