Global adoption of robotic technology into neurosurgical practice and research

Recent technological advancements have led to the development and implementation of robotic surgery in several specialties, including neurosurgery. Our aim was to carry out a worldwide survey among neurosurgeons to assess the adoption of and attitude toward robotic technology in the neurosurgical operating room and to identify factors associated with use of robotic technology. The online survey was made up of nine or ten compulsory questions and was distributed via the European Association of the Neurosurgical Societies (EANS) and the Congress of Neurological Surgeons (CNS) in February and March 2018. From a total of 7280 neurosurgeons who were sent the survey, we received 406 answers, corresponding to a response rate of 5.6%, mostly from Europe and North America. Overall, 197 neurosurgeons (48.5%) reported having used robotic technology in clinical practice. The highest rates of adoption of robotics were observed for Europe (54%) and North America (51%). Apart from geographical region, only age under 30, female gender, and absence of a non-academic setting were significantly associated with clinical use of robotics. The Mazor family (32%) and ROSA (26%) robots were most commonly reported among robot users. Our study provides a worldwide overview of neurosurgical adoption of robotic technology. Almost half of the surveyed neurosurgeons reported having clinical experience with at least one robotic system. Ongoing and future trials should aim to clarify superiority or non-inferiority of neurosurgical robotic applications and balance these potential benefits with considerations on acquisition and maintenance costs.


Introduction
Neurosurgery is one of the most complex and delicate surgical specialties because of the limited maneuverability determined by the small surgical fields of modern minimally invasive approaches. Furthermore, high-precision standards are required to obtain maximal therapeutic benefits without compromising the function of noble anatomical structures of Vittorio Stumpo and Victor E. Staartjes contributed equally to this work. the central and peripheral nervous system [1]. Recent technological advancements have led to the development and implementation of robotic surgery in several specialties including general surgery, urology, gynecology, endocrine surgery, and orthopedics [2]. In this regard, neurosurgery-despite lagging behind the other specialties in terms of robotic applications because of its very technical peculiarities-constitutes no exception [1], and the practical application of robotic surgery is increasingly reported in the medical literature for the treatment of adult cranial [3], spinal [3][4][5], and pediatric pathologies [6].
Another reason for the rising importance of robotic technology in surgery is the advent of artificial intelligence in medicine. These advances have paved the way for the development of concepts such as the smart operating room, a futuristic surgical theater where human intervention is minimal, information is processed by smart objects, and decisions are made in an automated way. In such a setting, robots will have a major role not only in carrying out the surgical steps according to protocol but also as an intrinsically intelligent mind which can assess the environment and adjust accordingly in real time, or take appropriate actions to prevent errors [7,8].
Even robotic technologies that have been widely applied in other specialties have often demonstrated less than satisfying clinical performance. In light of the increasing appeal that robotics is gaining in the neurosurgical field, its application in routine clinical practice needs to be solidly grounded on evidence, with proof of superiority or non-inferiority compared with traditional neurosurgical interventions [9]. Moreover, in addition to considerations of technical feasibility and possible impact on outcome improvement, the implementation of robotic technology has to take into account also the financial repercussions on the healthcare system inherent to the high acquisition and maintenance costs [10].
While other surveys have tried to describe the status of worldwide applications of new neurosurgical technologies like neuronavigation [11], and despite the encouraging apparent trend in increased applications of neurosurgical robotics with the resulting possible clinical benefit and research advancement, global data on the adoption of robotics in neurosurgical practice and research is currently lacking.
Our aim was to carry out a worldwide survey among neurosurgeons to assess the adoption of and attitude toward robotic technology in the neurosurgical operating room, and to identify factors associated with use of robotic technology.

