Abstract
Purpose
-
To determine whether using robots in spine surgery results in more clinical advantages and fewer adverse consequences.
Methods
-
Between October 1990 and October 2022, a computer-based search was conducted through the databases of PubMed, Cochrane Library, Embase, Web of Science, China National Knowledge Infrastructure, China Biology Medicine, VIP databases, and WAN FANG. The study only included randomized controlled trials (RCTs) comparing the clinical efficacy and safety of robot-assisted surgery with those of conventional spine surgery. The review was conducted following PRISMA 2020, and AMSTAR-2 was used to evaluate the methodological quality. R version 4.2.1 was used in the meta-analysis. The Cochrane Collaboration Tool was used for evaluating the risk of bias.
Results
-
This study analyzed 954 participants from 20 RCTs involving cervical spondylosis, lumbar degenerative disease, scoliosis, etc. The robot-assisted group outperformed the freehand group in terms of intraoperative blood loss, number of screws in grade A position, grade A + B position, radiation dose, and hospital stay. Operation duration, visual analog scale scores of low back pain, Oswestry disability index, and radiation exposure time did not significantly differ between the two groups.
Conclusions
-
Although robotic spine surgery is more accurate in pedicle screw placement than conventional methods, the robot group did not demonstrate an advantage in terms of clinical efficacy. Studies of complications and cost-effectiveness are still very rare.
Introduction
Miscellaneous spinal disorders, including degenerative diseases, deformities, and trauma, are among the major causes of disability worldwide (1, 2). The persistently high morbidity rates make it become an important contributor to the global healthcare burden (3). Surgical intervention is the main strategy when patients with severe neurological impairments and unresponsive to conservative therapeutic measures. Since Hibbs and Albee first reported spinal fusion surgery in 1911, this operative procedure has gradually become the most widely used for treating spinal pathologies (4). Pedicle screw placement played an important role in the reconstruction of spinal stability, promoting fusion and early rehabilitation, and has become an essential procedure in spine surgery (5). Rajaee et al. (6) revealed that with the widespread application of pedicle screws, the number of spinal fusions has also increased dramatically by 2.4 times annually from 1998 to 2008.
Conventional freehand screw placement relies on anatomical landmarks, intraoperative fluoroscopy, and clinical experience of the surgeons. For inexperienced doctors, the risk of screw placement failure is higher, which can lead to neurovascular injuries and other complications (7, 8). The pedicle screw misplacement rates of conventional techniques are 30% and 55% in the lumbar and thoracic spines, respectively (9, 10, 11). This contrasts the reported high success rate of robot-assisted pedicle screw placement from 91.5% to 94.4% (12). In nearly two decades, robotic-assisted spine surgery (RSS) has represented a revolution in new surgical technologies. Despite being an emerging technology, it has been widely adopted for various surgical procedures (13, 14). The precise positioning performance of robots has greatly promoted the development of minimally invasive surgery (15, 16). Recently, some new spinal robotic systems are introduced into clinical use (17, 18).
Robot-assisted surgical technology is generally thought to be more accurate and stable, and several recent data and meta-analysis confirmed the benefits of robotic surgery in accurate pedicle screw placement (19, 20). However, previous meta-analyses have focused on the accuracy of screw placement, with insufficient attention to clinical outcomes and complications, and have not analyzed the differences in outcomes between different robot types and surgical segments. Therefore, we comprehensively included high-quality randomized controlled trials (RCTs) for a comprehensive in-depth systematic review and meta-analysis. This study aimed to further clarify whether RSS could enhance efficacy and decrease complications by improving the accuracy of the pedicle screw. In addition, this study focused on whether different types of robots, operating sites, and regions/countries affect clinical outcomes. It contributes to the further understanding of the advantages and deficiencies of RSS and provides suggestions on the future directions of this field.
Methods
The work has been described following AMSTAR-2 (21) and PRISMA 2020 (22). The study is listed in the PROSPERO database (registration number: CRD42022375991).
Search strategy
Robot-assisted pedicle screw placement was the subject of an electronic literature search from October 1990 to October 2022 in eight databases, including PubMed, The Cochrane Library, Embase, Web of Science Core Collection, China National Knowledge Infrastructure, China Biology Medicine, Wanfang Digital Periodicals (WAN FANG), and VIP databases. Orthopedic robotic surgery, pedicle screws, randomized controlled trials, and other terms were included in the search, which were only restricted to English and Chinese. A similar approach was used for additional electronic databases. References of eligible studies were also searched. The complete search strategy for PubMed is provided in Supplementary File 1 (see section on supplementary materials given at the end of this article). The whole texts of any eligible studies, as well as their titles and abstracts, were independently evaluated by two researchers (W-X S and W-Q H). Disagreements were discussed with a third researcher (Y-P L).
Inclusion and exclusion criteria
The PICOS principle was used to set the following inclusion criteria: (i) participants: the participants who required pedicle screw insertion including degenerative spinal disease or spinal fractures, and whose age was restricted to 18–60; (ii) Interventions: the robot-assisted technique for treating diseases was used in the experimental group (RA group); (iii) comparisons: the conventional freehand technique was used in the control group (FH group); (iv) results: at least two of the following were used as outcome indicators: the number of screws in grade A and grade A + B positions, visual analog scale (VAS) of low back pain and Oswestry disability index (ODI) scores, radiation dose and exposure time, intraoperative blood loss, operating time, and hospital stay. Based on the Gertzbein–Robbins classification (9), postoperative follow-up CT was used to assess the accuracy of pedicle screw placement; and (v) RCTs were acceptable as study designs.
The exclusion criteria trials were as follows: participants with a history of prior spinal surgery or combined spinal tumors, infections, rheumatic immune diseases, etc. Other exclusion factors include studies by the same authors in different languages, studies conducted on the same subjects during the same period, and studies reported by the same author.
Data extraction
Two researchers (W-XS and H-YL) independently gathered baseline data and outcome markers. Disagreements were settled by conversing with a third researcher (Y-P L) or seeking outside counsel. Baseline information includes the first author, publication year, study country, research design, sex, robot type, screw section, etc. Clinical indicators (VAS scores of low back pain, ODI scores, grade A and A + B screw positions) were the main outcome markers of interest. Safety factors, such as radiation dose, radiation exposure time, intraoperative blood loss, operation time, and hospital stay, served as secondary outcome indicators.
