SurgiCorder® : Capturing the Value of Video and AI in Surgery

The SurgiCorder® Deep Research Report

Introducing «SurgiCorder® – for your surgical intelligence» — the next-generation AI-powered platform transforming surgical documentation, training, intraoperative assistance, and performance evaluation.

Harnessing the power of META ARIA Smart Glasses, NVIDIA AI technology, and advanced Large Language and Vision Assistant (LLaVa) models, SurgiCorder® delivers unmatched analysis and guidance of live surgical procedures.

SurgiCorder® seamlessly combines visual and textual data, ensuring consistent documentation and delivering objective, data-driven feedback. With its ability to generate instructional content and apply cutting-edge machine learning techniques, SurgiCorder® not only enhances surgical skills but also drives forward evidence-based surgery. By addressing technical performance bias and improving homogeneity in treatment groups, SurgiCorder® is set to transform surgical education, quality assurance, and patient outcomes.

«Elevate your surgical expertise with SurgiCorder® — where AI and surgical excellence converge»

Insight 1: Video Data – The Highest-Value Asset for Surgical AI

Rich Information Content of Video vs. Other Data: Surgical video is uniquely data-rich, capturing the full visual context and sequence of a procedure in a way that other data sources cannot. Electronic medical records (EMRs) and sensor logs contain important information (patient history, device readings, timestamps), but they only offer discrete snapshots or summaries. In contrast, video provides a continuous, detailed record of the surgery itself – every instrument maneuver, anatomical landmark, and team interaction. This makes video a goldmine for AI algorithms, which thrive on large, detailed datasets. As industry expert Joe Mullings observed in his SAGES 2025 debrief, “Video has the highest value of data when we think about the use of AI” (linkedin.com). By analyzing raw video, AI can infer granular events and context that would never be documented in an EMR. For example, an EMR might note that a procedure took 90 minutes, but only video can show how that time was spent (which steps were slow, where complications occurred, how tools were used). Video can also be re-analyzed for new insights without needing prospective data collection, essentially “unlocking” hidden information from past cases.

Comparing Video to EMR and Sensor Data: Traditional surgical data sources have significant limitations for AI use. EMR data is primarily designed for billing and record-keeping – not real-time analytics – and often suffers from delays or inaccuracies. As Houston Methodist’s Chief Innovation Officer explained, “Historically, we counted on the EHR for all of our data. But that’s meant for record-keeping and billing, not for management and analysis” (healthcareitnews.com). Key operational details (like precise start/end times of each surgical phase) may be inconsistently recorded if reliant on manual input. Sensor data (such as vital signs, or even robotic instrument telemetry) can provide continuous streams, but each sensor only measures a narrow slice of the OR (for instance, a robot’s force sensor won’t tell you if the team did a safety timeout). Video, on the other hand, captures multiple dimensions – the patient, the tools, the OR staff – all in one feed. Modern “OR Black Box” systems illustrate this by synchronizing video with other data: for example, Teodor Grantcharov’s OR Black Box records “OR audio-video recordings, patient vitals, and surgical instrument usage feedback” to get a comprehensive view of each surgery (facs.org). In effect, video serves as the master record that can be augmented with or validated against other data sources. This rich, unstructured nature of video was precisely why Mullings and other experts at SAGES emphasized its value: whoever controls the surgical video data can derive the most actionable insights for improving surgery.

Advances in Computer Vision Enabling Video Analytics: Until recently, video was an underutilized asset in surgery because we lacked the tools to effectively analyze it at scale. Now, breakthroughs in computer vision (CV) and deep learning have made it possible to extract structured information from surgical footage. AI algorithms can be trained to recognize surgical instruments, anatomical structures, and even surgical actions or steps in real time. For example, Medtronic’s Touch Surgery platform has introduced AI models that automatically detect surgical workflow phases, instrument usage, and anatomy in laparoscopic and robotic videos (surgicalroboticstechnology.com). These models can translate raw video into a timeline of labeled events – e.g. identifying when a cholecystectomy has moved from dissection to clipping, or how often and for how long a particular instrument is used. The availability of large datasets is fueling this progress: companies like Theator have amassed libraries of tens of thousands of surgical videos (over “30,000+ hours… with a billion frames”) to develop reliable computer vision algorithms for video review and metrics analysis (facs.org). Such high-volume, diverse video data allows AI to learn the visual patterns of surgical procedures with increasing accuracy. In addition, the advent of transformer-based vision models and multimodal AI (combining vision with text or audio) means an AI can not only see what is happening in the OR, but also contextualize it against surgical guidelines or dialogue. In short, the technical barriers to harnessing video are coming down – making surgical video a ripe target for AI-driven innovation in the OR.

Why Video AI Is Well-Suited to the OR Environment: The operating room is a complex, high-stakes environment where many things happen at once. AI that can interpret video in real time is especially useful here because so much of what matters in surgery is visual (or at least observable). A computer vision system can continuously watch the surgical field and the OR scene without distraction or fatigue, potentially catching subtleties that a human might miss. Modern ORs already have multiple cameras – e.g. overhead cameras, endoscopic feeds, laparoscopic cameras – providing vantage points for AI analysis. Unlike data that come in numerical streams, video can capture physical processes (a surgeon’s technique, a sequence of actions) that are directly tied to outcomes. For instance, by analyzing video an AI could notice that a surgeon skipped a step in a protocol or used an unconventional technique, whereas pure numeric data wouldn’t reflect that. Moreover, video analysis doesn’t require any change in surgical practice to collect – it’s passively recorded, not interrupting workflow the way manual data entry would. The OR environment also benefits from multimodal AI – combining video with audio analysis. In surgery, crucial information is often conveyed through discussion or audible cues (e.g. a monitor alarm sounding, or a surgeon saying “clamp please”). AI that fuses audio and visual data can form a more complete understanding of the situation. (For example, if an instrument is obscured on camera but the scrub nurse calls out its name, an AI could still log that the tool was used at that moment.) This multimodal capability is becoming feasible with new AI architectures. Overall, the OR’s dynamic nature makes real-time video analytics a powerful approach: it’s like having an ever-vigilant assistant observing the procedure from all angles, ready to derive insights that improve care.

