Computer-Assisted Orthopedic and Trauma Surgery

Abstract

Background

There are many ways in which computer-assisted orthopedic and trauma surgery (CAOS) procedures can help surgeons to plan and execute an intervention.

Methods

This study is based on data derived from a selective search of the literature in the PubMed database, supported by a Google Scholar search.

Results

For most applications the evidence is weak. In no sector did the use of computer-assisted surgery yield any relevant clinical or functional improvement. In trauma surgery, 3D-navigated sacroiliac screw fixation has become clinically established for the treatment of pelvic fractures. One randomized controlled trial showed a reduction in the rate of screw misplacement: 0% with 3D navigation versus 20.4% with the conventional procedure und 16.6% with 2D navigation. Moreover, navigation-assisted pedicle screw stabilization lowers the misplacement rate. In joint replacements, the long-term results showed no difference in respect of clinical/functional scores, the time for which the implant remained in place, or aseptic loosening.

Conclusion

Computer-assisted procedures can improve the precision of certain surgical interventions. Particularly in joint replacement and spinal surgery, the research is moving away from navigation in the direction of robotic procedures. Future studies should place greater emphasis on clinical and functional results.

cme plus

This article has been certified by the North Rhine Academy for Continuing Medical Education. Participation in the CME certification program is possible only over the Internet: http://daebl.de/RY95.

The deadline for participation is 19 November 2021.

Computer assisted orthopedic surgery (CAOS) comprises a range of computer-assisted applications aiming to improve precision and reduce the invasiveness and radiation exposure of surgical interventions. Computer assisted surgery links the patient’s anatomy with surgical instruments or implants, either on the basis of imaging or in an imageless approach.

The techniques can be categorized under different aspects:

• Intraoperative 3-D imaging

  • 3-D fluoroscopy (C-arm and full orbital rotation systems)

  • Robotic 3-D-C-arm systems

  • Intraoperative computed tomography (CT)

• Intraoperative navigation

  • Based on preoperatively collected imaging data: CT based navigation, 3D printed template-guided navigation

  • Based on intraoperatively collected imaging data: 2-D fluoroscopy navigation, sonography; 3-D navigation (3-D fluoroscopy, intraoperative CT)

  • Imageless navigation

• Robotic surgery

  • Active

  • Semi-active

  • Passive systems

Method

Navigation

A navigation device consists of a navigation camera, a reference base, and a workstation, as well as specific surgical instruments (e1).

Distinction is made between the following object tracking technologies:

• Electromagnetic tracking

• Optical tracking.

Electromagnetic systems are mainly used in craniofacial navigation and consist of a signaling device that is inserted into a site, plus several coils that detect the electromagnetic field. For the purposes of navigation in orthopedic and trauma surgery, optoelectronic systems have become the main approach (e1).

Optoelectronic navigation is based on referencing the anatomical region and associating the real anatomy with the x-ray anatomy in a virtual matrix (e2-e4). To this end, a reference base studded with ball markers is attached to the bony structures that are about to be operated on (e4). The workstation of the navigation system is then used to determine the position within the space on the basis of an analysis of the infrared signal emitted by the stereotactic camera after it hits the ball markers of the reference base and subsequent reflection.

In principle, the available software packages allow for anatomy independent navigation, depending on the associated imaging modalities, but over the years, relevant standard procedures have come to dominate. Figure 1 shows an example of how this is used in spinal surgery.

Robotic surgery

Robots are commonly understood to be technical apparatuses that take over mechanical tasks as instructed by humans. Distinction is made between differently intelligent classes of robots, from manual manipulators to simple handling devices to computer-assisted, sensor equipped, independently anticipatory systems. Surgical robots have been the subject of development for more than 20 years (2).

The requirements for using robotic assistance in a sensible way depends in particular on the objective of the surgical procedure. Surgical robots are specifically developed depending on their intended function. Surgical robots are constructed as macro-micro-manipulators in resective surgical disciplines like urology or abdominal surgery, to allow the transformation of blunt hand movements into subtle hand movements (e5).

A clear distinction needs to be made between these robots and orthopedic robots for the implantation of joint replacements. In principle, the joint (for example, the knee) is palpated or scanned by using CT/MRI (magnetic resonance imaging) in order to develop a computer based 3-D model (e6). A prosthesis is virtually planned on the model, and the necessary bone cuts are planned precisely. By using a navigation system, this model is then linked intraoperatively with the real situation. With conventional navigation systems, this implementation is done by manual bone cutting, whereas the surgical robot is able to perform some of these steps (figure 2) (3– 5).

Figure 2.

Example of knee replacement surgery in primary osteoarthritis (a, b) with robot-assisted surgery. A 3-D model of the individual anatomy is set out by using computed tomography scanning. This enables virtual prosthetic planning with an exact selection of size. Furthermore, resection levels and all 9 degrees of freedom of the prosthetic joint are planned in advance (c). Intraoperatively, joint mobility and the individual patient’s ligamentous tension are documented, and the desired knee laxity planned subsequently. The components can be moved virtually in 0.5 mm steps. Subsequently the plan is implemented by using a semi-autonomous haptic robotic system, with the surgeon guiding the saw blade suspended from the robotic arm.

Results

Navigation

Overall, healthcare delivery by means if navigation assisted procedures is based on results with a low evidence level.

• Prospective randomized controlled trials (RCTs) exist only for (Table 1a and b)

• Navigation procedures for osteosynthesis in posterior pelvic ring fractures (6)

• Navigation in pedicle screw insertion (7, 8)

• Replacement of the shoulder, hip, and knee joint (10– 22, E7-e12)

• Reconstruction of the anterior cruciate ligament (e13)

• Corrective osteotomy of the proximal tibia (23).

Table 1a. Outcomes of navigated trauma surgery.

