Discover the latest AI-enabled orthopedic technologies that are changing orthopedic care as we know it.
We’re at the cusp of a boom of artificially intelligent (AI) orthopedic technologies.
With a growing tidal wave of patient data, a constant industry-wide drive to improve surgical accuracy, and higher post-surgical outcome expectations from patients, AI is sure to have a growing presence in orthopedics — both inside and outside the operating room.
Because of strict review, research, and compliance standards for orthopedic devices, it’s still early days for AI in orthopedics. And while AI will not replace orthopedic surgeons or healthcare professionals altogether, it will undoubtedly have a major impact on the industry by:
- Reducing surgeon burnout: Surgeons can offload time-consuming and stressful tasks to AI systems or receive in-surgery support from surgical robots or augmented reality.
- Increasing surgical accuracy: Orthopedic templating, diagnostic imaging, and surgery-assisting robots will aid in positioning implants and saving healthy bone.
- Improving recovery outcomes: Better preoperative planning and postoperative monitoring will lead to better and less stressful patient outcomes.
1. Patient monitoring apps
To improve the accuracy of its models, AI needs data. The rise in popularity of patient monitoring apps provides a major new data source.
A 2021 study predicts that 70.6 million Americans (over 25% of the population) will use remote patient monitoring (RPM) tools by 2025. RPM uses wearables and apps to allow orthopedic doctors to monitor physiological patient data like vital signs and blood pressure outside a clinical setting. Clinicians can use this data to track at-home recovery and modify treatment plans.
Remote therapeutic monitoring (RTM) offers additional data on musculoskeletal activity and response to therapy. RTM data includes patient-reported outcomes (PROs), where patients can self-report pain intensity and interference.
But RPMs and RTMs aren’t just about monitoring an additional data stream — they can also enhance recovery progress with AI.
For example, Exer Health uses AI to track patients’ mobility and participation in their home exercise protocols (HEPs). Using any smartphone or tablet, Exer Health uses computer vision to accurately measure range of motion and ensure proper form when patients complete their HEPs. This data is then sent to the orthopedic surgeon in a personalized progress report.
The patient benefits from this too. They receive daily alerts to remind them to complete their HEPs, which helps them achieve positive outcomes.
This is an excellent example of where AI can provide additional diagnostic data that would never have been collected by orthopedic doctors or physical therapists — without needing additional expensive equipment.
2. Orthopedic templating software
Digital orthopedic templating, which uses radiographs to create a digital representation of an implant, has already been widely adopted, as it allows for more reliable and accurate sizing, positioning, and alignment of an orthopedic implant.
Studies suggest that 3D digital templating and operative planning achieves at least 90% accuracy, whereas traditional manual planning achieves only 57% accuracy. Now, we’re starting to see AI used to enhance orthopedic templating further.
For example, tools like PeekMed or OrthoPlan 2.0 use AI to automate templating. These tools use AI to perform bone segmentation and landmark detection to suggest the most suitable template and its optimal position.
3. Diagnostic imaging
Recent years have provided us with major academic breakthroughs for AI-powered diagnostic imaging. The primary use case for AI-powered diagnostic imaging in orthopedics is analyzing X-rays to identify implants visually and make clinical diagnoses faster than trained surgeons.
Implant identification
When planning for revisions and reoperations, correctly identifying a patient’s current failed implant is necessary for salvage option planning and equipment ordering.
One 2021 study reported that a deep learning algorithm could differentiate between nine knee arthroplasty implants with 99% accuracy using only plain radiographs. In another 2021 study, an artificial neural network (ANN) model was able to classify existing implants from patient anteroposterior pelvic radiographs with 95.15% accuracy on validation data and 91.16% accuracy in a set of prospective patients. Perhaps most impressively, the same ANN could recognize implants with an average time of only 0.96 seconds using an outdated iPhone 6.
Compared to the time it would take a clinician to identify an existing implant manually, this could be a huge time-saver.
Clinical diagnoses
AI-powered image detection can also be used to make medical diagnoses with similar or better accuracy to trained professionals.
In a 2017 study, a deep-learning network was able to diagnose fractures from orthopedic trauma radiographs with 83% accuracy. This was on par with the performance of two senior orthopedic surgeons who were presented with the same images. Another 2019 study showed that machine learning algorithms could grade osteoarthritis (OA) from radiographs as accurately as arthroplasty surgeons — but in a fraction of the time.
AI-powered image recognition has also been used to diagnose osteoporosis, joint infections, periprosthetic fractures, and periprosthetic component loosening with high accuracy.
4. Data interpretation
The digitization of medical records and new sources for collecting patient data means we suddenly have vast amounts of orthopedic data ready for analysis.