Sample population
The survey was distributed via the European Association of the Neurosurgical Societies (EANS) and Congress of Neurological Surgeons (CNS) in January, February, and March 2019. The EANS is the professional organization that represents European neurosurgeons. An e-mail invitation was sent through the EANS newsletter on January 28, 2019. Furthermore, the membership database of the CNS was searched for e-mail addresses of active members and congress attendants. The CNS is a professional, United States-based (US) organization that represents neurosurgeons worldwide. At the time of the search, the database contained 9007 members from all continents, a subset of which had functioning email addresses. The survey was hosted by SurveyMonkey (San Matea, CA (USA)) and sent by e-mail together with an invitation letter. Reminders were sent after 2 and 4 weeks to non-responders to increase the response rate. To limit answers to unique site visitors, each e-mail address was only allowed to fill in the survey once. All answers were captured anonymously. No incentives were provided.

Survey content
The online survey was made up of nine or ten compulsory questions, depending on the participants' choice of whether they had or had not used robotic technology in their neurosurgical practice. A complete overview of survey questions and response options is provided in Table 1. The order in which potential reasons for use/non-use are displayed was randomized to avoid systematic bias. The definition of robotic technologies that was provided within the survey was: "Any form of robotic assistance in neurosurgery, including but not limited to cooperative robot arms and modules ("cobots") assisting in surgical maneuvers such as pedicle screw placement, endoscopy, radiosurgery, microscopy, biopsy, or DBS electrode placement, etc." The survey was developed by the authors based on prior, similar surveys carried out in a similar population. This report was constructed according to the Checklist for Reporting Results of Internet E-Surveys (CHERRIES) guidelines [12].

Statistical analysis
Continuous variables are given as means ± standard deviations (SD), whereas categorical variables are reported as numbers (percentages). Countries were grouped by region (Europe/North America/Latin America/Asia & Pacific/ Middle East/Africa) according to a previous worldwide survey by Härtl et al. [11]. Fisher's exact test was applied to compare implementation incidence of robotics among regions. By use of a multivariate logistic regression model, we identified independent predictors of adoption of robotic technology into clinical practice and research, respectively. The importance of reasons for use or non-use of robotics was compared among regions using the Kruskal-Wallis H tests. When calculating the ratio of respondents who had applied robotic technology in research, we incorporated both respondents who had never used robotics in their research and those who do not participate in medical research into the denominator. R version 3.5.2 (The R Foundation for Statistical Computing, Vienna, Austria) was applied for all analyses, and the Type I error rate was defined as p ≤ 0.05 for two-tailed tests.

Response rate and respondent characteristics
From a total of 7280 neurosurgeons who were sent the survey, we received 406 answers, corresponding to a response rate of 5.6%. Detailed characteristics of the respondents are given in Table 2. The majority of respondents were in the 30-40 years age group (33%), and 88.7% of the answers were from male participants. Most of surveyed neurosurgeons were specialized in spinal surgery (34.5%). As far as the work setting was concerned, more than two-thirds of the neurosurgeons were practicing in an academic hospital (67.7%), followed by non-academic hospital (15.5%), private practice (15%), and other settings (1.7%). We also sought to describe the level of experience of the surveyed population. Participants were mostly board-certified/attending neurosurgeons (58.9%), while residents (20%), chairs of department (10.8%), fellows (4.7%), medical students (3.2%), and others (2.5%) were less represented. Geographic distribution of the answers was skewed in favor of North America (70.4%) and Europe (17.2%), while less answers were received from surgeons from Asia and Pacific (5.4%), Latin America (3.9%), Middle East (2.5%), and Africa (0.5%).

Robotics in clinical practice and research
When inquired about the use of robots in neurosurgical clinical practice and research, 48.5% and 61.5% of the surveyed population answered positively, respectively. Stratified by region ( Table 3), use of robotic technology in clinical practice was most common in Europe (54.3%) and North America (51.4%), followed by Asia and Pacific (31.8%), Middle East (20.0%), Latin America (18.8%), and Africa (0.0%). Figure 1 provides a graphical illustration of the worldwide clinical use of robotics in neurosurgery. Respondents were also asked to list which types of robots they had worked with ( Table 4). The most commonly used robotic devices were from the Mazor family (32%), followed by the ROSA robot (26.4%). A high proportion of the robot users did not identify the specific type of robots that they had used (33.5%).