Statistical analysis
Risk ratios (RRs) and 95% CIs were calculated for dichotomous variables. The mean difference (MD) was calculated for continuous variables, and the standard mean difference (SMD) and its 95% CI were used as statistics when the calculation method and units varied. A χ 2 test with α = 0.05 was used to examine heterogeneity between trials. If I 2 ≤ 50 and P > 0.05, heterogeneity between trials was deemed unimportant, and a fixed effects model was used for the meta-analysis. A meta-analysis employing a random-effects model was performed if I 2 > 50 and P ≤ 0.05, which indicated considerable heterogeneity, with P > 0.05 suggesting statistically significant differences in the outcome markers.
Egger’s regression was used to examine funnel plots to determine heterogeneity and bias, adding Begg’s test where there >10 studies were included. Sensitivity analysis (leave-one-out method) was used to evaluate the influence of each study to confirm the consistency of the meta-analysis’s conclusions. We will do subgroup analysis for the region, robot type, and screw section for the key outcome indicators. At P < 0.05, the differences between the analysis subgroups are deemed significant. R 4.2.1 was used to conduct meta-analyses (www.r-project.org).
The interquartile range and the median of some of the data in the available literature indicate that the data were not normally distributed. Box–Cox and Quantile estimates, which more precisely estimated means and s.d., were suggested by McGrath et al. (23) to modify the data. Using an online scientific calculator (https://smcgrath.shinyapps.io/estmeansd/), these data were converted into means and s.d. The s.e., which was calculated for the meta-analysis, could be applied to both the RA and freehand groups, using the P-value and sample size to calculate the s.d. to complete the data (24). This was done because some data lacked s.d., so the s.e. was assumed to be equal for each group.
Results
Study search
In total, 197 studies were retrieved, 68 of which were repeatedly published. Ultimately, 20 RCTs were selected, 9 of which were written in English (20, 25, 26, 27, 28, 29, 30, 31, 32) and 11 were in Chinese (33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43). Quality assessments of the studies are detailed in Figs. 1 and 2. Figure 3 shows the flow diagram for the study selection process. Three studies (44, 45, 46) were eliminated because they contained trials from the same period or the literature was written by the same authors, and one article (47) with an unclear indication that was a prospective research was also eliminated. In total, 1582 patients were registered, with 787 in the RA group and 795 in the FH group. The publication years were mostly in the range of 2012–2022. More information on the included articles and specific complications are provided in Tables 1, 2, and 3.
Demographic characteristics of the studies included in the meta-analysis.
| Study | Country | Study type | Sex | Cases | Average age (years) | BMI | ||||
|---|---|---|---|---|---|---|---|---|---|---|
| M | F | RA | FH | RA | FH | RA | FH | |||
| Cui et al. (25) | China | RCT | 10 | 38 | 23 | 25 | 51.3 ± 9.8 | 54.1 ± 10.2 | N/A | N/A |
| Fan et al. (26) | China | RCT | 82 | 45 | 61 | 66 | 49 | 49.5 | 23.65 ± 4.1 | 24.47 ± 3.94 |
| Feng et al. (20) | China | RCT | 31 | 49 | 40 | 40 | 63.45 ± 4.56 | 64.22 ± 6.19 | N/A | N/A |
| Han et al. (27) | China | RCT | 113 | 121 | 115 | 119 | 54.6 ± 11.3 | 56.1 ± 11.4 | 25.7 ± 4.1 | 24.9 ± 2.9 |
| Hyun et al. (28) | Korea | RCT | 17 | 43 | 30 | 30 | 66.5 ± 8.1 | 66.8 ± 8.9 | 24.7 ± 2.6 | 25.8 ± 3.3 |
| Kim et al. (29) | Korea | RCT | 41 | 37 | 37 | 41 | 65.4 ± 10.4 | 66.0 ± 8.6 | 25.9 | 25.3 |
| Li et al. (30) | China | RCT | 7 | 10 | 7 | 10 | 47.4 ± 12.9 | 49.9 ± 10.9 | 24.3 ± 1.8 | 24.6 ± 2.6 |
| Ringel et al. (31) | Germany | RCT | 26 | 34 | 30 | 30 | 68 | 67 | 26 | 28 |
| Roser et al. (32) | Germany | RCT | N/A | N/A | 18 | 10 | N/A | >18 | N/A | N/A |
| Wang et al. (33) | China | RCT | 63 | 19 | 40 | 42 | 42.7 ± 8.0 | 43.1 ± 9.1 | N/A | N/A |
| Hou et al. (34) | China | RCT | 49 | 13 | 32 | 30 | 43.1 ± 8.91 | 44.2 ± 9.10 | 26.15 ± 3.97 | 25.97 ± 4.02 |
| Wang & Kedong (35) | China | RCT | 27 | 33 | 28 | 32 | 55.7 ± 6.2 | 56.0 ± 7.4 | 26.3 ± 3.1 | 25.9 ± 2.9 |
| Huang et al. (57) | China | RCT | 33 | 27 | 28 | 32 | 54.32 ± 6.54 | 54.81 ± 7.56 | 26.22 ± 5.07 | 25.81 ± 3.42 |
| Xu et al. (37) | China | RCT | 25 | 18 | 24 | 19 | 45.2 ± 9.2 | 50.6 ± 10.7 | N/A | N/A |
| Wei et al. (38) | China | RCT | 17 | 23 | 23 | 17 | 55.5 ± 10.1 | 52.0 ± 14.1 | N/A | N/A |
| Zhai et al. (39) | China | RCT | 11 | 20 | 16 | 15 | 16.7 ± 7.0 | 15.4 ± 4.6 | N/A | N/A |
| Zhang et al. (43) | China | RCT | 47 | 43 | 44 | 46 | 54.82 ± 3.51 | 55.06 ± 3.57 | N/A | N/A |
| Yang et al. (40) | China | RCT | 49 | 39 | 45 | 43 | 45.27 ± 7.71 | 47.83 ± 6.82 | N/A | N/A |
| Cao et al. (41) | China | RCT | 19 | 11 | 15 | 15 | 16.08 ± 2.31 | 16.16 ± 2.26 | N/A | N/A |
| Li et al. (42) | China | RCT | N/A | N/A | 12 | 10 | N/A | N/A | N/A | |
FH, freehand; RA, robot-assisted; RCT, randomized controlled trial.
