Navigation in Musculoskeletal Oncology: An Overview.

Navigation in surgery has increasingly become more commonplace. The use of this technological advancement has enabled ever more complex and detailed surgery to be performed to the benefit of surgeons and patients alike. This is particularly so when applying the use of navigation within the field of orthopedic oncology. The developments in computer processing power coupled with the improvements in scanning technologies have permitted the incorporation of navigational procedures into day-to-day practice. A comprehensive search of PubMed using the search terms "navigation", "orthopaedic" and "oncology" yielded 97 results. After filtering for English language papers, excluding spinal surgery and review articles, this resulted in 38 clinical studies and case reports. These were analyzed in detail by the authors (GM and JS) and the most relevant papers reviewed. We have sought to provide an overview of the main types of navigation systems currently available within orthopedic oncology and to assess some of the evidence behind its use.

Introduction

On going technological advances have resulted in the development of new techniques and opportunities which have been incorporated into medical practice such as magnetic resonance imaging (MRI)-guided focused ultrasound therapy and proton beam therapy. The increased processing power of modern computers can permit the planning of complex surgical procedures such as deformity correction and tumor resection.1 The production of computed tomography (CT) and MRI scanners with more detailed, accurate imaging capabilities coupled with the development of precise intraoperative navigation technology and equipment have led to the adoption and incorporation of new techniques within the field of orthopedics. Intraoperative navigation software has been utilized within medicine for a number of years.23 Neurosurgery was first to adopt navigation into surgical practice, to map brain tumors preoperatively to determine their intracranial location and their surgical field providing more accuracy in their resections45 As a result subsequent studies have shown improved margins in neurosurgical tumor resection.678 Similarly, the adoption of navigation guided pedicle screw insertion has also been shown to reduce radiation exposure and screw malpositioning in spinal surgery.9 The benefits of the technology in this setting are obvious; providing real-time visual feedback to the surgeon working with a very small margin for error. Further surgical specialties have since incorporated navigation systems into their practice including ENT and urology.1011 General orthopedics has also embraced the use of this new technology not only in trauma1213 but also in joint arthroplasty where its use has increased substantially.14151617181920

The development of navigation systems in oncological surgery has been termed ’computer assisted tumor surgery’. Within the field of musculoskeletal oncology surgery, navigation systems have played an important role particularly within pelvic surgery, limb reconstruction, and limb salvage. Hufner and Krettek independently described the use of navigation-aided resections of pelvic tumors in 2004.23 While in 1999 Handels et al. described the virtual operation planning in orthopaedic surgery software which allowed the computer mapping of hip and pelvic tumors and the later construction of implantable allografts.21222324 The aim of surgery is to achieve resection of the tumor with clear margins to achieve a reduction in the risk of local recurrence but without sacrificing important structures and therefore function. Within the pelvis, access can be limited by anatomical constraints and, due to the extent of the tumor, intraosseous tumor margins can be difficult to appreciate. The role of navigation is to facilitate adequate surgical margins and safe resections, defined by a reduction in avoidable functional impairment. Preoperative imaging in the form of MRI, CT, and positron emission tomography-CT is performed to assess tumor placement, size, proximity to vital structures and extent of intraosseous disease. The location and involvement of vital structures then determine whether tumor resection with limb salvage is possible or whether amputation is the safest procedure. These imaging modalities can be fused to form a three-dimensional (3D) representation of the pelvis and tumor as described by Wong et al.2425 This can either be printed out in plastics or resins to form a physical model to assist in surgical preplanning (additive layer manufacturing), or it can be combined on the screen to form a visual representation. The benefit of this visual representation is that it allows detailed preoperative planning of osteotomies and their trajectories. This is of particular importance around the sacrum and posterior ilium as it can prevent the unnecessary resection of vital structures such as spinal nerve roots. Planned osteotomies can be marked on the 3D image or model to allow their visualization and preoperative rehearsal of their placement, which has been shown to improve surgical performance.26 The degree of the intraosseous disease cannot be appreciated with the naked eye, and therefore, the use of the preoperative scanning coupled with the preoperative resection planning has been shown to reduce the incidence of involved margins at resection.272829