Use Cases Unlocked by AI on Surgical Video: High-value video data, paired with AI, is enabling a range of impactful applications in surgery:

Surgical Skill Coaching & Training: Video provides a basis for objective feedback on surgical technique. Platforms like Caresyntax’s InfluenceOR™ and Johnson & Johnson’s C-SATS leverage video recordings to evaluate a surgeon’s performance and suggest improvements. InfluenceOR allows board-certified experts to review a surgeon’s video and provide analysis of technique and decision-making, turning operative footage into a personalized coaching tool (facs.org). J&J’s C-SATS (acquired in 2018) goes further by combining crowd-sourced expert ratings with AI; uploaded surgical videos are analyzed to produce “actionable and objective insight” on technical skill (facs.org). These services generate metrics like economy of motion, instrument handling, and adherence to best practices, helping surgeons (and residents in training) identify specific areas to improve. The key is that video doesn’t lie – it captures exactly what was done, enabling fair comparisons and benchmarking. Surgical societies are increasingly interested in video-based assessment as a more reproducible and educational form of evaluation, since it provides qualitative feedback in a non-punitive way (facs.org). In the future, AI might even automate parts of this assessment, instantly gauging proficiency or flagging errors in a recording. This use of video+AI directly feeds the goal of data-driven improvement of surgical skills.
Intraoperative Decision Support and Guidance: While still emerging, there are initiatives to use live video analysis to guide surgeons during a procedure. An AI watching the endoscopic feed or an overhead camera could potentially warn the surgeon of a risk (e.g. an instrument dangerously close to a critical structure) or highlight a region of interest (like a tumor margin that hasn’t been resected). Augmented reality (AR) overlays in the surgeon’s field of view are being explored to augment human perception with AI insights. For example, an AI might recognize from the video that a required surgical step hasn’t occurred yet and prompt the team (“Have you placed the safety clip on the artery?”). Such real-time guidance is challenging to implement (it requires extremely robust recognition and minimal latency), but it is a natural evolution now that video analytics are becoming faster and more accurate. Even audio can play a role – an “AI assistant” could listen to intraoperative conversations and remind the team of checklist items or milestones (“the sponge count hasn’t been confirmed yet”). Overall, leveraging video in real time can make the OR smarter and potentially head off mistakes before they happen.
Automatic Surgical Documentation: Video analysis can dramatically streamline documentation by automatically recording what happens during surgery. Today, surgeons and OR staff spend significant time after procedures documenting key events, which is tedious and prone to omissions. AI can use video (and audio) to generate an automatic timeline of the case – for instance, noting “incision made at 10:03, first instrument insertion at 10:07, closure started at 10:45,” etc., all by analyzing the video feed. Apella, for example, uses computer vision on OR video to log timestamps like wheels-in, incision start, closure, and wheels-out without anyone having to input data (healthcareitnews.com). These time stamps and events can be fed into the electronic record or a report. Further, an AI might transcribe spoken operative findings (via microphone input) or count how many times a device was used. The end result is a High-Fidelity Surgical Record, as Caresyntax calls it, which integrates video, imaging, device data and more for a complete case record (facs.org). Automating documentation not only saves time, but also improves accuracy – ensuring that the surgical record truly reflects what occurred. It can also facilitate better billing and quality reporting by capturing all billable events and any deviations or complications.
Workflow Optimization & OR Efficiency: One of the most pressing use cases – and a major focus at SAGES 2025 – is using AI on video data to optimize OR workflow and throughput. Hospitals are turning to video analytics to identify bottlenecks in surgical flow (e.g. prolonged turnover times, delays in incision after anesthesia, etc.) and then address them. Apella’s platform is a prime example: it employs ambient cameras and AI to “provide a 360-degree view of the operating room” and generate a real-time feed of operational metrics (healthcareitnews.com). At Houston Methodist Hospital, this approach yielded a 15% increase in OR capacity without additional staff, equating to ~33 more cases per month – a significant efficiency gain (healthcareitnews.com). The AI was able to capture events more reliably than manual logging, giving staff confidence to adjust schedules on the fly and fit in more cases (healthcareitnews.com). Similarly, UK-based VitVio uses ceiling-mounted cameras and computer vision to “track and analyze every aspect of the surgical procedure – from tool usage and surgical stages to adherence to safety protocols – in real time”, presenting live updates on a dashboard (businesswire.com). This helps OR managers predict case end times, reduce idle gaps, and coordinate staff more dynamically. By catching inefficiencies (like unneeded delays or idle times) through video, AI can recommend process changes that let surgeons safely do more procedures per day. Given that operating theaters account for ~40% of hospital expenses (businesswire.com), even modest improvements in utilization have huge financial and patient-access impact. AI-driven video analytics thus directly translate to doing more surgery with the same resources, a critical need as surgical demand rises.
Safety and Quality Improvement: Video review has long been used in other high-risk fields (like aviation) to investigate errors and improve safety. In surgery, AI-enhanced video capture enables a similar “black box” concept for continuous quality improvement. Every near-miss or adverse event in the OR can be studied after the fact by reviewing the synced video, audio, and device data, allowing teams to learn and prevent recurrence. Companies like OSI (Surgical Safety Technologies) with the OR Black Box have shown that continuous recording fosters a culture of accountability and transparency in the OR (facs.org). AI can aid by flagging portions of the video where an anomaly occurred – perhaps detecting a rapid change in vital signs on the monitors, a loud exclamation, or an instrument count discrepancy – to direct attention for review. Over time, aggregating this data can highlight systemic issues (for example, a certain instrument is frequently associated with delays or errors). Video-based analysis also supports compliance with safety protocols: an AI might check if the surgical timeout was performed (listening for it via audio) or if all team members were sterile at the field (via video). These applications improve patient safety by ensuring that best practices are followed and giving hospitals actionable intelligence on where complications originate.