Details of robot, number of screws used, type of disease, and outcome parameters reported in the studies included in the meta-analysis.
| Type of robot | Screw segment | Screws, n | Type of disease | Outcome parameters reported on | |||||||
|---|---|---|---|---|---|---|---|---|---|---|---|
| RA | FH | SPG | RAD | Pain scores* | Blood loss | Op time | HS | ||||
| Cui et al. (25) | Tirobot | L4-S1 | 92 | 100 | DDS | A | Y | Y | Y | Y | |
| Fan et al. (26) | Tirobot | Cervical vertebrae | 186 | 204 | DDS | A + B | Y | Y | Y | ||
| Feng et al. (20) | Tirobot | Lumbar vertebrae | 170 | 174 | DS | A + B | Y | Y | Y | ||
| Han et al. (27) | Tirobot | Thoracolumbar spine | 532 | 584 | DDS; Spinal fracture | A + B | Y | Y | Y | ||
| Hyun et al. (28) | Renaissance | Lumbar vertebrae | 130 | 140 | DDS | A + B | Y | Y | Y | ||
| Kim et al. (29) | Renaissance | L2-S1 | 158 | 172 | DDS | A + B | Y | Y | |||
| Li et al. (30) | Orthbot | Lumbar vertebrae | 32 | 50 | DDS | A + B | Y† | Y | Y | Y | |
| Ringel et al. (31) | SpineAssist | L2-S1 | 146 | 152 | N/A | A + B | Y | Y | Y | ||
| Roser et al. (32) | SpineAssist | T1-S1 | 72 | 40 | N/A | A + B | Y‡ | Y§ | Y | ||
| Wang et al. (33) | Renaissance | T10-L2 | 240 | 252 | Spinal fracture | A + B | Y | ||||
| Hou et al. (34) | Renaissance | Thoracic vertebra | 230 | 216 | Spinal fracture | A + B | Y | ||||
| Wang & Kedong (35) | Tirobot | Lumbar vertebrae | 112 | 128 | DDS | A + B | Y | Y | Y | ||
| Huang et al. (57) | Tirobot | Lumbar vertebrae | 112 | 128 | DDS | A + B | Y§ | Y | Y | Y | |
| Xu et al. (37) | Tirobot | Thoracolumbar spine | 132 | 106 | DDS | A + B | Y | Y | Y | ||
| Wei et al. {38) | Tirobot | Lumbar vertebrae | 102 | 88 | DDS; Spinal fracture | A + B | Y | Y | |||
| Zhai et al. (39) | Renaissance | Thoracolumbar spine | 276 | 255 | Scoliosis | A + B | Y | Y | Y | ||
| Zhang et al. (43) | Tirobot | Thoracolumbar spine | N/A | N/A | Spinal fracture | Y‡ | Y | ||||
| Yang et al. (40) | Tirobot | Thoracolumbar spine | 208 | 202 | DDS; Spinal fracture | A + B | Y‡ | Y | Y | ||
| Cao et al. (41) | Tirobot | Thoracolumbar spine | 212 | 217 | Scoliosis | A + B | Y | Y | Y | ||
| Li et al. (42) | XGK-6508A | Lumbar vertebrae | 58 | 44 | DDS | A + B | Y | Y | |||
Blood loss, refers to intraoperative blood loss (mL); DDS, degenerative disease of spine including cervical spondylopathy, spondylolisthesis and spinal stenosis; FH, freehand; HS, hospital stay in days; Op time, operation time (minutes); RA, robot assisted; RAD, radiation dose and exposure time (s); SPG, screw position grade; Y, yes .
*VAS and ODI; †exposure time only; ‡dose only; §VAS only.
Complications reported in the studies.
| Complications | RA (n = 787) | FH (n = 795) |
|---|---|---|
| Screw-related | ||
| Proximal facet violations | 5 (0.64%) | 33 (4.15%) |
| Screw repair during operation | 16 (2.03%) | 49 (6.16%) |
| Postoperative screw loosening | 0 | 1 (0.13%) |
| Injury of vertebral artery | 0 | 1 (0.13%) |
| Nerve root injury | 2 (0.25%) | 10 (1.26%) |
| Screw-unrelated | ||
| Infection | 2 (0.25%) | 5 (0.63%) |
| Numbness in the back and lower limbs | 1 (0.13%) | 0 |
| Delayed wound healing | 0 | 1 (0.13%) |
| Transfusion events | 10 (1.27%) | 14 (1.76%) |
| Cerebrospinal fluid leak | 1 (0.13%) | 3 (0.38%) |
| Decreased muscle strength | 1 (0.13%) | 1 (0.13%) |
| Dural injury | 0 | 1 (0.13%) |
| Total | 38 (4.83%) | 119 (14.97%) |
FH, freehand; RA, robot assisted.
Clinical efficacy evaluation
VAS scores of low back pain
Nine studies (20, 25, 28, 29, 35, 36, 37, 39, 41) discussed the VAS scores of low back pain, which included 490 patients, i.e., 241 in the RA group and 249 in the FH group. The two groups had high levels of heterogeneity (P < 0.05, I 2 = 80%), and a random-effects model was used. The results of the study (Fig. 4) indicated a statistical difference between the RA group and FH group (MD = −0.44, 95% CI (−0.89, 0.02), P < 0.01). No differences were found between the subgroups based on country, robot type, and screw segment (P > 0.05) (Supplementary Fig. 1).
VAS scores of low back pain forest plot.
Citation: EFORT Open Reviews 8, 11; 10.1530/EOR-23-0125
ODI scores
In this study, 430 patients (RA group, n = 213; FH group, 217) (20, 25, 28, 29, 35, 37, 39, 41). The two groups did not differ from one another (P> 0.05, I2 = 38%), and a fixed effects model was used. According to the analysis (Fig. 5), no difference was found between the two groups (MD = −0.79, 95% CI (−1.62, 0.04), P > 0.05). The country, robot type, and screw segment did not show differences between subgroups (Supplementary Fig. 2). These results suggest that ODI scores did not correlate with country, robot type, or screw segment.