Predictors of robotics use
Multivariate logistic regression analysis was used to investigate independent predictors of adoption of robotics into clinical practice and research (

Attitudes toward robotic technology in neurosurgery
The surveyed population was asked to rate the importance of the factors for and against the use of robotic technology in neurosurgical clinical practice (Table 6). Among those surgeons implementing the use of robotic technology, the perceived improved surgical outcome (3.3 ± 0.9) and marketing considerations for augmentation of patient referrals (3.2 ± 0.9) were rated the most important, followed by time savings (2.7 ± 1.0), lower risk of complications (2.7 ± 1.0), and costeffectiveness (2.3 ± 1.0). Only for time savings, we identified a significant difference in importance rating among the five regions (Kruskal-Wallis test, p = 0.003)-time savings were rated highly important in the Middle East and in Asia and Pacific, while this potential advantage was only of minor importance in Latin America. Among those neurosurgeons who had never used robotics in clinical practice, the most important factor prohibiting adoption of robotics into clinical practice was the inherent acquisition/maintenance costs (3.4 ± 0.9). Other considerations played a lesser role in this choice. Of note, a statistically significant imbalance was found among regions with respect to difficulties with staff training and device education and also of personal convincement of the added value granted by the implementation of robotics in surgical practice (Kruskal-Wallis test, p = 0.030 and p = 0.008 respectively).

Discussion
Our survey addressed a geographically diverse cohort of neurosurgeons at different levels of training. It is apparent that robotic surgery seems to have gained wide acceptance in neurosurgical practice as confirmed by the observation that almost half of the surveyed population have used robotic technology during neurosurgical procedures. Furthermore, around one-fifth of the surveyed population appears to have engaged in medical research using robotic technology. The overwhelming majority of robotics users was to be found in individuals under 40 years of age. Spinal surgery was the subspecialty that most often applied robotics, followed by neuro-oncologists, and cerebrovascular specialists. The most commonly used devices were the Mazor family and ROSA robots.
The proportion of neurosurgeons who reported having used robotic technology in clinical practice was very high and certainly higher than expected. Although, with recent trends, these numbers are conceivable, there are some factors that may potentially have led to a higher proportion of neurosurgeons reporting use of robotics in the surveyed population. First, the survey was circulated among EANS/ CNS members and congress attendants, by way of which a potentially more scientifically interested and academic population was selected for. As observed in our survey, academic neurosurgeons are far more likely to have had contact with robotic surgery than their non-academic counterparts are. Second, it is possible and conceivable that among the population that was sent this survey, the surgeons with prior experience with robotics were more interested in this topic and therefore more likely to fill in a survey on robotic surgery (response bias). Even though these potential biases may have increased the proportion of neurosurgeons reporting clinical use of robotic technology, our results demonstrate that in recent years, robotics has seen broad adoption into the neurosurgical operating rooms of particularly Europe and North America.
After adjustment for potential confounders, no subspecialty was found to be significantly associated with an increased or decreased robotics use, neither in clinical practice nor in research. This suggests that robotic technology has been rather broadly applied in many neurosurgical subspecialties and for the treatment of several different pathologies. The main reasons guiding the increased implementation into clinical practice were the perceived improved surgical outcome granted by robotics as well as marketing considerations, potentially  leading to more patient referrals. Predictably, adoption of robotic surgery into clinical practice was more frequent among younger surgeons, particularly those under 30, and less common in physicians practicing in non-academic centers. The fact that the use of robots in neurosurgery was particularly frequent in those < 30 years of age shows that neurosurgeons have increasingly contact with robotic technology during their residency training. The lower odds ratio identified for male respondents, may reflect an increased representation of the female population among the younger neurosurgeons and an encouraging trend in terms of closing the existing gender gap in neurosurgery [13,14]. A statistically significantly decreased application of robotic surgery into clinical practice was found in Asia and Pacific and the Middle East compared with Europe and North America. In addition, lower clinical adoption was observed in Latin America and Africa, but this effect was not statistically significant due to the low sample size. These findings are compatible to the potentially decreased availability of resources in some of the countries belonging to the aforementioned regions. This hypothesis is also confirmed by a trend toward higher scores obtained for acquisition and maintenance costs as a reason for non-use of robotics with respect to other countries.