There are two main types of navigation system currently in use in orthopedics; “image-based” and “image-less.” The two predominant techniques utilizing computer navigation within musculoskeletal oncology are patient-specific instrumentation (PSI) or intraoperative navigation. An additional form of surgery utilizing navigation technology is augmented reality as described most recently by Cho et al.30

Intraoperative navigation within oncology most often is an “image-based” system where the imaging (preoperative scan) is required to supply the software with data. This is in contrast to “image-less” systems which are more widely used in arthroplasty surgery. In this system the software is supplied with information intraoperatively during the set-up process allowing the software to calculate a patient's anatomy by registering established bone landmarks such as the tibial tubercle or tibial plateau.31 The software can then form an image of the patient's anatomy based on average appearances obtained from a large number of previous scans. The benefit of image-based systems is that it allows preoperative planning. In the technique of intraoperative navigation, the fusion of the preoperative CT and MRI images allows the tumor volume and extent to be mapped and color coded for easy identification on screen [Figures 1 and 2]. Navigation solely with MRI imaging has also been described as providing the added benefit of reduced radiation exposure to the patient.32

Figure 1

Preoperative resection templating on navigation software

Figure 2

Color coded preoperative planning of resections (green and purple) with tumor (yellow) and computer aided design implant (red)

Unlike PSI navigation, the intraoperative navigation software requires matching to the patient on the operating table. This allows the software to assimilate the information obtained from the scans and map it to the patient. This is done by marking a series of reproducible and identifiable bony landmarks on the preoperative imaging. Such points would include the anterior inferior and superior iliac spine, posterior superior iliac spine, and iliac crest tubercle. Bone landmarks are chosen as soft tissue points can move and therefore reduce the navigation accuracy. A stereotactic camera emitting infra-red light on a mobile gantry picks up signals from reflective markers on a held instrument and then displays the locations of the instruments on a screen. Adequate surgical exposure is then obtained, and the bone landmarks are registered using a navigation probe [Figure 3]. The software is then able to match the patient's landmarks touched by the probe to those previously registered on the imaging. This registration process allows the software to establish a link between the real coordinates on the patient and the virtual coordinates within the imaging data. Further surface matching is then carried out to reduce the registration error to < 1 mm. This is done by marking 100 or more points across a bone surface in different locations. Once this registration has been completed, an accurate image of the fused CT/MRI is displayed with an exact position of a hand-held probe placed on the surface of the exposed bone. This permits real-time interactive assessment of the surgeons probe in space relative to the patient. It is possible to delineate distances between bone and soft tissue elements of the tumor without the risk of accidental intralesional resection. Instruments, such as osteotomes, can be calibrated to allow an accurate visualization of the exact position of the cutting blade of the instrument in relation to the osseous component of the tumor while performing a bone cut. The use of such devices has demonstrated accurate reproduction of the planned and actual margin achieved at resection [Figure 4].3334

Figure 3

Intra-operative photograph demonstrating display screen and navigated probe and osteotome

Figure 4

Intra-operative photograph demonstrating display screen and navigated probe and osteotome

PSI requires preoperative scanning followed by production of a 3D physical model of the tumor and bony anatomy around which a precise cutting jig may be designed, also using ALM techniques. Using computer-aided design and computer-aided manufacturing (CAD-CAM) software the jig is developed to allow resection of the tumor to a predefined margin. The jig is pinned to the adjacent bone and acts as a guide for the osteotomy or bone saw. Laboratory studies have demonstrated reproducible margins when compared to the preoperative modeling without inadvertent intralesional resection.35 This was however a saw bone study that does not have the soft tissue constraints that are present in vivo. However, Gouin et al. published their results of a study of eleven patients who underwent pelvic tumor resection and found that the bone resection margins were clear in all cases, with an average error in the resection margin of 0.8 mm.36 The limitations of the use of PSI are apparent when the soft tissue extension of the tumor is taken into consideration. As the resection jig is applied for guidance of the bony resection, there is no aid to ensure an adequate soft tissue margin. As this is a static system, there is no real time imagery for intraoperative referencing as is seen with intraoperative navigation systems. Second, the time lag between the CT planning scan and the development of the patient-specific instruments may mean that the tumor has grown resulting in a mismatch between jigs and the tumor extension, which can result in intralesional resections. The cutting jigs are designed to fit to bony landmarks but can be difficult to fit accurately as a result of the soft tissue extension of the tumor or changes to the bony anatomy between the time of the scan and the time of surgery.36 Equally, operator error remains. If jigs are not applied in the exact position determined by the preoperative plan, this may again lead to intralesional or inaccurate resections.24