Pedicle screw placement accuracy
Grade A screw position
In total, 19 studies (20, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38, 39, 40, 41, 42) had grade A screw position. The total includes 6452 screws, of which 3200 were created with robot assistance and 3252 with manual methods. By using a random-effects model for meta-analysis, differences between the two groups that were statistically significant (P < 0.05, I 2= 81%) were examined. In this study, 1.13 times as many screws in grade A position in the RA group as there were in the FH group (Fig. 6) (RR = 1.13, 95% CI (1.07, 1.19), P < 0.01). Depending on the country and screw segment, differences were found between the subgroups (P < 0.01) (Supplementary Figs. 3 and 4). Based on the robot type, no differences were found between the groupings (P > 0.05) (Supplementary Fig. 5). These results indicate that the number of grade A screw positions was related to the country and screw segment and not to the type of robot.
Screw position grade A forest plot.
Citation: EFORT Open Reviews 8, 11; 10.1530/EOR-23-0125
Grade A + B screw position
A total of 6260 screws, of which 3108 were created with robot assistance and 3152 with manual methods, were described in 18 studies (20, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38, 39, 40, 41, 42). According to the analysis (Fig. 7), the RA group had 1.04 times more screws in grade A + B position than the FH group (RR = 1.04, 95% CI (1.02, 1.06), P < 0.01). Differences were found between subgroups based on country, robot type, and screw segment (P < 0.01) (Supplementary Figs. 6, 7, and 8). These results suggest that the number of grade A + B screw position correlated with the country, robot type, and screw segment.
Screw position grade A + B forest plot.
Citation: EFORT Open Reviews 8, 11; 10.1530/EOR-23-0125
Safety indicators
Radiation dose
Five studies (27, 32, 35, 40, 43) reported the radiation dose and analyzed a total of 500 patients, with 250 in each group. As shown in Fig. 8, the radiation dose was lower in the RA group than in the FH group (SMD = −1.78, 95% CI (−3.00, −0.95), P< 0.01).
Radiation exposure time
Five studies (27, 30, 31, 32, 36) reported the radiation exposure time and included a total of 399 patients, with 196 in the RA group and 203 in the FH group. The analysis showed (Fig. 9) that the duration of radiation exposure time in the RA group did not differ from that in the FH group (SMD = −0.08, 95% CI (−0.93, 0.77), P < 0.01).
Radiation exposure time forest plot.
Citation: EFORT Open Reviews 8, 11; 10.1530/EOR-23-0125
Intraoperative blood loss
Twelve articles (20, 25, 26, 27, 30, 35, 36, 37, 39, 40, 41, 42) reported intraoperative blood loss and included a total of 840 patients, with 414 in the RA group and 426 in the FH group. Heterogeneity between the two groups was high (P < 0.05, I 2 = 98%), and meta-analysis was performed using a random-effects model. The analysis showed (Fig. 10) that the incidence of intraoperative blood loss was less in the RA group than in the FA group (MD = −60.55, 95% CI (−112.10, −9.01), P< 0.01).
Intraoperative blood loss forest plot.
Citation: EFORT Open Reviews 8, 11; 10.1530/EOR-23-0125
Operation time
Seventeen articles (20, 25, 26, 27, 28, 29, 30, 31, 34, 35, 36, 37, 38, 39, 41, 42, 43) reported on the operation time and included a total of 1142 patients, with 565 in the RA group and 577 in the FH group. Heterogeneity between the two groups was high (P < 0.05, I 2 = 97%), and a meta-analysis was performed using a random-effects model. The analysis showed (Fig. 11) a statistical difference between the RA and FH groups; however, it could not be concluded that the operating time was shorter in the RA group than in the FH group (MD = −2.15, 95% CI (−19.50, 15.20), P < 0.01).
Hospital stay
Nine studies (25, 26, 27, 28, 30, 31, 36, 39, 42) reported the hospital stay and included a total of 1142 patients, 565 in the RA group and 577 in the FH group. Heterogeneity between the two groups was high (P < 0.05, I 2 = 86%) and meta-analysis was performed using a random-effects model. The analysis showed (Fig. 12) that the hospital stay was shorter in the RA group than in the FH group (SMD = −0.75, 95% CI (−1.40, −0.11), P < 0.01).
Sensitivity analyses and publication bias
VAS scores of low back pain, ODI scores, radiation dose, radiation exposure time, intraoperative blood loss, and hospital stay funnel plots suggested overall basic symmetry and further Egger’s regression (P > 0.05), further indicating that the findings were robust (Supplementary Fig. 9). Funnel plots of screw position grade A suggested overall asymmetry, further Egger’s regression (P < 0.05), and Begg’s test (P > 0.05), indicating that the findings were not robust. Grade A + B funnel plot suggested overall asymmetry, followed by Egger’s regression (P < 0.05), and Begg’s test (P< 0.05) indicated that the study results were not robust. Operation time funnel plot suggested overall asymmetry, further Egger’s regression (P > 0.05), and Begg’s test (P < 0.05), indicating that the study results may have publication bias (Supplementary Fig. 10).
Heterogeneity of VAS scores of low back pain, grade A screw position, grade A + B screw position, radiation dose, radiation exposure time, operation time, and hospital stay were high. After the removal of Hyun, the results of the sensitivity analysis showed that the heterogeneity of ODI scores decreased to 0 (28), suggesting that this study may have had a greater effect on the ODI scores, and the findings did not change after removal. Further sensitivity analysis of intraoperative blood loss showed a decrease in heterogeneity after Cui removal (25), suggesting that this study may have had a greater effect on the intraoperative blood loss results; however, the conclusions were not altered after removal.
Discussion
In this meta-analysis, the study participants were from different areas across three countries. The main diagnoses were lumbar spondylolisthesis, thoracolumbar fracture, scoliosis, cervical spondylosis, and spinal stenosis, which cover almost all common spinal diseases aside from tumors. The surgical segments covered the whole spine. Compared with the published meta-analysis, this study included the largest sample size of RCTs. Articles were carefully re-selected after the initial screening to minimize selection bias, reasons that studies had been culled including similar studies from the same hospital and literatures with overlapping study periods. Moreover, the recent findings of relevant studies on the two latest robots were included to further enhance the reliability of the meta-analysis.