Reference	Study design	Details	Comment
3-D navigation: osteosynthesis methods: pedicle screw osteosynthesis
Laine et al. Eur Spine J 2000 (7)	RCT	50 conventional versus 41 navigated pedicle screw osteosynthesis procedures of the thoracic spine/lumbar spine; 10.1 % medial or inferior pedicle screw perforation in conventional surgery group versus 0.4 % in navigated surgery group; medial or inferior perforation in 40 % of patients in conventional versus 2.4 % in navigated surgery group	Standard procedure, regular use, especially upper thoracic spine and cervicothoracic transition or in special questions such as ankylosing spinal disorders, occasional deformities
Rajasekaran et al., Spine 2007 (8)	RCT	Deformities thoracic spine; 16 patients (236 pedicle screws) versus 17 patients (242 screws); 54/236 pedicle perforations (23 %) in conventional versus 5/242 (2 %) in navigated surgery
Villard et al., Spine 2014 (9)	RCT	RCT: 11 conventional, 10 3-D navigated stabilizations of the thoracic spine/lumbar spine; radiation exposure; patient dose 1 884 cGy*cm2 in conventional and 884 cGy*cm2 in navigated surgery group (non-significant); reduction in radiation exposure (µSv) for the surgical team about 5-fold in the eye area, 10-fold in the chest area, and 6.5-fold in the forearm area
3-D navigation: osteosynthesis methods: screw placement in injuries of the posterior pelvic ring (sacroiliac screw osteosynthesis)
Matityahu et al., JOT 2014 (6)	RCT	58-times conventional, 18-times 2-D navigated, 54-times 3-D navigated procedure; screw misplacement 20.4 % (10/58) in conventional surgery group, 16.,6 % (3/18) in 2-D navigation and 0 % (0/54) in 3-D navigation groups	Standard procedure, used regularly
2-D navigation: corrective osteotomies of the proximal tibia or distal femurs
Akamatsu et al., Arthroscopy, 2016 (23)	RCT	RCT: corrective osteotomy proximal tibia; 62 patients, 31 conventional, 31 navigated surgical groups (combination: CT based and imageless); 24 months of follow-up; navigated operations: longer surgery time 100 versus 83 minutes; less radiation exposure 5.3 versus 3.6 minutes; smaller loss of tibial slope 0.2° versus 1.6°; more precise sagittal osteotomy angles 0.1° versus 12.1; no change: total axis, femorotibial angle, tibial posterior slope, lateral tibial fracture and functional scores (Lysholm Score and American Knee Society Scores)	Clinical use in individual centers
2-D navigation: navigated anterior cruciate ligament replacement surgery
Cheng et al., Knee 2012 (32)	RCT	Meta-analysis of RCTs: included were 151 conventional, 155 navigated procedures; no differences regarding knee stability (Lachman Test, Pivot-Shift Test) and functional assessment (IKDC-Score, Lysholm Score, Tegner Score)	No improvement as a result of navigation; no relevant clinical use

Additional RCTs: Plaweski 2006 (e13), Mauch 2007 (30), Chouteau 2008 (e26), Hart 2008 (e27), Endele 2009 (31)

CT, computed tomography; RCT, randomized controlled trial

Table 1b. Outcomes for navigated procedures in total joint replacement surgery.

Reference	Study type	Details	Comment
Navigation: total hip or knee replacement surgery
Kim et al., JBJS, 2012 (17)	RCT	RCT: mean follow-up 10.8 years; 520 conventional versus 520 navigated total knee arthroplasties; longer surgery time (88 versus 76 minutes) for navigated surgery; no differences in clinical functional scores (Knee Society Score, Range of Motion, WOMAC Score, UCLA Activity Score); radiography results for coronal and sagittal planes did not differ; no difference in the implant‘s lifespan (99.2 % placed conventionally and 98.8 % by navigated surgery)	Numerous RCTs regarding total knee replacement surgery, some regarding navigated total hip replacement surgery; increasing clinical availability and use, especially for total knee replacement arthroscopy; finally rivalry with robotic approaches
Parratte et al., CORR 2016 (11)	RCT	RCT: 30 conventional versus 30 navigated total cement free hip replacement operations; 10 years follow-up; functional outcome: Hip Disability and Osteoarthritis Outcome Score, Harris Hip Score (96+/- 5 conventional; 95+/- 6 navigated), or SF-12. Ni differences were observed in implant lifespans regarding aseptic loosening. Dislocation events occurred in neither of the groups.
Navigation: total shoulder replacement surgery
Kircher et al., JSES 2009 (10)	RCT	RCT: 10 conventional, 10 CT based, navigated implantations of the glenoid component for total shoulder replacement surgery. Navigated surgeries take more time: 169.5 versus 139 minutes. Increased precision of component placement with correction from 14.4° auf 7° retroversion conventionally and 15.4° to 3.7° in navigated surgery; 4/10 patient in the navigated group had navigation stopped because of technical problems	To date no relevant clinical use/application

Additional RCTs: Sparmann 2003 (e11), Bäthis 2004 (19), Decking 2005 (18), Kalteis 2006 (e10), Spencer 2007 (16), Harvie 2012 (20), Hetaimish 2012 (22), Huang 2012 (15), Pang 2011 (e9),

Lützner 2013 (e8), Cip 2014 (14), Lass 2014 (21), Weber 2017 (e7), Xu 2014 (12), Renkawitz 2015 (13);

CT, computed tomography; RCT, randomized controlled trial; SF, Short Form;

Numerous studies have been reported in the context of trauma related indications, which indicate improved precision regarding the identification of points, placement of wires or screws, or even the re-positioning and axis control for navigated approaches than for conventional methods in the laboratory model or in human specimens (e14– e18, 24).

Initial clinical observational studies and case series have found indications that support these results (6, 25, 26, e19– e22), especially for corrective osteotomy of the tibia (26), pedicle screw insertion, and pelvic osteosynthesis.

While clinical observational studies of treatment for femoral fracture by means of medullary pin/rod in the initial years provided indications of a possible advantage for navigation regarding torsion and length (27, e23– e25), a retrospective analysis with a larger number of patients did not indicate any improvement to torsion and length, nor radiation exposure (28). For this reason, 2-D navigation has not found its way into routine clinical practice for stabilizing the long bones. Studies of higher evidence levels are currently not available.