One advantage of navigation-assisted surgery is a perceived reduction in operative time. Wong et al. reported on a cadaveric study which investigated the time taken for resection and resection accuracy between PSI and intraoperative navigation for pelvic tumor resection.37 There was no statistical difference in the resection measurements between the two techniques, but there was a significant reduction in the time taken for the resection when using PSI. The anatomical challenges of surgery in the pelvis make accurate resections difficult. Cartiaux demonstrated that the probability of an experienced surgeon achieving a 10 mm surgical margin when working without navigation was 52%.38 While it is known that surgical margins do not affect life-expectancy, margins do predict local.3940 Jeys et al. demonstrated a reduction in intralesional margins in tumors excised from the pelvis and sacrum using navigation-assisted surgery. They showed a reduction from 29% before the use of navigation to 8.7% following the introduction of intraoperative navigation.41 This finding has been reproduced elsewhere. Young et al. demonstrated clear margins in all patients who underwent navigation assisted tumor resection not only from the pelvis but also of diaphyseal tumors.28 Cho et al. also showed clear margins were achieved in all 18 patients included in their study assessing the application of intraoperative navigation for both pelvic and metaphyseal tumors.42 Local recurrence occurred in only two patients both of whom had tumors excised from the pelvis.

Navigation has also been utilized in the field of limb reconstruction and limb salvage. Careful templating of tumors and soft tissues preoperatively can avoid unnecessary soft tissue resection and maintain function. Coupled with CAD-CAM software, allografts can be constructed that are tailor-made to fit into the precise resections that the navigation can provide [Figures 5 and 6]. These can be constructed for periarticular or diaphyseal tumors [Figures 7-10]. Li et al. have described very promising results in their use of navigation for performing complex juxta-articular resections in their limb salvage surgery with clear margins obtained in all cases both around the knee and the proximal humerus.434445 Furthermore, the use of navigation has shown to be useful in joint preservation surgery whereby tumors located in the metaphysis require accurate, precise resection to spare the joint or physis of the adjacent joint.4647 In a study of navigated chondrosarcoma excision around the knee Aponte-Tinao et al. compared the resected specimens with the preoperative planned resections and found a high level of accuracy between the two. The mean difference between the planned and actual resections was 2.43 mm.48

Figure 5

Resection specimen and computer-aided design implant

Figure 6

Navigation and computer-aided design production allows higher degree of anatomical conformity between resection and implant

Figure 7

Postoperative radiograph of pelvis with both hips anteroposterior view showing implant in situ

Figure 10

Postoperative radiograph of pelvis with both hip joints anteroposterior view showing insertion of custom implant following tumor resection

Radiograph of pelvis with both hip joints anteroposterior view showing lytic lesion in left acetabulum