Our results contribute further evidence supporting that the accuracy of robot-assisted pedicle screw placement (grades A and A + B) was superior to the freehand group, which reflects the main advantages of RSS that the robot can work consistently and accurately and eliminate medical errors (e.g. screw misplacement) due to surgeons’ inexperience and mental fatigue. A large, multicenter retrospective analysis showed that RSS was more accurate than freehand screw placement (48). This conclusion has also been confirmed by further prospective RCTs (19, 25, 26). Some studies point out the advantages of robotic-assisted screw placement accuracy in difficult surgeries such as scoliosis (39, 41). Although RSS has higher accuracy of pedicle screw placement, we did not find this apple on translating into clinical benefits.
Clinical efficacy is one of the most important dimensions in evaluating an innovative technique and should be considered a primary outcome indicator. Our findings suggest no significant differences in VAS and ODI scores between the RSS group and the conventional surgery group. This illustrates that greater precision in screw placement is not equivalent to better patient outcomes. In clinical practice, some patients with poor screw position grades (grade C or even grade D) do not have corresponding clinical symptoms during follow-up, which is consistent with the findings of previous studies (29, 49). Remarkably, the accuracy of the freehand screw placement was greatly dependent upon the surgeon’s experience and awareness of intraoperative fluoroscopy outcome. Experienced surgeons can insert screws more efficiently and safely (50). In addition, adequate nervous decompression is the key to ensure therapeutic potency. However, the robot cannot provide substantial assistance. This would explain why more accurate screw placement did not lead to better results.
The RSS had a lower radiation dose; however, no difference was found in the radiation exposure time between the groups. Adequate preoperative planning in the RSS group can reduce the number of intraoperative radiographs, and the surgeon can insert screws step-by-step under robotic guidance without the need for additional fluoroscopy (51). A prospective RCT found that the radiation dose in the freehand group was almost four times higher than that of the RSS group (28). Some types of robots require a three-dimensional CT in the prone position during surgery, which prolongs intraoperative radiation exposure (52). However, overall, the radiation dose for robot-assisted screw placement is reduced.
The RSS group had less intraoperative blood loss and shorter hospital stays, although the operation time did not differ between the groups. Most of the RSS is minimally invasive surgery; thus, it can reduce the incision and avoid prolonged muscle distraction, which contributes to reduced postoperative pain and shorter hospital stays (25). The large variation in the operation time between different spinal procedures and the long preparation and commissioning time required for the robot upfront prolong the duration of robotic surgery. In the future, the operation time can still be further reduced by optimizing the preparation process and skilled use of the robot.
Perioperative adverse events of RSS are another important concern that we should pay more attention; however, not all RCTs had reported complications. Some of them were unable to confirm whether all complications were disclosed. As a result, a complication summary was only conducted using the limited data available, not a meta-analysis. Collectively, the common complications of RSS including proximal facet violations, pedicle screw misplacement or loosening, vertebral artery and nerve root injury, infection, delayed wound healing, transfusion, dural injury, and cerebrospinal fluid leak (20, 25, 26, 27, 28, 29, 30, 31, 32), and the overall complication rate of RSS was 4.83%, with a screw-related complication rate of 2.92%, which was significantly lower than that in the traditional surgery group.
In addition, RSS has some flaws and questions. (i) At this stage, most of the clinical applications of RSS are focused on pedicle screw placement. Only a small number of cases of robotic-assisted vertebroplasty and spinal endoscopic surgery have been reported. Moreover, few or no studies have focused on other critical steps of spine surgery, such as robot-assisted spinal decompression, discectomy, interbody fusion, and osteotomy. (ii) Although RSS performs excellently in screw insertion, some problems with the practical application of orthopedic robots must be considered, which consisted mostly of unstable operation, time-consuming image alignment, and lateral slippage of the positioning screw (53). Robotic malfunction or operational error may cause serious complications and increase the operative duration. Page et al. found that the operative time was significantly longer in the RSS group than in the conventional surgery group (54). Such conditions could be associated with different design and manufacturing philosophy of different robot types, which require further exploration (55). (iii) The high cost of purchase and maintenance of robots is the major obstacle that limits its generalized application. Peter et al. found that the cost of a robot in lumbar fusion surgery was significantly higher than that of conventional open and minimally invasive surgery (28). Another study showed that robot-assisted minimally invasive transforaminal lumbar interbody fusion was cost-effective in 63% of simulations when patients were willing to pay $50 000/quality-adjusted life year (QALY) (56). (iv) In addition, the learning curves in RSS should not go unnoticed. RSS takes approximately 25 practice operations for a surgeon to become proficient (53). The learning curve for RSS has many variables, and similar nuances and inconsistencies in different types of robots are a relatively important factor among them (57).
This study is limited by the use of freehand screw placement, which comprised two different treatment paradigms, i.e., conventional percutaneous fluoroscopy-assisted screw placement and open surgical freehand screw placement, which has some consequences on the clinical efficacy evaluation. Their case numbers were not large enough to permit a subgroup analysis, and this may also contribute to some of the shortcomings in this study.
Conclusion
The accuracy of screw placement is the major advantage of RSS. The potential to decrease blood loss and hospital stays was noted. Moreover, the complication rates of RSS appear to be lower than conventional techniques. However, these strengths did not translate to better clinical efficacy. Therefore, how to broaden the application of RSS reasonably needs further research for greater clinical effectiveness, which included but is not limited to robot-assisted spinal decompression and tumors resection. More attention is needed to the efficacy, complications, and cost–effectiveness ratio. Further integration is still needed to enable automatic registration and automatic planning of surgical paths and improve the automation and intelligence of orthopedic surgical robots. Large, multicenter, high-quality, RCTs with uniform metrics are still needed to provide more robust evidence.
Supplementary materials
This is linked to the online version of the paper at https://doi.org/10.1530/EOR-23-0125.
ICMJE Conflict of interest Statement
The authors declare that there is no conflict of interest that could be perceived as prejudicing the impartiality of the research reported.
Funding Statement
This work was partially supported by a grant from National Natural Science Foundation of China (No. 82004385 to YPL), Excellent Talents Program of Guangdong Provincial Hospital of Chinese Medicine (No. ZY2022YL17 for LYP), The State Key Laboratories Development Program of Guangzhou (No.202102010012 to CBL), The Innovation Team Project of Guangdong Provincial Department of Education (No. 2021KCXTD020 to CBL), and Administration of Traditional Chinese Medicine of Guangdong Province (o. 20190407090652 to DJ).
