• Ajay Macherla

An Analysis of Artificial Intelligence and Robotics in Neurosurgery


In the technologically-driven society that has come upon humans today, there is a growing desire, or perhaps need, autonomy in every aspect of their lives. Although common in commercial and convenient needs, there is a new possibility of expanding autonomy to the healthcare industry, particularly surgery. Artificial intelligence, machine learning, and surgical robotics are slowly being implemented into surgeries through various system frameworks. However, with this advancement comes several strengths and limitations.

The aim of this study was to examine AI and surgical robotic advancements in neurosurgery and determine whether these technologies are feasible to implement. Through a consideration of pros and cons, the paper will recommend next steps for physicians to take in order to work towards making autonomy a reality in neurosurgery.

It was discovered that AI, ML, and surgical robotics have strengths in the neurosurgery field in that they can increase precision and hand dexterity skills, whether using a fully autonomous robotic system or a supervised one. Imaging enhancements will allow for more safe and accurate procedures as well. Furthermore, the speed of the delivery will mean earlier discharge for patients. There is also less pain associated with robotic equipment due to its precise nature and its ability to operate in tiny incisions through minimally invasive neurosurgery.

However, neurosurgical robotics also carry limitations in that they are not cost-friendly, which can increase surgical procedures prices for patients. Moreover, the lack of human emotion can lead to failures, as evidenced by several studies detailing the casualties that have occured due to robotic failures. Additionally, the technology’s sheer size can take up too much room in the surgery room. Lastly, neurosurgical robotics have the potential to eliminate neurosurgeon jobs, as more autonomy decreases the need for human presence.

However, after a careful analysis of both arguments, it was concluded that neurosurgical robotics are feasible and should be further experimented upon and subjected to multiple clinical trials before they are released to the public. Limitations of the technology were also ruled or made less exaggerated. For these reasons, AI and surgical robotics in neurosurgery is indeed attainable.



As society grows more dependent on technology, there is a question of whether one other aspect of human life can be automated: healthcare and surgery. At a superficial glance, it would appear as if the opportunity should be taken at once––not only does it provide an opportunity to minimize surgical errors and maximize manual dexterity, but it also allows more patients to be treated at once and with ease. However, automation of healthcare may replace human workers faster than it creates more job opportunities, and there comes ethical considerations of which areas would receive the expensive equipment (Carver et al., 2021).

Artificial intelligence (AI) is an ability for technology, computers, or machines to demonstrate natural intelligence comparable to that of the human mind. As a part of AI is machine learning (ML), which is an area of study responsible for developing computer algorithms that improve through experience and data input. AI and ML, when incorporated together more pervasively, could have the possibility of revolutionizing the healthcare industry, including neurosurgery.

AI and ML are already being incorporated in neurosurgery––just not to the extent as some would like, or too much for others, for that matter. In spine surgery, AI is utilized in robotic screw placement. By leveraging computer algorithms and ML, AI is able to analyze the patient’s computed tomography (CT) scan and accurately place the screws in the spine. AI in neuro-oncology imaging has also been utilized. AI can analyze imaging scans of a patient’s brain and compare it with a much larger database––and with greater speed. According to an article, AI was able to improve scan interpretation reliability by over 36% compared to human neuro-oncologists and radiologists (Smith, 2010).

However, at this point of time, AI is currently transforming the non interventional aspects of neurosurgery and healthcare––diagnosis and prognostication. Although it plays a vital role in this area currently, AI can make an even larger impact if combined with surgical robotics, potentially allowing it to enter the surgery room and perform interventional operations on a patient (Panesar et al., 2019).

Prototypes and machines are already being implemented in some facilities for neurosurgery. The telesurgical robot, or the NeuroArm, is a MRI-compatible robotic arm that follows the movement of the neurosurgeon’s arms. By leveraging piezoelectric motors, it is able to provide “tactile feedback” to the neurosurgeon. After showing promising results in animal research, it has made its way to human subjects and has even been involved in over 1000 neurosurgical procedures, including MRI-guided tumor biopsies, microsurgical dissection, and hematoma evacuations (Bagga & Bhattacharyya, 2018). Many more of these neurosurgical robots combine AI and ML to carry out procedures, such as the supervisory surgeon-controlled robot.

However, before AI, ML, and surgical robotics can be implemented further into healthcare, a critical assessment of the pros and cons of these prospects need to be considered. In this paper, we will provide the strengths and limitations of AI and surgical robotics in specifically neurosurgery. The aim of this literature review is, based on analysis, to propose an appraisal on the current situation.