For treating fractures of the posterior pelvic ring, a prospective randomized study has shown radiological advantages for 3-D navigation over 2-D navigation and conventional methods (6). Rates of misplaced screws were 20.5% for conventional approaches, 16.6% when 2-D navigation was used, and 0% for 3-D navigation. This approach of 3-D navigation assisted sacroiliac screw osteosynthesis has become the method of choice in clinical practice and is used regularly in many centers (etable 1) (6).

eTable 1. Results for navigation.

3-D navigation	Reference (first author, year, source)	Details	Comment
3-D navigation: osteosynthesis
Navigated osteosynthesis methods on limbs (eg, screws, wires)	Stöckle (2003) (e60), Hüfner 2003 (e61), Leung 2004 (e62), Rübberdt 2009 (e63), Leung 2010 (e64), Stübig 2013 (e65), Stübig 2013 (e66)	–	No level I or level II studies; regular use in clinical practice
Stabilization surgery of the spine (pedicle screw osteosynthesis)	Laine et al., Eur Spine J. 2000 (7)	RCT: 50 conventional versus 41 navigated cases of pedicle screw osteosynthesis thoracic spine/lumbar spine; 10.1 % medial or inferior pedicle screw perforation in conventional versus 0.4% in navigated surgery group; medial or inferior perforation in 40 % of patients in conventional versus 2.4 % in navigated surgery group	Standard procedure, regular use especially for the upper thoracic spine and cervicothoracic transition or in special questions, such as ankylosing spinal disorders, occasionally deformities
	Rajasekaran et al., Spine 2007 (8)	RCT: deformities thoracic spine; 16 patients (236 pedicle screws) versus 17 patients (242 screws); 54/236 pedicle perforations (23 %) in conventional versus 5/242 (2 %) in navigated technique
	Villard et al., Spine 2014 (9)	RCT: 11 conventional, 10 3-D navigated stabilization procedures thoracic spine/lumbar spine; radiation exposure; patient dose 1 884 cGy*cm2 conventional and 884 cGy*cm2 navigated surgery group (non-significant); reduction in radiation exposure (µSv) for surgical personnel about 5-fold in the eye area, 10-fold in the chest area, and 6.5-fold in the forearm area
	Further studies: Laine 1997 (e67), Kendoff 2007 (e68), Nakashima 2009 (e69), Silbermann 2011 (e70), Schouten 2012 (e71), Shin 2012 (e72), Kim 2014 (34), Bredow 2015 (e73)	–	No level I or level II studies
Screw placement in injuries of the posterior pelvic ring (sacroiliac screw osteosynthesis)	Matityahu et al., JOT 2014 (6)	RCT: 58 conventional, 18 2-D navigated, 54 3-D navigated procedures; screw misplaced 20.4 % (10/58) in conventional surgery group, 16.6 % (3/18) in 2-D navigation and 0 % (0/54) in 3-D navigation groups	Standard procedure, used regularly
	Further studies: Stöckle 2003 (e60), Citak 2006 (e74), Kendoff 2007 (e68), Rübberdt 2009 (e63), Behrendt 2012 (e75), Peng 2013 (e76), Zwingmann 2013 (e77), Richter 2016 (e78)	–	No level I or level II studies
Percutaneous screw placement in acetabular fractures	Hüfner 2004 (e79), Kendoff 2007 (e68), Lin 2008 (e80), Ochs 2010 (e81), Ruan 2012 (e82), Schwabe 2014 (e83), Wong 2015 (e84)	–	No level I or level II studies, regular use described in individual centers
Screw osteosynthesis in scaphoid fractures	Kendoff 2007 (e68), Citak 2010 (e85), Smith 2012 (e86)	–	No level I or level II studies; no relevant clinical use/application to date
Stabilization methods of the shoulder girdle	Gras 2013 (e87), Stübig 2013 (e65), Stübig 2013 (e66)	–	No level I or level II studies; clinical use only in individual centers
3-D navigation: drilling in osteochondral lesions and osteonecrosis, biopsies
Multiple localizations possible: eg, acetabulum, head of femur, distal femur, talus	Kendoff 2007 (e68), Kendoff 2005 (e88), Kendoff 2003 (e89), Citak 2007 (e90)	–	No level I or level II studies, but regularly used in certain centers
2-D navigation: methods to control length, alignment, and torsion of the limbs
Medullary pin/rod osteosynthesis of femur fractures	Leung 2004 (e62), Kendoff 2007 (e68), Gardner 2008 (e91), Gösling 2009 (e92), Frank 2010 (e93), Hawi 2012 (e94), Keast-Butler 2012 (e95)	–	No level I or level II studies; results inconsistent; currently altogether rarely used in clinical practice
Corrective osteotomy of the proximal tibia or distal femur	Akamatsu et al., Arthroscopy, 2016 (23)	RCT: corrective osteotomies of the proximal tibia; 62 patients, 31 having conventional, 31 navigated surgery (combination: CT based and imageless); 24 months follow-up; navigated operations: time taken 100 versus 83 minutes; less radiation time 5.3 versus 3.6 minutes; lower loss of tibial slope 0.2° versus 1.6°; more precise sagittal osteotomy angles 0.1° versus 12.1°; no change: total axis, femorotibial angle, tibial posterior slope, lateral tibial fracture and functional scores (Lysholm score and American Knee Society scores)	Clinical use in individual centers
	Further studies: Keppler 2004 (e96), Leung 2004 (e62), Hankemeier 2006 (e97), Kendoff 2007 (e68), Gebhard 2011 (e98)	–	No level I or level II studies
Osteosynthesis in proximal femur fractures	Leung 2004 (e62), Kendoff 2006 (e99), Chong 2006 (e100), Liebergall 2006 (e101), Weil 2007 (e102), Müller 2012 (e103), Wilham 2011 (e104)		No level I or level II studies; No relevant clinical use/application
Navigated anterior cruciate ligament replacement	Plaweski et al., AJSM, 2006 (e14)	RCT: 30 conventional, 30 navigated surgery; mean follow-up 24 months; no differences in laxity (IKDC Laxity Score) in 200 N: 1.5 mm conventional versus 1.3 mm navigated, but less variability in laxity in navigated surgery group, also more precise placement of drill channel in sagittal position of the tibia in navigated surgery (ATB measurement−1.2 conventional versus 0.4 navigated); no differences for femoral canals	No improvement as a result of navigation; no relevant clinical use/application