Author contribution statement
W-x Sun: investigation, data curation, formal analysis, writing the original draft; W-q Huang: investigation, data curation, formal analysis; H-y Li: investigation, data curation; H-s Wang: methodology, reviewing and editing; S-l Guo: analysis and interpretation of data and supervision; J Dong: methodology; B-l Chen: conceptualization, methodology, writing, reviewing and editing, and supervision; Y-p Lin: conceptualization, methodology, writing, reviewing, and editing, supervision.
Acknowledgements
We appreciate the team at the Guangzhou Key Laboratory of Integrated Traditional Chinese and Western Medicine for Prevention and Treatment of Lumbar Degenerative Diseases.
References
- 1.↑
Ross PD, Davis JW, Epstein RS, & Wasnich RD. Pain and disability associated with new vertebral fractures and other spinal conditions. Journal of Clinical Epidemiology 1994 47 231–239. (https://doi.org/10.1016/0895-4356(9490004-3)
- 2.↑
Prevalence of disabilities and associated health conditions among adults–United States, 1999. JAMA 2001 245 1571–0000. (https://doi.org/10.1001/jama.285.12.1571-jwr0328-3-1)
- 3.↑
GBD 2017 Disease and Injury Incidence and Prevalence Collaborators, Afshin A, & Agesa KM. Global, regional, and national incidence, prevalence, and years lived with disability for 354 diseases and injuries for 195 countries and territories, 1990–2017: a systematic analysis for the Global Burden of Disease Study 2017. Lancet 2018 392 1789–1858. (https://doi.org/10.1016/S0140-6736(1832279-7)
- 4.↑
Reid PC, Morr S, & Kaiser MG. State of the union: a review of lumbar fusion indications and techniques for degenerative spine disease. Journal of Neurosurgery. Spine 2019 31 1–14. (https://doi.org/10.3171/2019.4.SPINE18915)
- 5.↑
Magerl FP. Stabilization of the lower thoracic and lumbar spine with external skeletal fixation. Clinical Orthopaedics and Related Research 1984 189 125–141. (https://doi.org/10.1097/00003086-198410000-00014)
- 6.↑
Rajaee SS, Bae HW, Kanim LEA, & Delamarter RB. Spinal fusion in the United States: analysis of trends from 1998 to 2008. Spine 2012 37 67–76. (https://doi.org/10.1097/BRS.0b013e31820cccfb)
- 7.↑
Sun J, Wu D, Wang Q, Wei Y, & Yuan F. Pedicle screw insertion: is O-arm-based navigation superior to the conventional freehand technique? A systematic review and meta-analysis. World Neurosurgery 2020 144 e87–e99. (https://doi.org/10.1016/j.wneu.2020.07.205)
- 8.↑
Joseph JR, Smith BW, Liu X, & Park P. Current applications of robotics in spine surgery: a systematic review of the literature. Neurosurgical Focus 2017 42 E2. (https://doi.org/10.3171/2017.2.FOCUS16544)
- 9.↑
Gertzbein SD, & Robbins SE. Accuracy of pedicular screw placement in vivo. Spine 1990 15 11–14. (https://doi.org/10.1097/00007632-199001000-00004)
- 10.↑
Weinstein JN, Spratt KF, Spengler D, Brick C, & Reid S. Spinal pedicle fixation: reliability and validity of roentgenogram-based assessment and surgical factors on successful screw placement. Spine 1988 13 1012–1018. (https://doi.org/10.1097/00007632-198809000-00008)
- 11.↑
Xu R, Ebraheim NA, Ou Y, & Yeasting RA. Anatomic considerations of pedicle screw placement in the thoracic spine. Roy-Camille technique versus open-lamina technique. Spine 1998 23 1065–1068. (https://doi.org/10.1097/00007632-199805010-00021)
- 12.↑
Gao S, Wei J, Li W, Zhang L, Cao C, Zhai J, & Gao B. Accuracy of robot-assisted percutaneous pedicle screw placement under regional anesthesia: a retrospective cohort study. Pain Research and Management 2021 2021 6894001. (https://doi.org/10.1155/2021/6894001)
- 13.↑
Mikhail D, Sarcona J, Mekhail M, & Richstone L. Urologic robotic surgery. Surgical Clinics of North America 2020 100 361–378. (https://doi.org/10.1016/j.suc.2019.12.003)
- 14.↑
Lane T. A short history of robotic surgery. Annals of the Royal College of Surgeons of England 2018 100 5–7. (https://doi.org/10.1308/rcsann.supp1.5)
- 15.↑
Pérez De La Torre RA, Ramanathan S, Williams AL, & Perez-Cruet MJ. Minimally-invasive assisted robotic spine surgery (MARSS). Frontiers in Surgery 2022 9 884247. (https://doi.org/10.3389/fsurg.2022.884247)
- 16.↑
Stull JD, Mangan JJ, Vaccaro AR, & Schroeder GD. Robotic guidance in minimally invasive spine surgery: a review of recent literature and commentary on a developing technology. Current Reviews in Musculoskeletal Medicine 2019 12 245–251. (https://doi.org/10.1007/s12178-019-09558-2)
- 17.↑
Li J, Huang L, Zhou W, Wang Z, Li Z, Zeng L, Liu Z, Shen H, Cai Z, Gu H, et al.Evaluation of a new spinal surgical robotic system of Kirschner wire placement for lumbar fusion: a multi-centre, randomised controlled clinical study. International Journal of Medical Robotics + Computer Assisted Surgery 2021 17 e2207. (https://doi.org/10.1002/rcs.2207)
- 18.↑
Gao S, Zhou L, Cao Y, Li W, Tao H, & Yang Q-g. The accuracy and safety of robot-assisted and free-hand placement of pedicle screws by meta analysis. Journal of Cervicodynia and Lumbodynia 2022 43 1–9. (https://doi.org/10.3969/j.issn.1005-7234.2022.01.001)
- 19.↑
Li HM, Zhang RJ, & Shen CL. Accuracy of pedicle screw placement and clinical outcomes of robot-assisted technique versus conventional freehand technique in spine surgery from nine randomized controlled trials: a meta-analysis. Spine 2020 45 E111–E119. (https://doi.org/10.1097/BRS.0000000000003193)
- 20.↑
Feng S, Tian W, & Wei Y. Clinical Effects of Oblique Lateral interbody Fusion by Conventional Open versus Percutaneous Robot-Assisted Minimally Invasive Pedicle Screw Placement in Elderly Patients. Orthopaedic Surgery 2020 12 86–93. (https://doi.org/10.1111/os.12587)
- 21.↑
Shea BJ, Reeves BC, Wells G, Thuku M, Hamel C, Moran J, Moher D, Tugwell P, Welch V, Kristjansson E, et al.AMSTAR 2: a critical appraisal tool for systematic reviews that include randomised or non-randomised studies of healthcare interventions, or both. BMJ 2017 358 j4008. (https://doi.org/10.1136/bmj.j4008)