Neurosurgery involves navigating and treating some of the most anatomically complex structures in the human body, such as the brain and the spine. Errors in this type of surgery can be extremely fatal, which is why AI and surgical robotics are emerging as potential alternatives or enhancement devices. In this section, we will attempt to overview the strengths of implementing these robotic solutions in neurosurgery.

Among the most commonly cited reasons for implementing surgical robotics into neurosurgery is the improvement of manual dexterity. An article from Gilreath and Associates cites that medical errors, including errors in dexterity when operating on a patient in the surgery room, accounts for 9.5% of deaths in the United States. This number is alarming considering that it is higher than deaths caused by stroke, accidents, or Alzheimer’s Disease (Behind the Numbers, n.d.). These medical errors are called preventable medical errors for a simple reason: they can be prevented. Proponents of AI and neurosurgical robotics in neurosurgery believe that the pinpoint accuracy of equipment will decrease the probability of preventable medical errors in the surgery room. Furthermore, surgical robotics are more precise in movement and have greater visibility of the patient. They can cross-reference information in large databases and make more informed decisions at lightning speed.

A system called the da Vinci Surgical System offers a unique way to implement surgical robotics to improve dexterity and precision in surgical operations. First, the surgeon works from the computer console in the operating room, controlling “miniaturized instruments mounted on three robotic arms to make tiny incisions in the patient.” After that, the surgeon looks through a 3D camera that is attached to a fourth robotic arm, allowing for greater magnification of the surgical site––this is where increased visibility comes into play. The surgeon’s limb movements are processed and transmitted through the console to the robotic arms. Imitations of the movements allow for motion of the arms and the execution of the surgical procedure (Benefits of Robotic, n.d.).

The da Vinci Surgical System also offers the ability to make less incisions on the patient to access body structures, which decreases the risk of blood loss and infection that may be transmitted from the surgeon. Below is a diagram detailing how the da Vinci system leverages minimally invasive surgery (in the example, it is shown with incisions to the abdomen) to decrease blood loss and pain (Surgical Services, n.d.):

Figure 1: Diagram depicting the difference

in number and size of incisions in surgery by utilizing da Vinci Surgical System

The same article also mentions that oftentimes, patients have an earlier discharge from the hospital because the operation takes less time to complete. This is because manual dexterity enhancements and greater visibility offered by robotic solutions will easen and expedite the surgical process (Surgical Services, n.d.).

An article from the National Library of Medicine, National Institutes of Health, cites three robotic systems that aid in this process: supervisory controlled robotic system, robotic telesurgical system, and shared control system. The supervisory controlled robotic system is a type of solution where robotic intervention into the brain is pre-programmed by the surgeon. There is no input from the surgeon. The robotic telesurgical system, on the other hand, is manipulated in real time by the surgeon. The last system, the shared control system, is a type of robot that enhances the surgeon’s skills through dexterity improvements, such as physiologic tremor reduction (Karas & Chiocca, 2007).

The same article cites that coupling these systems with image-based navigation systems, which utilize AI for cross-referencing capabilities and greater visibility, will allow neurosurgeons to gain “precise target-acquisition attempting intracranial procedures” (Karas & Chiocca, 2007).

Cyberknife, a type of supervisory-controlled system, is frequently used in spinal surgeries. This allows surgeons to monitor the robot as it carries out the procedure. It also allows multiple procedures, such as retraction and dissection, to occur simultaneously without the need of multiple surgeons in the surgical room. Another such technology in spinal surgery is SpineAssist, from MAZOR Surgical Technologies, which is a type of robot for “pedicle and translaminar facet screw placement” (Karas & Chiocca, 2007). The device is basically a passive arm that has pre-defined movement capabilities. One study tested SpineAssist and received stellar results on its functionality and success, and since then, the device is FDA-approved (Choi et al., 2000; Lieberman et al., 2006).

The decrease in blood loss and pain, along with enhancements in manual dexterity and speed of the surgical procedure, indicates that generally, robotic solutions are correlated with better clinical health outcomes in the patient.

Another type of advanced surgical tool is called Robotic Operating Surgical Assistant (ROSA) Brain, which has the ability to autonomously perform minimally invasive procedures in the brain. In fact, ROSA is frequently used by neurosurgeons in treating children with epilepsy and other neurological diseases.