	Mauch et al., AJSM, 2007 (30)	RCT: 24 conventional, 29 navigated; radiological comparison of tibial tunnel placement; no differences between navigated and conventional procedure
	Chouteau et al., AOTS, 2008 (e26)	RCT: 36 conventional, 37 navigated; center of drill channel deviated from planned: conventional 7 ± 1.5 mm, navigated: 2.5 ± 1.1 mm; no functional differences at 24 month follow-up (IKDC Score 89.5 conventional, 89.7 navigated); knee stability (KT 1000): 8.9 mm conventional; 9.5 mm navigated
	Hart et al., Arthroscopy, 2008 (e27)	RCT: 40 conventional, 40 navigated; 28 months follow-up; no difference in laxity measurement (KT 1000, IKDC Laxity Test); functional scores identical (IKDC, Lysholm Score); radiologically more precise placement of femoral tunnel
	Endele et al., Arthoscopy, 2009 (31)	RCT: 20 conventional, 20 navigated; at 24 month follow-up: no differences in tunnel placement and functional scores (IKDC, Tegner, Lysholm Score)
	Cheng et al., Knee, 2012 (32)	Meta-analysis RCTs: altogether included: 151 conventional, 155 navigated; no differences regarding knee stability (Lachman Test, Pivot Shift Test) and functional assessment (IKDC Score, Lysholm Score, Tegner Score)
	Further studies: Picard 2001 (e105), Kendoff 2006 (e106), Pearle 2008 (e107), Kendoff 2009 (e108), Pearle 2009 (e109), Lee 2012 (e110)	–	No level I or level II studies
2-D navigated stabilization in vertebral/spinal fractures	Holly 2003 (e111), Izadpanah 2009 (e112), Shin 2012 (e72), Njoku 2016 (e113)		No level I or level II studies; very little use in clinical practice
2-D navigated SI screw osteosynthesis	Matityahu et al, JOT 2014 (6)	See 3-D navigation	Made obsolete by 3-D methods, use no longer relevant
	Further studies: Stöckle 2001 (e114), Grützner 2002 (e115), Mosheiff 2004 (e116), Briem 2007 (e117), Zwingmann (2013) (e77)	–
Acetabulum osteosynthesis	Stöckle 2004 (e118), Stöckle 2007 (e119), Gras 2008 (e120), Hong 2010 (e121)	–	No level I or level II studies; use described in individual centers
Patient specific implants, 3-D pressure
Control of length, alignment, torsion	Bruns 2019 (e122), Krettek 2019 (e123), Liodakis 2019 (e124)	–	No level I or level II studies; relatively new approach
Tumor resections and prosthetic replacement in tumors	Hüfner 2004 (e125), Krettek 2004 (e126), So 2010 (e127), Cartiaux 2013 (e128), Wong 2014 (e129), Sternheim 2015 (e130)	–	No level I or level II studies; individual use in specialist centers
Navigation: total joint replacement surgery
Total replacement surgery for hip or knee	Sparmann et al., JBJS Br, 2003 (e11)	RCT: 120 conventional versus 120 navigated total knee arthroplasty, of which 16/120 with > 3° varus or valgus in the mechanical leg axis after conventional and 0°after navigated surgery; no difference in sagittal tibial component placement (slope); general surgical complications the same 6/120 conventional, 5/120 navigated)	Numerous RCTs regarding total knee arthroplasty, some regarding navigated total hip replacement surgery; increasing clinical availability and use, especially for total knee arthroplasty; rivalry with robotic approaches
	Bäthis et al., JBJS Br 2004 (19)	RCT: 80 conventional versus 80 imageless navigated total knee arthroplasty; improved frontal plane alignment (2.1° conventional versus 1.5°); 74/80 (94 %) navigated surgery patients and 69/80 (86 %) within varus/valgus alignments of +3° to −3°
	Decking et al., J. Arthroplasty, 2005 (18)	RCT: 25 conventional versus 27 navigated total knee arthroplasty; improved reconstruction of the mechanical leg axis in 22/27 navigated versus 20/25 conventional surgery patients (1.5 ° varus ± 2.1 navigated versus 2.3° ± 3.5° conventional); no differences in functional scores after 3 months (WOMAC, Knee Society Score)
	Kalteis et al., JBJS Br, 2006 (e10)	RCT: navigated total hip replacement surgery; 30 conventional, 30 CT based, 30 imageless navigation; better placement of acetabular component in CT based and imageless navigation versus conventional implantation (25/30; 28/30; 14/30 correct angles with 40° inclination ± 10° and 15° anteversion ± 10°); increased duration of surgery versus conventional with 8 minutes for imageless navigation and 17 minutes CT based navigation
	Spencer et al., JBJS Br, 2007 (16)	RCT: total knee arthroplasty. 36 conventional versus 35 imageless navigated; 2 year follow-up; no changes in functional results: Knee Society Score (158.9 vs 156.4), WOMAC Score (13.6 versus 23.4), Oxford Knee Score (20.1 versus 26.7) or Bartlett Patellar Score (23.8 versus 23.0)
	Harvie et al., J. Arthroplasty, 2012 (20)	RCT: 5 years of follow-up in total knee arthroplasty. 22 conventional and 24 navigated implantations; improved frontal alignment femoral and tibial; no functional differences (Knee Society Score, WOMAC, SF-36) between conventional and navigated surgery
	Hetaimish et al., J. Arthroplasty, 2012 (22)	Meta-analysis of 23 RCTs for conventional versus navigated Implantation of total knee arthroplasty; In the conventional group the total leg axis deviated in 30.1 % (349/1 160) at more than 3° varus/valgus versus 12.8 % (158/1 234) in the navigated group. The same results applied to the femoral and tibial component alignment. No conclusions regarding functional scores.
	Huang et al., J Arthroplasty, 2012 (15)	RCT: 5 year follow-up; 44 conventional versus 46 imageless navigated total knee arthroplasty procedures; improved frontal alignment with ± 3° varus/valgus for 42/46 patients in the navigated surgery group versus 27/44 patients in the conventional surgery group