- 22.↑
Page MJ, McKenzie JE, Bossuyt PM, Boutron I, Hoffmann TC, Mulrow CD, Shamseer L, Tetzlaff JM, Akl EA, Brennan SE, et al.The PRISMA 2020 statement: an updated guideline for reporting systematic reviews. BMJ 2021 88 105906. (https://doi.org/10.1136/bmj.n71)
- 23.↑
McGrath S, Zhao X, Steele R, Thombs BD, Benedetti A & DEPRESsion Screening Data (DEPRESSD) Collaboration. Estimating the sample mean and standard deviation from commonly reported quantiles in meta-analysis. Statistical Methods in Medical Research 2020 29 2520–2537. (https://doi.org/10.1177/0962280219889080)
- 24.↑
Maggio S, & Sawilowsky SS. A new maximum test via the dependent samples t-test and the Wilcoxon signed-ranks test. Applied Mathematatics 2014 5 110–114. (https://doi.org/10.4236/am.2014.51013)
- 25.↑
Cui GY, Han XG, Wei Y, Liu YJ, He D, Sun YQ, Liu B, & Tian W. Robot-assisted minimally invasive transforaminal lumbar interbody fusion in the treatment of lumbar spondylolisthesis. Orthopaedic Surgery 2021 13 1960–1968. (https://doi.org/10.1111/os.13044)
- 26.↑
Fan M, Liu Y, He D, Han X, Zhao J, Duan F, Liu B, & Tian W. Improved accuracy of cervical spinal surgery with robot-assisted screw insertion: a prospective, randomized, controlled study. Spine 2020 45 285–291. (https://doi.org/10.1097/BRS.0000000000003258)
- 27.↑
Han X, Tian W, Liu Y, et al.Safety and accuracy of robot-assisted versus fluoroscopy-assisted pedicle screw insertion in thoracolumbar spinal surgery: a prospective randomized controlled trial. Journal of Neurosurgery: Spine 2019 1–7. (https://doi.org/10.3171/2018.10.SPINE18487)
- 28.↑
Hyun SJ, Kim KJ, Jahng TA, & Kim HJ. Minimally invasive robotic versus open fluoroscopic-guided spinal instrumented fusions: a randomized controlled trial. Spine 2017 42 353–358. (https://doi.org/10.1097/BRS.0000000000001778)
- 29.↑
Kim HJ, Kang KT, Chun HJ, Hwang JS, Chang BS, Lee CK, & Yeom JS. Comparative study of 1-year clinical and radiological outcomes using robot-assisted pedicle screw fixation and freehand technique in posterior lumbar interbody fusion: a prospective, randomized controlled trial. International Journal of Medical Robotics + Computer Assisted Surgery 2018 14 e1917. (https://doi.org/10.1002/rcs.1917)
- 30.↑
Li Z, Chen J, Zhu QA, Zheng S, Zhong Z, Yang J, Yang D, Jiang H, Jiang W, Zhu Y, et al.A preliminary study of a novel robotic system for pedicle screw fixation: a randomised controlled trial. Journal of Orthopaedic Translation 2020 20 73–79. (https://doi.org/10.1016/j.jot.2019.09.002)
- 31.↑
Ringel F, Stüer C, Reinke A, Preuss A, Behr M, Auer F, Stoffel M, & Meyer B. Accuracy of robot-assisted placement of lumbar and sacral pedicle screws: a prospective randomized comparison to conventional freehand screw implantation. Spine 2012 37 E496–E501. (https://doi.org/10.1097/BRS.0b013e31824b7767)
- 32.↑
Roser F, Tatagiba M, & Maier G. Spinal robotics: current applications and future perspectives. Neurosurgery 2013 72(Supplement 1) 12–18. (https://doi.org/10.1227/NEU.0b013e318270d02c)
- 33.↑
Wang Y, Liu H, Shao S, Fu S, Wang Z, &Hou H.Clinical application of Renaissance spinal surgical robot in thoracolumbar fractures. Journal of Spinal Surgery 2021 19 89–93. (https://doi.org/10.3969/j.issn.1672-2957.2021.02.004)
- 34.↑
Hou H, Shao S, Wang Z, ,Fu S, Liu H, Huang X, Wang H, & Wang L. Application of Renaissance spinal robot in middle and upper thoracic fractures. Chinese Journal of Orthopaedics 2021 41 763–769. (https://doi.org/10.3760/cma.j.cn121113-20210209-00163)
- 35.↑
Wang C, & Kedong H. Orthopedic robot-assisted modified transforaminal lumbar interbody fusion for the treatment of degenerative Safety and clinical efficacy of retrograde lumbar spondylolisthesis. China Medical Journal (English Edition) 2022 57 220–223. (https://doi.org/10.3969/j.issn.1008-1070.2022.02.028)
- 36.↑
Huang J, Han B, & Jixuan L. Clinical effect for orthopaedic robot assisted minimally invasive lumbar internal fixation surgery. Beijing Biomedical Engineering 2020 39 145–151. (https://doi.org/10.3969/j.issn.1002-3208.2020.02.006)
- 37.↑
Xu P, Ge P, & Ren-jie Z. Effect of robot assisted pedicle screw fixation in the treatment of thoracolumbar fracture. Journal of Cervicodynia and Lumbodynia 2018 39 687–690. (https://doi.org/10.3969/j.issn.1005-7234.2018.06.004)
- 38.↑
Wei T, Mingxing F, Xiaoguang H, Zhou J, & Liu Y.Pedicle screw insertion in spine: a randomized comparison study of robot-assisted surgery and fluoroscopy-guided. Journal of Clinical Orthopaedics and Related Research 2016 1 4–10. (https://doi.org/10.19548/j.2096-269x.2016.01.003)
- 39.↑
Zhai G, Gao Y, Gao K, & Zhang J. Robot-assisted versus traditional posterior pedicle screw internal fixation in the treatment of scoliosis. Journal of Clinical Practice Diagnosis and Therapy2019 33 636–640. (https://doi.org/10.13507/j.issn.1674-3474.2019.07.003)
- 40.↑
Yang Rui LYZK. Clinical Applicationand Experienceof PedicleScrewInsertion assisted by Tianjiorthopaedic robot. Journal of Practical Orthopaedics 2019 25 892–897. (https://doi.org/10.13795/j.cnki.sgkz.2019.10.008)
- 41.↑
Cao Z, Xuan T, Yu M, Luo R, & Lu W.Clinical application of TIANJI orthopedic surgical robot in patients treated by adolescent idiopathic scoliosis surgery. Medical Equipment 2021 34 3–6. (https://doi.org/10.3969/j.issn.1002-2376.2021.17.002)
- 42.↑
Li J, Yu M, Liu Z-j, Jin Z, Zeng L, & Li J.Comparison of accuracy of pedicle screw placed by a novel robotic system versus free-hand technique. Orthopedic Journal of China 2020 28 1941–1944. (https://doi.org/10.3977/j.issn.1005-8478.2020.21.05)
- 43.↑
Zhang Y, Zhang Z, & Chaoqun Y. Clinical application and experience of robot-assisted pedicle screw placement in Tiangui. Women's Health 2021 16 25.