Industry Momentum and Strategic Importance of Video-AI Capabilities: The recognition that “video has the highest value of data” (linkedin.com) for surgical AI has led to a flurry of activity in the medtech ecosystem. Established medtech companies and startups alike are racing to build or integrate digital surgery platforms that capitalize on video analytics. At SAGES 2025, Mullings noted the surge of digital surgery booths and predicted that major “visualization” vendors (like Karl Storz, Olympus, Stryker – companies known for endoscopic towers and cameras) may soon ramp up acquisitions in this space (linkedin.com). The strategic rationale is clear: an endoscopic tower sits in virtually every OR and is a hub for video data. If that video feed can be captured and analyzed by AI, it becomes incredibly valuable. Companies that traditionally only sold imaging hardware are realizing they may need to provide AI and data services on top or risk being left behind. On the other side, the large surgical robotics players – Intuitive Surgical, Johnson & Johnson, Medtronic – also recognize that owning the surgical video and analytics is key to “owning” the digital OR ecosystem. Matt Krueger (President at Caresyntax) commented that while the robot companies dominate robotic surgery, they “lack the experience in how to sell and deploy IoT and workflow solutions” across all ORs (linkedin.com). He pointed out that Intuitive, J&J, and Medtronic each have digital OR offerings (Intuitive’s “Hub”, J&J’s surgical platform referenced as Polyphonic (“The digital ecosystem for surgery: Unlocking the power of surgical data – Video is a rich data source in the OR” thenext.jnjmedtech.com/digital), and Medtronic’s Touch Surgery) but to truly win in this space, “you also need data rights to the video for that tech stack and footprint to be meaningful.” (linkedin.com) In other words, controlling surgical video data is the linchpin of the future surgical platform – it’s not enough to have smart devices; one must capture the procedure itself. This has led to notable acquisitions and partnerships: J&J acquiring C-SATS for video-based skill assessment, Medtronic acquiring Digital Surgery (Touch Surgery’s maker) to build its AI analytics, and most recently a proliferation of partnerships between hospitals and AI startups to pilot these technologies, e.g. Cleveland Clinic with Caresyntax, Royal Orthopaedic Hospital with VitVio (businesswire.com), etc.. Even vendor-neutral platforms like Proximie (originally focused on telepresence) have started emphasizing their role in capturing and leveraging procedural recordings across institutions, because that content can be mined for insights and training. Open, interoperable platforms are gaining traction (linkedin.com), suggesting that hospitals prefer solutions that can plug into any camera or imaging system rather than proprietary closed loops. This openness accelerates innovation and data sharing – for example, a hospital might use Stryker cameras with a third-party AI analytics tool like Apella or VitVio to get the best of both worlds.

For MedTech companies, the strategic value of video-AI capabilities lies in the data network effects: the more surgical video they analyze, the smarter their algorithms get and the more value they can deliver (creating a competitive moat). Additionally, offering AI analytics can transform a one-time sale (of a device or camera) into a recurring revenue service model, which is very attractive business-wise. In summary, the entire medtech ecosystem sees surgical video analytics as a critical frontier – one that will shape which companies lead in the next decade of digital surgery.

Insight 2: AI-Powered Autonomous Monitoring of OR Procedures

“Autonomously track & analyze procedures, tool usage, processes in the OR, using audio and video along with AI to increase number of procedures and optimize OR efficiencies.” (linkedin.com)

This second key insight from SAGES 2025 describes a future (already becoming reality) where surgeries are continuously observed by AI “agents” that log every instrument and every step, ideally improving throughput and efficiency without human intervention. Breaking this down technically and operationally:

From Video/Audio to Real-Time OR Intelligence: Autonomous monitoring means the AI system watches and listens to the OR in real time, and interprets the surgical process automatically. This involves several layers of AI capability:

