Artificial intelligence, robotics may offer improvements to orthopedic surgery
Click Here to Manage Email Alerts
Artificial intelligence and robotics may provide accurate analysis and improve surgical outcomes of patients undergoing orthopedic surgery, according to presenters.
With no objective digitized analyzing system, Stefan Nehrer, MD, noted clinical assessment of osteoarthritis of the knee through X-rays is subjective and has shortcomings, such as not being able to assess the progression of OA.
“Some studies do not reach clinical evidence because our outcome parameter, the Kellgren-Lawrence score, is not objective enough to do so,” Nehrer said in his presentation at the European Federation of National Associations of Orthopaedics and Traumatology webinar on Practical Application of Artificial Intelligence. “When we look at the joint space narrowing, which is one of the parameters we look at often, we see that it is not assessed by a digital system but only manually measured.”
According to Nehrer, use of artificial intelligence, such as machine learning and deep learning, can assess digitized datasets and identify patterns found in patients with knee OA. He added support of a software system has also been shown to increase the accuracy and precision of knee OA diagnosis, as well as increase the predictability of knee OA.
“If you want to stop the progression of OA, we have to take the earliest timepoint of intervention,” Nehrer said. “To come from a more subjective manual [analysis] to a digitized standardized analysis, that is the goal, and artificial intelligence is promising with that and also already in routine use in some of our institutions.”
Robotics in TKA
Use of robotics with imaging in total knee arthroplasty can also allow surgeons to plan in advance with 3D reconstruction, compare bone registration and ligament evaluation intraoperatively, perform the bone resection and collect feedback, according to Francesco M. Benazzo, MD. He added imageless robotics can allow surgeons to acquire bony landmarks to reconstruct the geometry and morphology of the knee, evaluate ligaments and perform the bone resection with the robotic arm collecting feedback intraoperatively.
“So, the empowering of the surgeon is in data collecting, in planning, in the execution with high accuracy of the surgical act, and we can intraoperatively and immediately postoperatively [check] the results,” Benazzo said in his presentation.
According to Benazzo, robotic surgery can lead to fewer outliers and less damage to the soft tissues of the knee, as well as improve implant alignment, flexion-extension gap and gap asymmetry. However, Benazzo noted robotic TKA does have its limitations, such as installation and maintenance cost, need for additional preoperative imaging, increased operating time and additional incisions for insertion of the registration pins. He added more high-quality studies with longer-term follow-up are needed on functional outcomes, implant survivorship, complications and effectiveness.
“The long road must be followed before we can be sure to say that robotic surgery is better than conventional,” Benazzo said.
Holomedicine in trauma
Finally, holomedicine is a new technology that uses medical software to support physicians during surgery and aids in telehealth and education, according to Lars G. Grossterlinden, MD. With the help of artificial intelligence, Grossterlinden noted surgeons can improve accuracy and accessibility in surgery by merging MRI and CT images to create 3D holograms that can be controlled with gestures and speech commands.
Grossterlinden noted holomedicine has the advantages of viewing patient images from every perspective, which can help as a teaching tool or with educating patients on their injury and surgery. Holomedicine also allows surgeons to connect with colleagues around the world by inviting them to review a specific case, he said.
“If you need the help of a colleague from Bangkok or New York, just invite him and he may have the critical tip on how to ... improve your fracture fixation,” Grossterlinden said.
Despite the advantages with holomedicine, Grossterlinden noted there can be loss of information with the segmentation technique and volume rendering. Holomedicine also requires an ambitious IT service to assist in its use, and some of the software programs are cloud based, which might lead to issues with privacy protection, according to Grossterlinden.
“A dream of mine is for perfect matching and merging of 3D data in the OR ... and then you can mitigate every screw and everything you want to do,” Grossterlinden said. “But still, it is more or less like the same in robotics.”
Perspective
Back to Top
We are entering a new era in orthopedic surgery as we are at the tipping point for using technology in joint replacement: We have surpassed 10% of surgeons using some form of robotics and computer assistance in hip and knee replacement. The next 5 to 10 years in joint replacement will be about data, data acquisition and data science. There will soon be a day when a patient will walk into your office, you will make an analysis and create a patient-specific risk profile, find a target for the surgery, achieve that target with high accuracy and precision using robotics/technology, and then monitor that patient’s outcome, which will then feed back to the original algorithm. This is the core philosophy behind robotics and artificial intelligence in joint replacement.
However, along with this shift come new challenges. With respect to artificial intelligence, there is no question that machine learning and deep learning can digitize datasets and run analyses that no human can ever dream of doing. But there are reports of significant racial biases of the underlying algorithms, skewed against minorities. We must overcome these biases for everyone to trust the information and use it effectively; work is currently underway in these regards.
Collapse
Grossterlinden LG. Holomedicine in trauma care. Presented at: Practical Application of Artificial Intelligence EFORT Webinar. Jan. 25, 2021 (virtual).
Nehrer S. Applications of AI on osteoarthritis. Presented at: Practical Application of Artificial Intelligence EFORT Webinar. Jan. 25, 2021 (virtual).
Read more about
Click Here to Manage Email Alerts