	Pang et al., KSSTA, 2011 (e10)	RCT: 70 conventional and 70 navigated total knee arthroplasties with soft tissue balancing. 2 years of follow-up; the mechanical axis showed fewer deviations of more than 3° in the navigated surgery group(10 % versus 21 %). Oxford Knee Questionnaire Score showed improved results in the navigated surgery group (16.4 versus 19.1), no differences SF-36
	Kim et al., JBJS, 2012 (17)	RCT: mean follow-up 10.8 years; 520 conventional versus 520 navigated total knee arthroplasty operations; longer surgery time (88 versus 76 minutes) for navigated surgery; no differences in clinical-functional scores (Knee Society Score, Range of Motion, WOMAC score, UCLA activity Score); radiographic results for coronal and sagittal planes showed no differences. No differences in the implants‘ life spans (99.2 % conventional and 98.8 % navigated)
	Lützner et al., KSSTA, 2013 (e8)	RCT: 5 years of follow-up; 33 conventional versus 34 imageless navigated total knee arthroplasty procedures; improved frontal alignment with ± 3° varus/valgus 31/34 after navigated surgery versus 24/33 patients the conventional surgery group; no functional differences regarding Knee-Society Score (149.0 versus 150.2 navigated), Knee Score (36.1 versus 42.9 navigated) Function Score (65.5 versus 66.5 navigated) or Euroquol CAS Scale (61.7 versus 66.8 navigated)
	CIP et al., J Arthroplasty, 2014 (14)	RCT: minimum follow-up 5 years for total knee arthroplasty; 87 conventional versus 87 conventional; no differences in implant lifespans (95.4 % conventional, 98.9 % navigated); 4.6 % revision rate after conventional versus 1.1 % after navigated surgery; deviations less pronounced in the mechanical leg axis at the frontal plane of 1.7° navigated to 2.4° after conventional surgery; still higher deviations in tibial slope (5.3° conventional, 6.5° navigated) after conventional surgery, improved clinical outcome for Insall Knee Score in navigated surgery group; improved HSS Knee Score in navigated surgery group, no difference in WOMAC Score
	Lass et al., J Arthroplasty, 2014 (21)	RCT: 63 conventional versus 62 navigated total hip replacement operations; mean follow-up 1.5 Jahre; duration of surgery in navigated group longer (104.2 versus 122.3 minutes); no changes in functional scores (Harris Hip Score, WOMAC Score); radiologically no changes in inclination (37.7° conventional versus 38.6° navigated); more precise setting of anteversion – 19.5° ± 4.6° navigated versus 17.3° ± 10.4° conventional; also the navigated surgery group showed fewer gross deviations of the acetabular inclination – 7.9 % (5/63) conventional and 0 % for the navigated surgery group; also fewer outliers regarding antetorsion angle setting for the navigated surgery group 6/62 (9.7 %) compared with conventional placement – 23/63 (36.5 %); furthermore no differences in leg length between groups
	Weber et al., CORR, 2017 (e7)	RCT: 62 conventional fluoroscopy controlled versus 53 navigated total hip replacement operations; no changes in overall leg length or total offset; in the navigated surgery group the difference in leg length was smaller, at 1.8 ±0.2 mm than the 3.5 mm ±0.2 mm for conventional surgery; measurements were taken with differences of less than 5 mm in the leg length and less than 88 mm offset difference in 93 % and in 95 % in the navigated and 54 % and 95 % in the conventional fluoroscopy controlled group
	Xu et al., IJS, 2014 (12)	Meta-analysis RCTs: total hip replacement surgery: differences were seen between the rate of acetabular placement outside the safe zone in the conventional surgery group; further the surgery took longer in the navigated group than in the conventional surgery group; the difference in leg length was reduced after navigated surgery; no differences were found regarding acetabular inclination, anteversion, incidence of dislocations, or deep vein thromboses of the lower limb
	Renkawitz et al., Bone & Joint J, 2015 (13)	RCT: 69 conventional versus 66 navigated total hip replacement operations; improved potential relevant movement range in everyday life (> 110° flexion, 30° extension, 45° external rotation in 0° flexion, 30° internal rotation in 90° flexion, 50° abduction and 30° adduction) in 86 % of navigated and 66 % conventional surgery patients; no clinical-functional differences (Hip Osteoarthritis Outcome Score, Harris Hip Score, Euro-Qol 5D, Manucuso THA Expectations Score) after 6 months or 1 Jahr
	Parratte et al., CORR, 2016 (11)	RCT: 30 conventional versus 30 navigated implantations of cement free hip replacements. 10 years of follow-up; functional outcome: Hip disability and Osteoarthritis Outcome Score, Harris Hip Score (96 ± 5 conventional; 95 ± 6 navigated), or SF 12. No differences were seen in lifespans as regards aseptic loosening. No dislocation events in either group.
	Further studies: u. a. Kelley 2009 (e131), De Steiger 2015 (e132), Wasterlain 2017 (e133), Bohl 2019 (e134)	–	No level I or level II studies
Total shoulder replacement surgery	Kircher et al. JSES, 2009 (10)	RCT: 10 conventional, 10 CT based, navigated implantations of the glenoid component for total shoulder replacement surgery; navigated operations take longer: 169.5 versus 139 minutes; imp[roved precision of component placement with correction of 14.4° to 7° retroversion for conventional surgery and 15.4° to 3.7°for navigated surgery; 4/10 patients in navigated surgery group had navigation stopped because of technical problems.	No relevant clinical use/application to date
	Further studies: Briem 2011 (e135), Stübig 2013 (e136), Theopold 2016 (e137)	–	No level I or level II studies