Procedure Tracking: The AI must recognize which stage of the operation is happening at any given time. For example, in a laparoscopic cholecystectomy, stages might include: port placement, exposure, clipping, cutting, gallbladder removal, closure. Vision algorithms classify video frames (or short video segments) into these phases based on patterns of motion and surgical scene context. By tracking the progression, the system can estimate how much time is left or detect if a step is taking unusually long.
Tool Usage Monitoring: The system uses object detection on video (and possibly instrument RFID or sensor data, if available) to identify each surgical instrument and track its usage frequency and duration. For instance, the AI can detect when a scalpel or a laparoscopic grasper appears in view and log “Instrument X in use from 10:07 to 10:12”. Modern computer vision models are capable of detecting dozens of tool types from video frames, even in the presence of blood or poor lighting, after being trained on annotated surgical videos. Audio can assist too – if a surgeon says “scalpel” or a nurse confirms an instrument count, that audio cue can corroborate the visual detection.
Process/Workflow Analysis: Beyond the surgical task itself, autonomous systems monitor the surrounding OR workflow. This includes events like patient entering room, anesthesia start, first incision, closing incision, patient exit, room cleanup, etc. Some of these are visible (patient on table), some are auditory (anesthesia says “patient asleep”), and some might be inferred (no activity for a while might mean a delay). AI models can be configured to recognize these operational events. For example, Apella’s system classifies events from video such as “Patient Draped” as a distinct timestamp, which helped one hospital split surgical processes into more precise phases than their EHR had captured (healthcareitnews.com). By logging every step of the perioperative process, the AI builds a timeline that can be analyzed for inefficiencies.
Anomaly and Deviation Detection: An autonomous platform not only tracks normal procedures, but also watches for deviations from the norm or from protocol. This could mean flagging a potential error (e.g. an instrument not accounted for, a count discrepancy, a break in sterility if someone contaminates the field) or simply noting a variance (a step took 2× longer than average, or the surgeon skipped a usual step). Audio analysis is valuable here: the AI could catch if a required verbal confirmation (like the surgical timeout) didn’t occur, or if team communication indicates confusion. In real time, the system might alert the team to certain deviations (for instance, an alert on a dashboard if an unexpected instrument is opened or if the procedure is running behind schedule). Many current systems focus on post hoc analysis rather than live alerts, but the trend is toward more immediate, AI-driven feedback loops in the OR.

Multimodal Learning: Fusing Video and Audio for Context: The inclusion of audio in “using audio and video along with AI” is notable – it acknowledges that OR intelligence is not just about what an AI can see, but also what it can hear. Research in multimodal learning combines visual and auditory streams to better understand activities. In an operating room, certain critical information is only available via sound: machine alarms (indicating patient vital issues), spoken commands (“clamp!”), or even the tone of voices (which can indicate urgency or trouble). By training AI on synchronized video and audio, developers enable detection of events that a single modality might miss. For example, the AI might learn that a particular beeping pattern from a monitor (audio) combined with surgeons stepping back (video) indicates the patient went into an arrhythmia – a complex event that neither modality alone fully captures. Likewise, a sudden silence or change in conversation might precede a critical step, providing context to the visual data. Some cutting-edge AI models, often called “ambient intelligence” in healthcare, take in audio, video, and other sensor inputs (like RFID tags or infrared motion) to create a holistic picture. An example is the OR Black Box system, which records audio alongside video and sensor data to analyze human factors; it has been used to study communication errors and compliance lapses during surgery. While much of multimodal OR AI is still in R&D, the concept is clearly aligned with this SAGES insight – that by leveraging all available data (visual and auditory), an AI can autonomously understand the surgical process end-to-end. The use of large-scale vision-language models (like those that describe images in natural language) hints at how AI might even explain in words what’s happening in the OR video, effectively generating a narrative of the procedure that includes spoken context. This convergence of modalities is what will make AI agents truly powerful assistants in the OR rather than just simple counters of time or tools.

Current Platforms Demonstrating Autonomous OR Analysis: Several pioneering platforms are already implementing aspects of this autonomous tracking and analysis in live OR settings:

VitVio: A recently launched platform (highlighted by Mullings) explicitly embodies this vision. VitVio installs cameras and IoT sensors in operating rooms and uses AI to “track and analyze every aspect of the surgical procedure – from tool usage and surgical stages to adherence to safety protocols – in real time.” (businesswire.com) In a pilot at the Royal Orthopaedic Hospital (UK), VitVio provided a dashboard to OR staff with live surgery progress updates, estimates of remaining time, and alerts for deviations from standard protocols (businesswire.com). It essentially functions as an AI co-pilot for the OR, giving coordinators and administrators a constant stream of information to make decisions (e.g. start prepping the next patient because this case is almost done, or call for assistance if an unexpected delay occurs). After surgery, VitVio produces a detailed report highlighting successes and areas for improvement, fueling post-op debriefs and continuous learning for the team (businesswire.com). This matches exactly what the SAGES insight describes – autonomous tracking through audio/video to optimize efficiency – and shows that the technology is no longer theoretical. VitVio’s founders emphasize that it’s non-intrusive and doesn’t add tasks to the surgical team (businesswire.com); all data capture is ambient. Such systems are likely to become more common as hospitals seek to streamline operations.
Apella: Apella is another platform deployed in U.S. hospitals that uses an array of wall-mounted 360° cameras and “ambient” sensors to monitor OR workflow. It integrates computer vision with the hospital’s existing systems to automatically log key events and timestamps. In practice, Apella’s AI “passively and objectively” records events as they happen, which proved “much more reliable” than staff manually entering data in the EHR (healthcareitnews.com). In Houston Methodist’s pilot, Apella was able to identify previously untracked events (like the “Patient Draped” milestone between anesthesia and incision) to refine the surgical timeline (healthcareitnews.com). The system then provided real-time visibility: staff could see on a monitor exactly which phase a case was in and adjust scheduling accordingly (healthcareitnews.com). The results were tangible – improved coordination and a significant rise in utilization, as mentioned earlier (15% more capacity) (healthcareitnews.com). Apella shows that autonomous monitoring isn’t just about the surgical steps themselves, but the entire orchestration of the OR day. It’s an existence proof that AI can sit quietly in the corner of the OR, watching and learning, and deliver value by optimizing workflow without human prompting.
Caresyntax: While Caresyntax started as a broader surgical data integration platform, it has moved into real-time analytics through its Digital OR product suite. It can ingest live video, device feeds, and sensor data in a vendor-neutral manner. Caresyntax’s platform has been used to stream procedures in real-time for education (facs.org) and to analyze operational metrics (like delays and turnover times) across thousands of ORs. Although much of its analysis has been retrospective, Caresyntax recently introduced real-time elements – for instance, an app for intraoperative use that shows ongoing metrics. Caresyntax has also embraced AI for risk management (identifying potential issues during surgery) and announced collaborations to integrate more AI into its workflow solutions (caresyntax.com) (caresyntax.com). As Matt Krueger (the president of Caresyntax) highlighted, the battleground is the OR integration market – meaning whoever can best embed into the OR’s hardware and capture its data will have the advantage (linkedin.com). Caresyntax’s large installed base (over 2,800 ORs by 2023 (facs.org)) gives it a strong foothold to deploy autonomous tracking widely by software upgrades, illustrating how incumbents are adapting.
Touch Surgery (Medtronic): Medtronic’s Touch Surgery Enterprise initially focused on recording and analyzing surgical videos (especially from its own surgical robots and scopes) post-operatively. However, it is moving toward more real-time and integrated use. In 2024, Medtronic launched Touch Surgery Live Stream with new AI-driven “Performance Insights” algorithms (surgicalroboticstechnology.com). While the live stream aspect is for remote observation, the AI Performance Insights include 14 algorithms that automatically detect surgical workflow events, instruments, and anatomical landmarks during procedures (surgicalroboticstechnology.com). These algorithms work behind the scenes as the surgery is streamed or recorded, and they enable very detailed post-case analytics without manual review. For example, the system could tell a surgeon how long they spent dissecting tissue vs. suctioning, or how many times they used the cautery tool, and even identify anatomical structures touched during the case. This level of granular analysis is a form of autonomous understanding of the procedure – and in time, one can imagine it providing live feedback (“the lymph node count in view is X”) during a case. Medtronic’s push here underscores that major industry players are embedding autonomous video analysis features right into the surgical workflow (especially to complement their surgical devices). The strategic goal is similar: improve outcomes and efficiency by knowing everything that happens in the OR.
Proximie: Proximie is best known for remote surgery collaboration (letting surgeons virtually join another OR). Its platform records procedures that can later be reviewed or edited. Recently, Proximie has been positioning itself as a repository of surgical content that can be analyzed for insights. While Proximie itself hasn’t publicly detailed AI that autonomously tracks tools or steps, it partners in the ecosystem where such analysis could be applied to its recorded videos. Proximie’s value for autonomy lies in its potential to network multiple ORs globally – if autonomous AI agents were watching all surgeries on the Proximie network, they could derive large-scale insights (like global benchmarks for procedure times or best techniques). In essence, Proximie could serve as the connective tissue, while specialized AI modules (like those from Theator or others) perform the analysis. This modular, open approach reflects Mullings’ note that “open platforms are gaining traction” in digital surgery (linkedin.com), meaning the future may see different systems interoperate (one for capture, one for AI, one for data storage) to achieve autonomous OR monitoring.

Efficiency Gains and Workflow Improvements: The ultimate promise of autonomous AI tracking is a step-change in OR efficiency and consistency. By increasing the number of procedures that can be done (throughput) and reducing wasted time or supplies, hospitals can tackle backlogs and lower costs. The SAGES insight explicitly ties AI-driven tracking to “increasing # of procedures and optimizing OR efficiencies.” This has been borne out in early deployments:

At Houston Methodist, after adopting an AI vision system, the hospital saw more on-time starts and the ability to proactively adjust schedules, resulting in more surgeries completed per day. A 10% increase in monthly case volume was observed in the pilot ORs (healthcareitnews.com), and system-wide a 15% capacity gain was realized (healthcareitnews.com). Financially, even a few extra cases per OR per month can translate to significant revenue (hundreds of thousands of dollars per OR annually), not to mention treating more patients.
VitVio’s platform aims to address huge surgical waitlists in the NHS by wringing out inefficiencies. By tracking things like instrument preparation and protocol adherence, it helps ensure that no time is lost due to avoidable human delays. The BusinessWire report notes that eliminating unplanned OR costs (often due to delays or misused resources) could save on the order of $1.4 million per OR per year in the US (businesswire.com). Automating processes via AI is one of the few ways to capture these savings without simply demanding staff “work harder,” since it systematically identifies where improvements can be made.
Another efficiency aspect is reducing documentation burden on clinicians. If the AI automatically documents the case, the surgical team can finish up and move to the next case faster, rather than spending 10-15 minutes charting. Over a full day of back-to-back cases, those minutes add up to potentially fitting in an extra procedure. This is part of the “optimized processes” that AI tracking can provide – essentially automation of administrative tasks in the OR.
Moreover, by analyzing tool usage, hospitals can streamline their supply and instrument workflows. Knowing exactly which instruments were used (and which never opened) in each case can reduce unnecessary instrument kits or allow OR turnover staff to pick instruments more efficiently for the next case. VitVio specifically markets “cutting costs from equipment waste through AI-powered surgical tool tracking.” (vitvio.com). If each case runs with an optimal set of tools and minimal waste, turnover is faster and inventory costs drop, contributing further to efficiency.

In summary, autonomous audio/video analysis in the OR is demonstrating a virtuous cycle: better real-time information leads to better coordination, which leads to time saved, which allows more cases and better resource use. These efficiency and productivity gains are extremely attractive to hospitals and health systems under pressure to improve outcomes while lowering costs.