Magnetic resonance imaging sagittal cut showing extent of tumor mass

The complication rate of endoprosthetic reconstruction following pelvic resections is high,4950 in which the use of precise fitting allografts or custom implants is crucial to maintain implant longevity [Figures 11-14]. Nonunion in bulk allograft reconstruction has been reported as being as high as 27%, a study by Lall et al. demonstrated that the use of navigated resections can increase contact between resected bone and allograft compared to freehand technique.51 This in theory should reduce the incidence of nonunion. Chen et al. have demonstrated a three to five-fold improvement in implant implantation precision compared when using navigation compared to conventional techniques.52 Wu et al. describe the use of their “virtual bone bank” system to improve allograft selection time and matching accuracy. Donor bone allografts are scanned as DICOM images which were then reconstructed as 3D virtual models. A navigation system was then utilized to map the patient's bone defects and facilitate the accurate implantation of the allografts.53 It is not only in the management of malignant tumors that navigation has been incorporated.54 A study by Wong et al.55 has demonstrated the use of CT based navigation with arthroscopic techniques in performing curettage of benign bone tumors of the extremities. This small study notes the benefits of minimally invasive technique and the reduction in radiation dose through lack of continual intraoperative fluoroscopy. The margins of the tumor and tumor wall can also be better appreciated intraoperatively and thus ensure a more thorough debridement. These findings are reproduced by Lee et al. who conducted a study on 8 patients with deep benign bone tumors who underwent arthroscopic curettage with a navigated burr.56

Figure 11

Radiograph of pelvis with both hip joints anteroposterior view showing large pelvic chondrosarcoma

Figure 14

Postoperative radiograph of pelvis with both hip joints following navigated internal hemi-pelvectomy with irradiation and re-implantation

Magnetic resonance imaging scanning revealing extent of tumor spread

Magnetic resonance imaging showing sciatic notch involvement

The development of augmented reality surgery described by Cho et al.30 offers an opportunity to simplify the image-patient registration process. In their study, they describe a system of navigation that relies on precise measurements of the affected bone along with the distances of the tumor margins to the bone ends. A computer generated bone model is then generated using a handheld tablet camera which takes images of the limb in various positions. A diagrammatic representation of a tumor in the bone is then generated by the software. The osteotomy through preplanned section of the bone is guided real-time monitoring with the software. They have shown an improvement in margins versus the use of conventional tumor excision techniques. However, this system can currently only be utilized in long bones and does not take into account any soft tissue tumor involvement.

There remain a number of limitations to the widespread adoption of navigation technology, particularly the use of intraoperative navigation. These include cost, increased preoperative planning time, the learning curve for development of surgical skills and the lack of evidence for long term outcome benefit. As with all technology, there is a time-dependent decrease in cost. Larger, higher volume hospitals are likely to find the acquisition of navigation equipment more affordable, especially if they can be utilized across specialties. With regards to operating time, Young found that the time for intraoperative registration was on average 30 minutes, but this decreased to 20 minutes after the fifth patient.28 Aponte-Tinao et al. found similar results in a study of 69 patients that showed that navigation added an average of 35 minutes to the operating time.57 In a meta-analysis of navigated knee replacements, it was found that the average increase in surgical time was 23% or 17 min.58 In another study looking at navigated total knee arthroplasty, it was found that the operating time was significantly longer, but this leveled off after the first 30 procedures.59 In contrast, Fafalli reported that the surgical operating time was reduced; although set-up time was not taken into account.60

The learning curve should be considered with all new technological developments and new techniques; interestingly, Fafalli did not show an appreciable difference in the learning curve of registration matching over time.

The longer-term benefits of the use of navigation in oncological surgery will only be able to be measured with time. It is only sensible to presume that surgical technologies that enhance patient safety, facilitate surgical accuracy and lead to improved patient care will become more commonplace in the future of musculoskeletal oncology surgery.

Summary Points

The navigation in musculoskeletal oncology has following advantages

(a) Optimal surgical margins and therefore reduced local recurrence. (b) Reduction in operative time. (c) Beneficial in complex pelvic or periarticular resections. (d) More accurate tumor resection and allograft implantation.

The disadvantages of navigation system are (a) Cost of equipment (b) Increase setup time (c) Learning curve.

Future potential

The future potentials of navigation system in musculoskeletal oncology are (a) Reduction in equipment costs (b) Smaller, more portable equipment (c) Potential incorporation of intraoperative robotics to further reduce surgical resection error.

Declaration of patient consent

The authors certify that they have obtained all appropriate patient consent forms. In the form the patient(s) has/have given his/her/their consent for his/her/their images and other clinical information to be reported in the journal. The patients understand that their names and initials will not be published and due efforts will be made to conceal their identity, but anonymity cannot be guaranteed.

Financial support and sponsorship

Nil.

Conflicts of interest

There are no conflicts of interest.