ROSA combines imaging capabilities through AI and ML along with robotic arm work through a supervisory robotic system. An article from Seattle Children’s Hospital cites these benefits of ROSA (Robot-Assisted Neurosurgery, n.d.):

  • Avoids removing part of the skull (craniotomy)

  • Allows for more precise procedures

  • Shortens the time your child is in surgery and under anesthesia

  • Reduces the number of stitches, which means faster recovery and less scarring

  • Lowers the risk of infection because the cuts (incisions) are so small

  • Lowers your child’s pain and their need for pain medicine

However, the article also stresses that ROSA is an option for patients. Multiple meetings, imaging studies, and other sessions are held to help families and children learn more about neurorobotics and how they can help the patient’s condition. Physicians are sure to give strong reasons as to why robotic technology, a seemingly experimental tool at a superficial glance, is the best method for treating the patient.

Multiple robotic technologies are already being implemented and have received compelling results from clinical trials and studies. Of course, more research and experiments must be executed in order to increase the feasibility of the method in neurosurgery, a task that requires precision and accuracy. However, the benefits of robotic tools in neurosurgery is surprisingly plentiful.



While advancements in AI, ML, and surgical robotics seem compelling in neurosurgery––and they have already been implemented with gleaming results––one must also consider the disadvantages of adopting this type of technology in such an important speciality in the medical field. In this section, we will attempt to provide limitations to surgical neurorobotics and provide empirical evidence to support detractors of this technology.

One commonly cited reason is the economic burden of such technologies in neurosurgery. With all of the medical equipment already required for most neurosurgical procedures, along with the costs of paying neurosurgeons, there seems to be little room to invest in such an advanced, sophisticated, and expensive robotic technology. According to an article on the National Library of Medicine, National Institutes of Health, the majority of the economic burden of neurorobotics can be attributed primarily to the initial cost of the equipment, which ranges from about one million to 1.2 million dollars, and yearly maintenance, which can cost upwards from one hundred thousand dollars (Morris, 2005; Tan et al., 2018). With more technological advances, updates, and innovations, the price is expected to increase. This increase in price will ultimately mean that patients will have to pay more for surgical procedures, and pay-cuts from the neurosurgeons end.

Another minor drawback cited by detractors is the sheer size of equipment. Sometimes, surgical robotics can be extremely bulky and take up the majority of the surgery room. Combined with the other equipment the surgeons need, along with ample space for the patient bed and patient, it seems that it is not physically feasible for such technologies to be implemented into surgery rooms without increasing the size of the room––again, another economic burden that contributes to the technology’s infeasibility (Morris, 2005).

It must not be forgotten that just because AI and surgical robotics are robots, they can still make mistakes. Multiple studies have investigated errors in regards to robotic surgery. A study conducted in 2016 by a team of researchers attempted to understand the causes behind robotic surgery failures by performing a comprehensive analysis of adverse events reported publicly in the MAUDE database, regulated by the FDA. They reported that out of the adverse events reported in the database, 144 deaths, 1391 patient injuries, and 8061 (75.9%) robotic device malfunctions. This analysis was performed at the 95% confidence interval. The study notes that instrument malfunctions, such as broken pieces, electrical arcing, unintended operation of robots, software errors, and imaging problems contributed to these adverse events the most. The study concluded that “[d]espite widespread adoption of robotic systems for minimally invasive surgery in the U.S., a non-negligible number of technical difficulties and complications are still being experienced during procedures” (Alemzadeh et al., 2016).

There are still thousands of deaths that are due in part to the nature of AI, ML, and surgical robotics. They do not have human emotions and thus cannot make moral and ethical decisions instantly. Robots follow programming, and should the patient experience any difficulties with the procedure, the robot cannot detect this and will simply proceed with the procedure. This possibility can lead to many deaths and adverse outcomes in the future. An article published in 2015 notes that decision making by surgeons is affected by “tactile perception, visual perception, motor skill, and instrument complexity,” all of which could be skewed by the implementation of neurosurgical robotics. The study concluded that perhaps robotic surgery may not be as obvious as originally thought, as increased precision offered by the technology can be offset by blurred decision making (Randell et al., 2015).

A popular con of robotics in neurosurgery is the elimination of neurosurgeon jobs. As these technologies become more autonomous, there will be less of a need for these physicians to monitor the robots as they perform the procedure. Granted, this possibility is surely not here yet, but it is a legitimate concern for those evaluating the long-term feasibility of neurosurgical robotics.

All of these results hint that the implementation of neurosurgical robotics in neurosurgery may not be as simple as one thought. There are many limitations to the technology that, if not considered and addressed, will lead to significantly more adverse events in the future. A careful analysis of both sides of this issue is necessary in order to come to a favorable and feasible compromise.



Considering both the strengths and limitations of AI, ML, and surgical robotics, a decision as to whether this possibility is feasible and safe needs to be made as soon as possible before further advancements are made in neurosurgical robotics. In this section, we will provide a recommendation as to whether this technology should be implemented based on previous analyses.