Implications for the Medtech Ecosystem: For technology providers, delivering autonomous OR monitoring is now a key differentiator. Companies that can offer an “AI co-pilot” for surgeries provide not just a product but a solution to hospital operational pain points. This has strategic implications:

Traditional device companies are partnering with digital startups to incorporate these capabilities (e.g., Medtronic’s partnership with Karl Storz for imaging combined with Touch Surgery AI, or J&J’s collaboration with Caresyntax for data integration). Everyone wants to avoid being left out of the data loop when AI agents become standard in every OR.
New business models are emerging, such as “AI-as-a-service” for the OR, where hospitals might pay subscription fees for analytics platforms (as they do for Caresyntax or Apella) and in return get efficiency gains that far exceed the cost. If an AI platform can help an OR do even one extra surgery a week, it more than justifies its expense.
There’s also a push for open data standards in surgical video and events, so that different systems (cameras, EMRs, AI algorithms) can talk to each other. Hospitals prefer not to be locked into a single vendor solution, which is why vendor-neutral platforms (like Proximie, Caresyntax, and others) stress integration. This openness accelerates adoption because the AI can tap into existing OR infrastructure (using whatever cameras or microphones are already there).
Ethically and regulatorily, autonomous OR tracking raises questions about data privacy and surgeon performance monitoring. Hospitals and companies must navigate how video data is stored (ensuring patient confidentiality by de-identifying video, for instance) and how it’s used (for quality improvement, not punitive measures against staff). So far, successful implementations have focused on the positive – using data to help staff, not to police them – which helps gain clinician buy-in (healthcareitnews.com). Finding that balance will be crucial for ecosystem players providing these solutions.

In essence, the medtech industry sees autonomous video/audio analysis not as a gimmick but as the next essential component of surgical systems. It delivers tangible operational value and generates big data that can further feed R&D, training, and product development. Little wonder that Mullings’ debrief called out these AI-powered OR platforms (VitVio, Apella, Proximie) as “tremendous value add” partners in the near-term future of surgery (linkedin.com). They extend the capabilities of surgical devices and imaging systems into the realm of intelligence and automation, which is where the future of surgery is headed.

SurgiCorder™ : Leveraging Video/Audio and AI in an AR Glasses Platform

SurgiCorder™ is a conceptual platform that embodies the above insights by using wearable technology and advanced AI to assist during surgery. It combines Meta’s Project Aria research glasses, NVIDIA’s edge AI computing, and vision-language AI models (LLaVA) to create a system that records surgeries from the surgeon’s perspective and provides real-time or after-action insights. In simpler terms, SurgiCorder™ is like a “smart surgical black box” worn by the surgeon – capturing everything they see and say – and an AI that can understand that multimedia data to help with documentation, guidance, and evaluation.

Design and Key Components: The hardware backbone of SurgiCorder™ is Meta’s Project Aria glasses, a lightweight wearable device loaded with sensors. These glasses are built for research and contain multiple outward-facing cameras, eye tracking, depth sensors, and microphones (reddit.com). Worn by the surgeon like a pair of spectacles, Aria glasses can record high-definition video from the surgeon’s point of view (POV) along with audio of the OR environment. The POV video is particularly valuable – it sees exactly what the surgeon is focusing on, which often includes the surgical field in open surgeries (where no laparoscope is present) or a heads-up view of laparoscopic monitors in minimally invasive cases. This perspective is ideal for reviewing surgical technique and context (it captures, for instance, where the surgeon’s attention was at each moment). The inclusion of multiple microphones allows capturing not just the surgeon’s voice but also other team communications and ambient OR sounds in a spatial manner.

On the processing side, SurgiCorder™ leverages NVIDIA AI computing, likely via an edge device such as the NVIDIA Jetson Orin platform. The Jetson Orin is a compact, powerful AI computer that can be connected to the Aria glasses to process the video/audio streams in real time. It provides GPU-accelerated computing on the edge (in the OR itself) so that heavy AI models can run without needing to stream data to a distant server (which could introduce latency or security concerns). In fact, NVIDIA has demonstrated that its Jetson edge devices can run sophisticated vision-language models with low latency: for example, a developer showcase in early 2024 ran the LLaVA model on live camera video via a Jetson Orin, achieving “interactive” speeds thanks to optimized 4-bit quantization of the model (forums.developer.nvidia.com). LLaVA stands for Large Language and Vision Assistant – it’s a multimodal model that extends a large language model (LLM) with the ability to interpret visual input. Essentially, LLaVA can look at an image (or video frame) and generate a human-like description or answer questions about it, combining visual understanding with language generation. By running a model like LLaVA, SurgiCorder’s system can not only capture video, but also analyze and talk about that video. This is a powerful capability: instead of just storing footage for later, the AI can be queried about what’s happening or can summarize events as they occur.

Surgical Documentation via AI: One of SurgiCorder’s core uses is automated surgical documentation. As the surgeon operates, the AR glasses’ video and audio feed into the AI, which can identify and log key moments. For example, the AI can note, “Surgeon made an incision at 10:05, using a #10 scalpel” by recognizing the action and instrument in the video. It could also transcribe spoken dictations – if the surgeon says “The appendix is inflamed” while looking at it, the system can record that observation. Using an AI like LLaVA, SurgiCorder™ can generate a textual summary of the procedure afterwards. Imagine an automatic operative report that reads: “Procedure Summary: The patient was brought into the OR and general anesthesia induced. The surgical team performed a laparoscopic appendectomy. Key steps: Initial inspection showed an inflamed appendix. The appendix was mobilized and a clamp applied to the base at 10:22. The appendix was divided and removed by 10:25. Hemostasis achieved and the specimen was passed off. No complications noted. The procedure concluded at 10:40 with closure of incisions.” Such a report can be drafted by the AI by analyzing the timeline of events and even recognizing anatomical structures (with the help of the surgeon’s gaze via eye-tracking data). The surgeon would then just verify and edit any nuances before finalizing it. This dramatically reduces the documentation workload. Given that documentation can take up a significant portion of a surgeon’s day (and poor OR record-keeping was noted to cause delays (healthcareitnews.com)), SurgiCorder’s automated logging is a big productivity win. It also ensures more thorough records – if an issue arises later (e.g. a legal question or a complication), the video-backed log can provide evidence of what was done or observed. In essence, SurgiCorder™ creates a real-time operative log, similar to what a diligent scribe might do, but powered by AI vision and language understanding.