- 44.↑
Kim HJ, Lee SH, Chang BS, Lee CK, Lim TO, Hoo LP, Yi JM, & Yeom JS. Monitoring the quality of robot-assisted pedicle screw fixation in the lumbar spine by using a cumulative summation test. Spine 2015 40 87–94. (https://doi.org/10.1097/BRS.0000000000000680)
- 45.↑
Kim HJ, Jung WI, Chang BS, Lee CK, Kang KT, & Yeom JS. A prospective, randomized, controlled trial of robot-assisted vs freehand pedicle screw fixation in spine surgery. International Journal of Medical Robotics + Computer Assisted Surgery 2017 13. (https://doi.org/10.1002/rcs.1779)
- 46.↑
Park SM, Kim HJ, Lee SY, Chang BS, Lee CK, & Yeom JS. Radiographic and clinical outcomes of robot-assisted posterior pedicle screw fixation: two-year results from a randomized controlled trial. Yonsei Medical Journal 2018 59 438–444. (https://doi.org/10.3349/ymj.2018.59.3.438)
- 47.↑
Liu B, Li JJ, Yao SQ, Yang YJ, Wu R, Cong B, Wu QS, Wang ZW, Yu CJ, & Zhou JP. Effectiveness of Robot-Guided versus Freehand Placement of Pedicle Screws in Thoracolumbar Fracture Surgery comparative analysis. Diet and Health 2020 4 44–46.
- 48.↑
Lee NJ, Leung E, Buchanan IA, Geiselmann M, Coury JR, Simhon ME, Zuckerman S, Buchholz AL, Pollina J, Jazini E, et al.A multicenter study of the 5-year trends in robot-assisted spine surgery outcomes and complications. Journal of Spine Surgery (Hong Kong) 2022 8 9–20. (https://doi.org/10.21037/jss-21-102)
- 49.↑
Liu H, Chen W, Wang Z, Lin J, Meng B, & Yang H. Comparison of the accuracy between robot-assisted and conventional freehand pedicle screw placement: a systematic review and meta-analysis. International Journal of Computer Assisted Radiology and Surgery 2016 11 2273–2281. (https://doi.org/10.1007/s11548-016-1448-6)
- 50.↑
Kosmopoulos V, & Schizas C. Pedicle screw placement accuracy: a meta-analysis. Spine 2007 32 E111–E120. (https://doi.org/10.1097/01.brs.0000254048.79024.8b)
- 51.↑
Yu L, Chen X, Margalit A, Peng H, Qiu G, & Qian W. Robot-assisted vs freehand pedicle screw fixation in spine surgery - a systematic review and a meta-analysis of comparative studies. International Journal of Medical Robotics + Computer Assisted Surgery 2018 14 e1892. (https://doi.org/10.1002/rcs.1892)
- 52.↑
Zhang Q, Han XG, Xu YF, Fan MX, Zhao JW, Liu YJ, He D, & Tian W. Robotic navigation during spine surgery. Expert Review of Medical Devices 2020 17 27–32. (https://doi.org/10.1080/17434440.2020.1699405)
- 53.↑
Schatlo B, Martinez R, Alaid A, von Eckardstein K, Akhavan-Sigari R, Hahn A, Stockhammer F, & Rohde V. Unskilled unawareness and the learning curve in robotic spine surgery. Acta Neurochirurgica 2015 157 1819–1823. (https://doi.org/10.1007/s00701-015-2535-0)
- 54.↑
Zhang Q, Han XG, Xu YF, Liu YJ, Liu B, He D, Sun YQ, & Tian W. Robot-assisted versus fluoroscopy-guided pedicle screw placement in transforaminal lumbar interbody fusion for lumbar degenerative disease. World Neurosurgery 2019 125 e429–e434. (https://doi.org/10.1016/j.wneu.2019.01.097)
- 55.↑
Ghasem A, Sharma A, Greif DN, Alam M, & Maaieh MA. The arrival of robotics in spine surgery: a review of the literature. Spine 2018 43 1670–1677. (https://doi.org/10.1097/BRS.0000000000002695)
- 56.↑
Garcia D, Akinduro OO, De Biase G, Sousa-Pinto B, Jerreld DJ, Dholakia R, Borah B, Nottmeier E, Deen HG, Fox WC, et al.Robotic-assisted vs nonrobotic-assisted minimally invasive transforaminal lumbar interbody fusion: a cost-utility analysis. Neurosurgery 2022 90 192–198. (https://doi.org/10.1227/NEU.0000000000001779)
- 57.↑
Huang J, Li Y, & Huang L. Spine surgical robotics: review of the current application and disadvantages for future perspectives. Journal of Robotic Surgery 2020 14 11–16. (https://doi.org/10.1007/s11701-019-00983-6)