Beyond the time savings we gain from human-caliber diagnostic imaging, AI-powered data analysis will also lead to discoveries and predictive models that go far beyond what human minds could achieve. AI has also been used in a preoperative capacity to improve predictions on length of stay and episode-of-care costs. This will lead to advancements in surgical planning, decision-making, and outcome predictions.
For example, a 2019 study used supervised learning to predict the outcomes of total shoulder arthroplasty procedures based on several patient variables. The machine learning model outperformed standard models for predicting adverse events and surgical site infections.
5. In-surgery tools
We’re just starting to see the potential of AI used during surgeries. So far, the two main use cases of AI for surgeries are in surgery-assisting robots and intelligent operating room systems.
Orthopedic surgery-assisting robots
Surgery-assisting robots are only in their infancy, but they’re expected to break out quickly over the next decade.
These robots work by taking a CT scan before the surgery to plan how much bone should be removed and where the implant should be placed. During surgery, the robot helps the surgeon follow the plan by providing tactile, visual, and auditory feedback to improve stability and mobility.
As of 2021, the orthopedic market had only 15 robots assisting with joint replacement, spine, and trauma procedures. But the overall AI-based surgical robots market is expected to grow to $56.7 billion by 2030 — increasing over 6.5 times from $8.8 billion in 2020.
This predicted growth is based on the fact that surgery-assisting robots benefit both orthopedic surgeons and their patients. For orthopedic surgeons, robots provide:
- More accurate positioning (to an accuracy of 0.5 millimeters)
- Decreased soft tissue damage
- Increased saving of healthy bones
- Better-positioned implants for stability, range of motion, and patient recovery
For orthopedic patients, a 2021 study found that 94% of patients were satisfied after their robotic-assisted total knee replacement (TKR) surgery — significantly higher than the 82% of patients who underwent traditional surgery without robots.
Current surgery-assisting robots include:
AI-assisted surgery platforms
We can also see an emerging class of AI-assisted orthopedic surgery software suites that provides automated support to surgical staff throughout the surgical process.
Zimmer Biomet released Omni Suite, an “intelligent operating room” that assists surgical staff by collecting data on key surgical events, keeping tabs on checklist completion, automating text communication between team members, and providing data displays that help optimize the operating room workflow.
The OrthoGrid AI platform also augments operating room workflows by providing implant-agnostic guidance control, position tracking, real-time image analytics, and enhanced imagery.
6. Augmented reality
Augmented reality (AR) superimposes digital images onto a real space, typically with smart glasses or a headset.
It’s also early days for AR in orthopedics, as most AR systems are still not approved for clinical use. However, a broad range of technology is in development to improve surgical accuracy, decrease operation times, and reduce radiation exposure.
According to a 2021 review, there are currently four significant uses for AR in orthopedics:
- Fracture care: In cadaveric studies, orthopedic surgeons used a see-through, head-mounted display with superimposed 3D images to insert screws with high precision.
- Adult reconstruction: AR has been used in medical training to improve the accuracy of implant orientation in total hip replacements.
- Oncology: Low-cost AR systems have been used as a workstation and position tracker to improve accuracy in tumor resection.
- Spine surgery: One 2019 clinical study saw 253 lumbosacral and thoracic pedicle screws placed with 94.1% accuracy in 20 live patients using an AR surgical navigation system.
In terms of commercially available tools, Surgalign’s HOLO Portal became the world’s first surgical guidance system to incorporate AI and AR in late 2022. The system uses AI to segment and label the anatomy to plan screw trajectories. It then uses AR to overlay the AI-generated plan during the surgical procedure, helping surgeons visualize the trajectories and guide their surgical instruments.
7. Recovery assistance
AI will also lead to more timely and consistent recovery assistance, improving outcomes and the patient experience.
A 2020 clinical study comparing AI-powered telephone follow-up to manual follow-up found that the manual follow-ups took 9.3 hours of human time, whereas the AI system took close to zero. Patient feedback was also more positive for the AI system.
In another 2021 clinical study, patients who used an AI-powered decision aid to inform their selection of TKR for advanced knee OA (compared to those who used only educational materials) showed better decision quality, collaborative decision-making, and satisfaction.
AI-powered patient monitoring apps can also play a role in assessing recovery. Gait monitoring apps, such as Exer Gait or Zimmer Biomet’s WalkAI, can identify patients who may be behind typical recovery curves.
Get ahead of the AI curve
From diagnosis and treatment planning to surgery and rehabilitation, AI will help orthopedic surgeons improve patient care and outcomes over the next few decades. But some AI-powered orthopedic technologies are already available today.
Exer Health captures accurate motion health data for patients recovering at home without needing expensive new technology or hardware. Thanks to in-app form correction and weekly reminders, patients stay motivated and engaged with their treatment plan.
To learn how Exer Health can bring the power of AI to your patient recovery program, request a demo today.