Robotics in neurosurgery
The very definition of robotics poses some difficulties in identifying how neurosurgery is adapting to this increasingly evolving field. To date, most surgical robotics are very limited in their ability to perform procedures and make decisions automatically without major human intervention. Therefore, several other classifications have been proposed to describe surgical robots, based on one side on the device's function and application, and on the other on the surgeon-robot interaction [15]. In fact, robotics far from only substituting and transforming the surgical act of the physician through automation and remote control has also been increasingly adopted for assisting specific surgical tasks, for example, anatomical localization of the lesion, stabilization of the surgeon's hand during prolonged microsurgical work, or pedicle screw insertion [16,17]. Moreover, the inherent complexity of neurosurgical procedures often requires different robotic competencies in different phases of surgery [1]. This kind of robotic aid is more precisely referred to as "cobot surgery", where robotics enhance and maximize specific parts of the surgical procedure without performing automatic actions. Regardless, the use of robotic systems has been increasingly often reported in the neurosurgical literature, both for cranial and spinal applications [16,18]. Table 7 provides an overview of relevant publications on the most recent developments of robotics in the field of neurosurgery.

Spinal applications
Several robotic systems are available for spinal interventions, mostly for assistance in pedicle screw placement [19]. Recent literature reported that robot-assisted screw placement is at least non-inferior if not superior with respect to accuracy than conventional free-hand technique and potentially decreases the rate of revision procedures [5,17,[20][21][22][23][24]. A recent paper byJosephetal.systematicallyreviewedapplicationsofrobotics in spinal surgery [18]. The authors reported that most comparative studies-apart from 1 RCT [25]-demonstrated that robotics can provide increased radiological accuracy with respect to free-hand placement both with the Mazor family and ROSA robots. A recent meta-analysis investigating clinically relevant pedicle screw revision in robotic-guided, navigated andfreehandthoracolumbarinstrumentationsfoundthatboth robotics and navigation reduced post-operative revisions, but statistical significance was lost at sensitivity analysis for the former [9]. When length of hospital stay and overall complications were evaluated, Siccoli et al. showed that free-hand thoracolumbar screw insertion had worse results with respect to navigation, while no difference was found with robotguidedsurgery [26].Onthecontrary,nosignificantdifference was found when radiation exposure was compared between robot-guided, navigated surgery, and free-hand approach [26]. More recently, a meta-analysis by Fatima et al. reported that perfect and acceptable pedicle screw accuracy as categorized by Gerztbein-Robbin classification was higher in robotassisted than in free-hand surgery; complication rate, proximal facet joint violation, and intra-operative radiation time and exposure were significantly lower, while length of surgery was significantly higher [27]. Table 8 summarizes the results of most recent meta-analyses comparing robotassisted spine surgery with navigated and free-hand technique.
Highly powered ongoing prospective studies like the European Robotic Spinal Instrumentation (EUROSPIN) [12] and MIS-ReFRESH [7] studies are necessary to investigate if these potential benefits warrant the high acquisition and maintenance costs of these systems.