CT, computed tomography; RCT, randomized controlled trial; SI, sacroiliac

In trauma surgery, studies of higher evidence levels are needed, as are meaningful and valid studies that investigate clinical-functional results.

Regarding stabilization in spinal surgery, several RCTs have shown a greater precision for navigation assisted pedicle screw placement—in the sense of a lower rate of misplacements (7, 8). This approach is used especially for posterior stabilization of the cervical spine, the cervicothoracic transition, pelvic stabilization, and certain uses in ankylosing spinal disorders and surgery for scoliosis of the spine (29, e22).

For the navigation assisted treatment of ruptures of the anterior cruciate ligament, promising descriptions transpired at the beginning, but subsequent RCTs showed no great improvement of the radiological results (30, 31, e13, e26, e27). Furthermore, no relevant improvement of knee stability or functional scores was achieved. For this reason, this method is used very rarely these days. The situation for 2-D based approaches to control axis, length, and torsion in corrective osteotomy of the proximal tibia is similar. Several retrospective analyses or prospective cohort studies initially described beneficial advantages for the procedure (26), but an RCT did not confirm these results (23).

Several studies of evidence levels I and II exists regarding orthopedic joint replacements. For navigation assisted knee replacement surgery, the literature contains more than 30 prospective RCTs (14– 22, e7– e11); for hip replacement surgery there are more than 10 RCTs (11– 13, 21, e10). Study results for knee replacement surgery showed an improved leg alignment at the frontal plane and a lower number of gross malalignments of more than 3 ° varus or valgus when navigated approaches were used. For hip replacement surgery, a more precise placement of the acetabulum components and a lower rate of gross deviations for inclination and antetorsion were observed. For both modalities (total knee and hip replacements), no differences were found at the 10 year follow-up regarding of rates of aseptic loosening. For hip replacements, no difference was seen in the rate of dislocations (11, e7). Both navigated modalities have in common that long-term results do not indicate any differences in the clinical-function scores either.

For shoulder replacement surgery, only one prospective randomized study exists of the implantation of conventional total shoulder replacement, with a small number of cases (10). In addition to the fact that the surgery took longer, improved precision of component placement was seen for the glenoid component. However, dropout rates of 40% were described, owing to technical problems. Thus far, no results exist regarding functional results; furthermore, the method is used only rarely.

Level III studies found a reduction in intraoperative radiation exposure for navigated procedures (9, 28, 33, e22), especially for the surgical staff (9, 33, e28). An RCT of spinal surgery with fluoroscopically assisted stabilization of the lumbar spine showed a mean trend towards a lower patient dose of 1884 cGy*cm² for conventional procedures and 884 cGy*cm² in the navigated group; in particular, the study found that radiation exposure (µSv) was up to 9.96 times lower for the surgical personnel—with 25 µSv for a navigated approach and 270 µSv for the conventional technique (9).

Regarding a cost analysis of navigated techniques, no valid conclusions can be drawn because of the complexity and susceptibility to different factors of influence. Some economic aspects have been described in the literature, however (box).

BOX. Economic aspects of navigated procedures.

• High initial financial outlay for 3-D imaging and navigation devices (e29, e30)

• Flat learning curve during implementation with extended time spent operating and associated opportunity costs (35, e31)

• Cost per use between 300 and 780 € (e32, e33) depending on the healthcare system

• Potential savings as a result of avoiding revision surgery, which is not remunerated (e29, e30)

• No relevant economic remuneration for the additional surgical effort (e34, e59)

Robotic surgery

Joint arthroplasty robots are articulated arm robots which have undergone different stages of autonomy. The first system deployed in a clinical setting was an exclusively autonomous, modified industrial robot (36). Minimal planning errors were actively corrected by the robot during the procedure. When these systems became widely used, they quickly displayed a higher complication rate than manual surgery. In particular, they incurred intraoperative fractures and soft tissue injuries. Such systems are nowadays still in use in Asia.

Semi-active systems are an alternative. A combination of visual controls and semi-autonomous saw/surgical fraise function with options for restricting their activity provides the greatest possible safety, in the sense of a “haptic robot” (e35).

In particular the virtual limitation of the saw cut, which can be pre-set on the control computer, makes it possible to limit the amount of trauma inflicted on the soft tissues and protecting the ligaments from being cut (e36). As the adjustment of the position of the prosthesis can be planned, effects on the soft tissue tension can be directly visualized and monitored (e37). Such prospective simulations are not possible when traditional manual knee implantation techniques are used.

Contrary to the classic surgical experience that innovations are associated with occasionally huge learning curves (e38), robotic systems show minimal learning curves, which speaks in favor of the robotic systems available today (37, e39). The implant can often be correctly positioned at the first attempt (e40, e41). Especially for unicondylar sledge prostheses, more precise positioning has been described for prostheses that were implanted by a robotic system (e42). Comparison studies between robot assisted systems and conventional implantation of sledge prostheses have shown improvements to the reconstruction of the joint line for robotic systems (e43, e44).