Intraoperative Guidance and AR Feedback: Another envisioned feature of SurgiCorder™ is to provide guidance during the operation using the AI’s understanding of the video. Since Project Aria glasses are a research device without a built-in display, the guidance could be delivered through audio cues (e.g., a discrete voice in the surgeon’s earpiece) or via an external screen/tablet that displays insights. For example, as the surgery progresses, SurgiCorder™ might notice something like “a missing step” or simply be monitoring for safety. If the surgeon is about to close but the AI did not detect that a final sponge count occurred, it could gently remind: “Please verify sponge count before closing.” If the surgeon is in a training scenario, the AI could act like a tutor: “Next, you may want to isolate the cystic artery – it should be located here,” possibly even highlighting an area if the video is streamed to a tablet. For experienced surgeons, guidance could take the form of real-time metrics (“30 minutes have elapsed since incision, which is 5 minutes longer than usual at this stage”) helping them pace the procedure. The combination of AI and AR opens the door to contextual overlays – for instance, projecting anatomy labels or segmentation on the surgeon’s view. If integrated with pre-operative imaging, the system might even project where a tumor margin is relative to what the surgeon sees, though that requires advanced registration of images to real-world view. While these are forward-looking ideas, they are within reach: augmented reality research has already shown that head-mounted displays can overlay navigation cues in surgery, and AI can provide the content for those overlays. SurgiCorder’s approach is unique in that it tries to do this with a relatively lightweight, glasses-based system rather than a bulky headset, making it less obtrusive for surgeons.

Performance Evaluation and Post-Op Debrief: After the surgery, SurgiCorder™ turns into a powerful assessment tool. It can generate analytics about the procedure similar to what systems like C-SATS or Touch Surgery do, but tailored to the individual surgeon’s performance. Because it captures the surgeon’s perspective, it knows exactly where the surgeon was looking and what they were doing at each moment. The AI can compile metrics such as:

Total procedure time vs. baseline average (for that type of case).
Time spent on each major step (e.g., “dissection took 15 minutes, closure 5 minutes”).
Instruments used and how many times (e.g., “the cautery hook was activated 8 times, total cautery time X seconds”).
Any detected errors or drops in efficiency (perhaps the video shows the surgeon searching for a tool for 30 seconds – the AI could flag that as a delay).
Adherence to protocols (did the team perform the timeout? Did the surgeon follow all critical safety steps?).

Using an LLM-based model, SurgiCorder™ could even provide a narrative critique or coaching tips in natural language. For example: “Performance Feedback: The procedure was completed successfully. One observation: during the arterial clipping step, you spent longer than usual possibly due to difficulty in exposure – consider adjusting retraction technique to improve speed. Instrument usage was efficient, with minimal idle time. All safety checks were completed. Overall, compared to your past 10 cases, this case was slightly longer but had no major issues.” This kind of feedback, if accurate, is immensely valuable for continuous improvement. It’s like having a personalized coach who watched the entire case (which in fact, the AI did). Surgeons could review the annotated video, jump to moments the AI flagged (say, when a complication happened or an instrument was dropped), and reflect or discuss with peers. This overlaps with what tools like InfluenceOR™ offer (expert video review) but tries to automate the analytics portion with AI for scalability. For surgical training programs, such a platform could objectively chart a resident’s progress over time with concrete data.

Technical Feasibility and Challenges: SurgiCorder’s vision is ambitious, but the components are falling into place. On the feasibility side:

Hardware: Meta’s Project Aria glasses are real and have been distributed to researchers. A second generation (“Aria Gen 2”) was announced with even more capabilities and on-device processing (reddit.com). They are light and designed for comfort and hands-free data capture, which is perfect for the OR setting where surgeons cannot fuss with equipment. The glasses by themselves record data; they rely on a companion device (like a smartphone or small computer) for storage and processing. Here is where NVIDIA’s Jetson comes in, acting as the companion that can both store high-bandwidth video and run AI algorithms on it. Ensuring a secure, sterile setup is a consideration – the hardware must be draped or designed to meet OR sterile field requirements (likely the glasses stay on the surgeon’s head, which is usually outside the sterile field, so that’s manageable).
AI Models: Running an advanced model like LLaVA in real time is cutting-edge but has been demonstrated. NVIDIA’s optimization managed to get response times low enough for interactive use (forums.developer.nvidia.com). That said, analyzing a continuous video stream (rather than answering one-off questions about single images) is a heavier lift. SurgiCorder™ might not need to process every frame; it could sample frames or events. Also, models can be tailored – for instance, instead of a general LLaVA, a smaller specialized model could be trained to detect surgical instruments and phases, which would be faster. A likely approach is a hybrid: use fast CV models to detect specific known events/instruments, and use the LLM-based model for higher-level reasoning or summarization when needed. Technical hurdles include ensuring the AI’s computer vision accuracy in a sometimes chaotic OR scene (blood, occlusions, sudden camera motion when the surgeon moves their head, etc.). Extensive training on surgical data and perhaps using depth/IMU from the glasses to stabilize the view can mitigate this.
Integration: SurgiCorder™ must also interface with hospital IT for certain features – e.g., pulling case schedules (to know what procedure is planned, which can inform the AI’s expectations) or pushing documentation into the hospital’s record system. Building those integrations and ensuring compliance with health data security (HIPAA etc.) is a non-trivial but well-trodden path for digital health devices.
Current Capabilities: While SurgiCorder™ as described might be in prototype or concept phase, each of its individual functions has seen proof-of-concept. For example, on the NVIDIA forums a developer showed a “live LLaVA” system describing a camera feed (forums.developer.nvidia.com) – imagine that applied to a surgical field (“The surgeon is tying a suture”). Speech-to-text is a mature tech, so capturing spoken words is straightforward. The novelty is in combining these with domain-specific understanding. We can expect initial versions of SurgiCorder™ to focus on a subset of tasks, such as record & summarize. Real-time complex guidance will require more validation. However, even an early version that reliably generates an op note draft and timestamps of key events would be highly valuable to users.
Limitations: There are challenges: POV video can be jarring as the head moves – the AI will need to handle motion (perhaps using Aria’s inward eye-tracking to know where the surgeon is really focusing). Audio in OR can be noisy (beeps, multiple people talking) – isolating the surgeon’s voice might need multi-microphone array processing. Also, an AI model might misidentify things (imagine it labels fat as a tumor; such errors must be caught by human review). Therefore, SurgiCorder’s outputs would likely be suggestions or drafts, with the surgeon remaining the final decision-maker, especially for any clinical guidance. From a regulatory standpoint, using AI for live surgical guidance would classify as a medical device function that needs rigorous testing for safety. Documentation and analysis features might face less regulatory hurdle (they can be decision-support or post hoc analysis tools).