Neuro-oncology
Robotic applications can also find applications in neuro-oncology. Most notably-of course also because invented by a neurosurgeon-the CyberKnife is one worldwide-adopted robot that is frequently used to treat tumors of all kinds using frameless stereotactic radiosurgery [28]. As other examples, robot-guided convection-enhanced delivery of chemotherapy for brainstem glioma was reported whereby the feasibility of accurately and safely delivering very small diameter catheters to deep targets within the brainstem was demonstrated [29]. Another example is the NeuRobot, a remotely controlled endoscope for tele-controlled tumor resection [30], which has been proven to be useful also for intraventricular dissections [31].

Cerebrovascular/endovascular neurosurgery
Robotics is also gaining momentum in cerebrovascular and endovascular neurosurgery [32]. Currently tested applications (in vitro and in vivo) include cerebral angiography (also a robotic digital subtraction angiography (DSA) system), Time savings 2.7 ± 1.0 2.9 ± 0.9 2.4 ± 1.1 1.7 ± 0.5 3.5 ± 0.7 3.0 ± 1.4 NA 0.003* Improved surgical outcome 3.3 ± 0.9 3.4 ± 0.9 2.9 ± 1.1 2.9 ± 1. Importance is presented as mean ± SD. The importance of reasons for use or non-use of robotics was compared among regions using the Kruskal-Wallis H tests *p ≤ 0.05 robot-assisted operating microscopes for the treatment of arteriovenous malformations and cavernomas, mechanical coil insertion systems for aneurysm treatment (reducing the number of operators needed for the procedure from two to one), and robotic endoscopic aneurysm clipping [33][34][35]. Moreover, several robotic systems that are already approved for clinical applications in other specialties like interventional cardiology and radiology may find fertile soil in neurosurgery after appropriate modifications [36].

Other cranial applications
Other clinical applications of robotics systems in cranial neurosurgery include stereotactic biopsy targeting, deep brain stimulation (DBS) electrode placement, radiosurgery, placement of stereoelectroencephalographic (SEEG) electrodes for investigation of refractory epilepsy, ventricular catheter placement, and laser ablative procedures [16]. Growing interest is currently being placed on exoscopic camera systems to improve illumination and depth-of-field when difficult-to-access or deep lesions limit the visibility, although their potential advantages over traditional operating microscopes still remain questionable. For example, several small case series have addressed the efficacy and safety of the Synaptive Modus V exoscope system in both spinal and cranial surgery, with encouraging results [37].

Limitations
Survey-based studies, while providing important insights, have inherent limits because of several potential biases. During survey distribution, selection and response bias are possible. Time constraints on responders may have limited their ability to answer with maximal accuracy, and in fact, concerning the adoption of robotic systems into clinical research, we obtained several incomplete or blank answers. The data is mostly based on subjective impressions of surgeons. Knowing this, bias could arise from the fact that surgeons who are more exposed to neurosurgical robotics can value it more positively than those who do not routinely make use of it, and vice-versa. However, reasons for advantages and disadvantages were specifically captured separately for users and non-users. Additionally, the relative percentage of geographic regions was skewed in favor of western countries, limiting the sensitivity of our survey for what concerns regions such as Asia and Pacific, South America, and in particular Africa.

Conclusions
Our study provides a worldwide overview of neurosurgical adoption of robotic technology. Robotic systems have the technical potential to improve surgical procedures in terms of efficacy and safety by several means, spanning from indirect assistance of surgeons in complex parts of the operation (such as lesion localization) to more or less integral substitution of the manual skills required by the surgical task. Our survey sheds light on the diffusion of such technology and their general perception by neurosurgical specialists. Almost half of the surveyed neurosurgeons reported having clinical experience with at least one robotic system. The Mazor family and ROSA robots were most commonly applied. Before a consistent and widespread shift in clinical practice, superiority or non-inferiority of neurosurgical robotic applications needs to be established by high level of evidence studies and, at the same time, carefully balanced with considerations on costs of implementation. The results of ongoing and future trials will clarify which neurosurgical robotic applications can routinely enter clinical practice and can determine the relative extent of the potential clinical benefits granted by the integration and technical refinement of robotic technology.