One advantage of such robotic systems consists of the fact that iatrogenic soft tissue injuries can be prevented by limiting the workspace of the robotic saw cuts virtually (e45). Because of the improved soft tissue balancing, robotically implanted sledge prostheses produce a more natural gait pattern than manually implanted prostheses (e46). Furthermore, a more rapid return to sports activities has been documented (e47). Further to a multitude of studies that showed the superiority or equivalence of robot assisted sledge prosthesis implantation (e35, e42, e45– e50), initial results from the Australian National Joint Replacement Registry showed lower rates of revisions and non-septic complications after three years (e51). Only for septic complications within the first three months did the registry indicate poorer outcomes for robot assisted prosthetic implantation (e51). With regard to the revision rate, the registry did not show any difference between robot assisted implantation of sledge prostheses and a particular manually implanted model (0–9 months: hazard ratio 1.14; 95% confidence interval [0.71; 1.83]; p=0.596; ≥ 9 months: HR 0.66; [0.42; 1.02]; p=0.058). Compared with all other manually implanted sledge prostheses, the implant delivered by a robotic system was found to have a significantly lower revision rate after three years (HR 0.58; [0.42; 0.79]; p<0.001) (e44). Rates of revision surgery on the basis of aseptic loosening were lower for prostheses implanted by using a robotic system than for all other non-robot-assisted procedures (total time period: HR 0.34; [0.17; 0.65]; p=0.001), but not when compared with a certain manually implanted model (e44). Revision rates on the basis of an infection were significantly higher in the group treated with robot assisted surgery than in the comparison groups (ZUK, Zimmer unicondylar knee): total time period: HR 2.91; [1.22; 6.98]; p=0/.016); other, non-robot-assisted unicondylar knee arthroplasty (UKA): 0–3 months: HR 5.57; [2.217; 14.31]; p<0.001) (e44).

Altogether 22 retrospective cohort studies of unicondylar knee replacement have been published to date, as have three case series, four RCTs, and one case-control study (e52).

For total knee replacement surgery, initial cohort studies have described better short term results compared to conventional total knee arthroplasty. Particularly quicker early rehabilitation, lower blood loss, and better short term knee function have been described (38). These effects are explained with the more precise position of the implant compared with manual implantation (e41) and more precise soft tissue balancing compared with manual approaches (e53). An in vitro study showed significantly better pressure distribution and lower pressure peaks in the knee joint across the whole range of movement for robot-assisted surgery versus manually executed total knee replacement surgery. To date, nine studies have been published on the commonly used robot-assisted systems used in knee replacement surgery, which reported patient outcomes. RCTs and long term results are lacking (39).

The current systems are used mostly in knee replacement surgery and hip replacement surgery (38, 40, e54). The technologies are suitable not only for joint surgery but also for corrective osteotomy and spinal surgery (e55– e58) (etable 2).

eTable 2. Orthopedic robotic systems and their areas of use.

System	Joint	Cutting and controlling principles	Platform	Model base
Tcat/TSolution (Nachfolger RoboDoc) Think Surgical, Fremont, CA, USA	Hip: mills medullary cavity femur Knee: mills resection areas	Autonomous milling by means of a robotic -arm guided ball mill, computer controlled limitation/restriction, surgical control through emergency off-switch	Open	CT
MAKO Stryker, Mahwah, NJ, USA	Hip: acetabular milling, leg length navigation Knee bicondylar: cuts complete resection areas Knee unicondylar: mills/ cuts complete resection area	Semiautonomous bone saw or mill: oscillating bone cutter with thick-cut cutting blade suspended from robotic arm, ball mill suspended from robotic arml Computer controlled limitation, surgeon permanently in control of the switch	Closed, only implants by the manufacturer	CT
NAVIO Smith + Nephew, Memphis TN, USA	Knee bicondylar: mills distal resection area Knee unicondylar: mills complete resection area	Semiautonomous mill: hand tool with ball mill, which is drawn away by the robot Computer controlled limitation, surgeon permanently in control of the switch	Closed, only implants by the manufacturer	Surface palpation
OMNIBOT OMNI, Raynham, MA, USA	Knee bicondylar: mills complete femoral resection area	Mini-robotic-arm positioned saw jig – manual cutting process Computer controlled ligament balancing tool	Open	Surface palpation
ROSA Zimmer Biomet, Warsaw, IN, USA	Knee bicondylar: cuts complete resection area Spine: positions drilling/screwing tool	Robotic-arm-positioned saw jig/drilling jig – manual cutting/drilling process	Closed	Optional: imageless or radiograph assisted (CT)
Excelsius GPS Globus Medical	Spine: positions drilling/screwing tool	Robotic-arm-positioned drilling jig – manual cutting/drilling process	Closed	Image based
Orthotaxy DePuy Synthes	Spine: positions drilling/screwing tool	Robotic-arm-positioned saw jig – manual cutting process	Closed	Image based
Mazor X Stealth Medtronic	Spine: positions drilling/screwing tool	Robotic-arm-positioned saw jig – manual cutting process	Closed	Image based
Pulse NuVasive	Spine: positions drilling/screwing tool, navigates drilling depth	Robotic-arm-positioned drilling jig, navigated depth of drilling/screwing	Closed	Image based

CT, computed tomography

Compared with adverse safety experiences (soft tissue injury owing to the system working autonomously) in earlier systems (2), today’s systems are of adequate technical robustness and are therefore suitable as models for future applications. Compared with autonomous systems, today’s systems provide a combination of virtual control and semi-autonomous cutting/milling function with limitation options.

As robot assisted surgery may become established in the future, the role of the surgeon is set to change. If as expected the robotic system performs surgery at a constant quality level, the expectation of success may rise, but the responsibility for the surgery will partly lie with the technology. It should not be forgotten that the overall surgical result rests only to a small extent with the manual execution. The majority consists of induction and anticipation of the operation—that is, patient selection, basic planning, and detail planning. Today’s robotic systems do not provide any additional intelligent services that absolve surgeons from planning concepts. On the contrary—the surgeon has to master the complexity of virtual planning, the machine takes over only the execution and therefore does not prevent planning errors.

At the same time the technology provides enormous evaluation potential for the future. Documenting many surgical parameters—such as balancing the knee joint—provides an opportunity to correlate these parameters in future with the results reported by patients (patient reported outcome measures, PROMS). The surgeon of the future can then be presented with sensible recommendations, for example regarding the individual positioning of the implant (e58).