It must be admitted that the strengths of AI and neurosurgical robotics is compelling and cannot be refuted, as past literature and empirical studies have proven its success in several clinical trials. Of course, more research and empirical evidence needs to be produced in order to fully accept these technologies in all surgical procedures. However, some of the limitations can be re-examined and re-considered. Let us consider the first limitation.

The economic burden of these technologies cannot be refuted––it is a fact and proponents of these technologies must accept increasing surgical procedure costs. However, as more advancements in robotics occur, the sheer abundance of AI will lower costs to an extent. In short, the economic burden may not be as significant as detractors of neurosurgical robotics originally think, but it is still a noteworthy downside. A similar argument can be put forth for the sheer size of these robotic technologies; although significant, advancements will ultimately decrease its size in the surgery room.

Medical errors due to neurosurgical robotics can occur, but at the same time, similar errors may occur if humans are to operate on patients. Both possibilities must be embraced. To fix this potential problem, we propose a solution. Whenever a patient is admitted into a surgery room, they can fill out a form requesting whether a surgeon or a robot can operate on them––depending on the type of procedure involved. This way, the patient is comfortable with his or her decision, and there is less liability on the surgeon’s end.

As for the elimination of neurosurgeon jobs, it must be noted that most robotic systems are still being controlled by humans. Autonomously running robots are extremely complicated and rare; they are only implemented in select facilities. Although surgeons may need to be trained more in order to operate these robots, they will not completely eliminate jobs in the surgery industry; surgeons will still be needed and they will be useful. Therefore, detractors of AI and ML in neurosurgery may be slightly exaggerated in their views.

For these reasons, we have come to an answer: the implementation of AI, ML, and surgical robotics in neurosurgery should be executed. However, more research and clinical trials need to be conducted in order to validate safety and other concerns before proceeding to actual surgery rooms.



Alemzadeh, H., Raman, J., Leveson, N., Kalbarczyk, Z., & Iyer, R. K. (2016). Adverse Events in Robotic Surgery: A Retrospective Study of 14 Years of FDA Data. PLOS ONE, 11(4), e0151470.

Bagga, V., & Bhattacharyya, D. (2018). Robotics in neurosurgery. The Annals of the Royal College of Surgeons of England, 100(6_sup), 23-26.

Behind the Numbers: Medical Malpractice Death Statistics. (n.d.). Gilreath & Associates.

Benefits of Robotic Surgery. (n.d.). UC Health.

Carver, N., Gupta, V., & Hipskind, J. E. (2021). Medical Error. StatPearls.

Choi, W. W., Green, B. A., & Levi, A. D. O. (2000). Computer-assisted Fluoroscopic Targeting System for Pedicle Screw Insertion. Neurosurgery, 47(4), 872-878.

Karas, C. S., & Chiocca, E. A. (2007). Neurosurgical robotics: A review of brain and spine applications. Journal of Robotic Surgery, 1(1), 39-43.

Lieberman, I. H., Togawa, D., Kayanja, M. M., Reinhardt, M. K., Friedlander, A., Knoller, N., & Benzel, E. C. (2006). Bone-mounted Miniature Robotic Guidance for Pedicle Screw and Translaminar Facet Screw Placement. Neurosurgery, 59(3), 641-650.

Morris, B. (2005). Robotic surgery: applications, limitations, and impact on surgical education. MedGenMed, 3(72).

Panesar, S. S., Kliot, M., Parrish, R., Fernandez-Miranda, J., Cagle, Y., & Britz, G. W. (2019). Promises and Perils of Artificial Intelligence in Neurosurgery. Neurosurgery, 87(1), 33-44.

Randell, R., Alvarado, N., Honey, S., Greenhalgh, J., Gardner, P., Gill, A., Jayne, D., Kotze, A., Pearman, A., & Dowding, D. (2015). Impact of Robotic Surgery on Decision Making: Perspectives of Surgical Teams. AMIA Annu Symp Proc, 1057-1066.

Robot-Assisted Neurosurgery. (n.d.). Seattle Children's Hospital Research Foundation.

Smith, A. (2010, August 10). Artificial Intelligence in Neurosurgery. Rocky Mountain Brain & Spine Institute.,place%20screws%20in%20the%20spine

Surgical Services. (n.d.). MedStar Health.

Tan, Y. P. A., Liverneaux, P., & Wong, J. K. F. (2018). Current Limitations of Surgical Robotics in Reconstructive Plastic Microsurgery. Frontiers in Surgery, 5.

2 views0 comments