Conclusions

Computer-assisted methods provide a multitude of options in orthopedic and trauma surgery, which can support the surgeon in planning and executing a procedure. In principle it was shown that no universally valid conclusions can be drawn about the use of CAOS systems across the entire spectrum of orthopedic surgery and trauma surgery. In order to achieve improved outcomes by using computer-assisted approaches, the learning curve should be completed and the technique applied regularly.

Individual applications can lead to improved precision and a lower radiation exposure in certain areas. This is particularly the case for 3D navigated surgery of the spine and pelvis, as well as for hip and knee replacement surgery, but also robot-assisted applications in the area of joint replacement surgery.

The evidence is rather modest for most areas of application. Furthermore, clinical-functional improvements were not seen for any area, with the result that some 2-D navigated surgical procedures (for example, corrective osteotomies, anterior cruciate ligament surgery, trauma navigation) are only rarely used these days owing to the tendentially greater effort involved in planning and performing the procedure. The trend towards research innovations in the area of joint replacement surgery and the spine has in recent years moved from navigated approaches to robot-assisted approaches. Future studies should focus on clinical-functional results, with medium term and long term investigations to a greater extent.

Figure 1.

Navigated stabilization surgery in a spinal injury

The patient was a 73 year old man who had fallen down a staircase under the influence of alcohol. The patient was awake and alert at hospital admission; preoperatively he had no neurological deficits. Computed tomography scanning (a) revealed polytrauma to the cervical spine with a tear drop fracture of the 2nd vertebral body, segmental instability at the level of the 6th and 7th cervical vertebral bodies with bilateral dislocation of the facet joints, arch fractures of the 7th cervical vertebral body, and a complete burst fracture of the first thoracic vertebral body. After preoperative preparation in the operating theater, the navigation camera and work station were positioned at the side of the surgeon, so as to enable intraoperative handling. After open preparation, the dislocation was repositioned, followed by decompression, and subsequently a reference clamp was attached to the vertebral body that was to be treated. After referencing a preoperative computed tomography scan by using a surface matching approach, lateral mass screws and pedicle screws were placed in the area of the cervical spine, assisted by navigation. Additionally, pedicle screws were attached in the area of the thoracic spine, again by using navigation, and the screws were connected by rods (b). The postoperative x-ray and CT images showed rule conform screw placement; no further neurological impairments were found (c).

Key Messages.

• Computer-assisted methods can improve radiological precision and minimize gross axial deviations.

• No relevant improvement has been shown for clinical-functional outcomes so far.

• Navigation devices and robotic surgery are associated with a great initial financial outlay; thus far, no additional remuneration is paid.

• Computer-assisted methods can pave the way for research innovations and the development of new surgical techniques, and they can be used for simulations or medical education and training.

• Although navigation assisted methods are already established for certain uses, most recent developments and innovations have been described for robotic surgery

Test questions for the article in issue 47/2020:

Computer assisted orthopedic and trauma surgery

CME credits for this unit can be obtained via cme.aerzteblatt.de until 19 November 2021 Only one answer is possible per question. Please select the answer that is most appropriate.

Question 1

What does the acronym “CAOS” stand for?

1. computational autonomous orthopaedic surgery

2. cyber-assisted orthopaedic surgery

3. complex automated orthopaedic surgery

4. computer-assisted orthopaedic surgery

5. computer-assisted organisation of surgery

Question 2

Which drawbacks have registry studies shown for sledge prostheses implanted by a robot?

1. A higher rate of necessary revision surgery.

2. Poorer positioning

3. poorer reconstruction of the joint line

4. A less natural gait pattern

5. Notably greater costs compared with manual techniques

Question 3

For which types of surgical procedures are navigation-assisted approaches the default method?

1. Stabilizing joint fractures

2. Predicle screw osteosynthesis

3. Tratment of anterior cruciate ligament rupture

4. Stabilizing femur and tibia

5. Treatment of posterior cruciate ligament rupture

Question 4

What posed a particular problem when using early autonomous surgical robots compared with manual techniques?

1. Intraoperative fractures

2. Postoperative joint misalignment

3. Planning errors were increased by the system

4. Higher mortality rates

5. Postoperative fractures

Question 5

What was shown by randomized controlled trials of total knee replacement surgery when comparing navigated and conventional implantation?

1. A longer lifetime of the prosthesis after navigated implantation.

2. A better Knee Society Score after navigated implantation.

3. A poorer result in terms of movement range after navigated implantation.

4. Poorer radiographical results after conventional implantation.

5. More time spent doing surgery in navigated implantation.

Question 6

A randomized controlled trial of fractures of the posterior pelvic ring determined the rate of misplaced screws. What was this rate for 3-D navigated surgery?

1. 5.2 %

2. 20.4 %

3. 0 %

4. 16.6 %

5. 9.8 %

Question 7

Which approach is regarded as the standard procedure in stabilizing spinal, as supported by good evidence?

1. 2-D navigation

2. Surgery performed by an autonomous robot

3. Conventional surgery

4. Surgery performed by an articulated arm robot

5. 3-D navigation

Question 8

In which area did a randomized controlled trial show a dropout rate of 40% for navigated surgery?

1. Sacroiliac screw osteosynthesis

2. Total shoulder replacement surgery

3. Pedicle screw osteosynthesis in the sacral spine area

4. Total hip replacement surgery

5. Total knee replacement surgery

Question 9

The results of initial cohort studies of robot-assisted total knee replacement surgery are available. These results imply which differences to manual surgery?

1. Poorer knee function in the short term

2. Greater blood loss

3. Quicker early rehabilitation

4. Poorer soft tissue balancing

5. Unsatisfactory positioning of the prosthesis

Question 10

Which beneficial advantage is navigated surgery and robot-assisted surgery most likely to confer on the surgeon?

1. The additional surgical effort is remunerated appropriately.

2. The radiation dose for the eye area, thorax, and forearms can be reduced.

3. Planning is less laborious than in conventonal surgery.

4. Patients are selected by the systems.

5. The surgeon needs less experience as the system takes over many complex tasks.