Market Potential and Differentiation: SurgiCorder™ sits at the intersection of several hot areas in medtech: augmented reality, AI, and digital surgery. Its potential market includes:

Surgeons and Surgical Educators: Individual surgeons who want to document and review their cases could use SurgiCorder™ as a personal tool (much like how some surgeons record their surgeries on GoPro or head-mounted cameras today, but with the benefit of instant AI analytics). Training programs could give these to residents to accelerate learning – every case becomes a teachable recording with AI feedback.
Hospitals/OR Managers: Hospitals could deploy SurgiCorder™ systems to surgeons in operating rooms that are not yet covered by bigger integrated solutions. It’s relatively low infrastructure – no need to outfit the whole OR with new equipment, just provide the glasses and a processing unit. This could be attractive for ambulatory surgical centers or smaller hospitals that can’t invest in full-scale OR integration platforms. By improving documentation and reducing errors, it could also lower malpractice risk and improve quality metrics, which hospital administrators care about.
Medtech Companies: For device makers, partnering or acquiring a solution like SurgiCorder™ could add value to their offerings. For instance, a company selling surgical implants could bundle an AI recording system to help surgeons using their implant to follow optimal steps (and gather usage data for the company). As Mullings and Krueger indicated, there is competitive pressure to have an ecosystem play – a company with SurgiCorder™ technology could license data or analytics to bigger firms or work with them.

The market potential is significant given the broad applicability – essentially any surgical procedure could benefit from better documentation and analysis. That said, SurgiCorder™ will also have to prove its value over existing solutions and navigate competition:

Some surgeons might question wearing glasses – comfort and not distracting the surgeon are paramount. Project Aria is designed to be lightweight and has no heads-up display, which is good for minimizing distraction. Still, surgeon adoption would require change management, similar to how surgeons had to get used to using scopes or robots at first.
From a business perspective, SurgiCorder™ could adopt a subscription or per-use model (as software) or hardware sales model. The latter might be challenging unless the hardware is unique; since Aria glasses are an off-the-shelf research product, the IP likely lies in the software/AI and workflow integration.

Comparing SurgiCorder™ to Other Digital OR Platforms: It’s useful to see how SurgiCorder’s approach contrasts with other prominent platforms in the digital OR space:

Platform	Data Capture Method	AI Focus & Capabilities	Key Use Cases / Features	Ecosystem Integration
SurgiCorder™	Surgeon POV via wearable AR glasses (Meta Aria) capturing video & audio	Multimodal AI (computer vision + language) for real-time understanding; LLM-based summarization of video; instrument and step recognition in surgeon’s field of view	Intraoperative: AI-driven guidance cues; hands-free video recording. Postoperative: Automatic op note generation, performance analytics, surgical skill feedback.	Stand-alone, surgeon-centric tool (vendor-neutral); potential to feed results into EMR or training databases. Leverages consumer/edge hardware (glasses, Jetson) for easy deployment.
Caresyntax	OR-installed cameras (overhead or endoscopic) plus integration with medical devices and EMR data	Workflow analytics, risk prediction (using ML on timestamps and sensor data); dashboarding of OR metrics; (Newer features include video-based skill review via InfluenceOR)	Operational: OR scheduling optimization, OR utilization tracking, safety check compliance. Analytics: Enterprise-level insights across many surgeries (e.g. which factors drive complications).	Enterprise platform deployed in thousands of ORs; vendor-neutral, pulling from various devices. Integrates deeply with hospital IT (EMR, PACS). Often sold to hospital administrators/perioperative management.
Touch Surgery (Medtronic)	Direct surgical video feeds (endoscopic, laparoscopic, or robotic camera output) recorded via a dedicated system; cloud-based video storage	Computer vision algorithms detect surgical phases, instruments, anatomy in recorded video; AI “Performance Insights” post-case; (real-time analysis emerging for guided live streaming)	Surgeon-focused: Video replay and sharing, automated surgical skill metrics (time on task, tool usage), educational content from recorded cases. Teleeducation: Live streaming of surgery with AR annotations for remote proctoring.	Integrated with Medtronic’s devices (e.g. Hugo RAS, laparoscopic towers), but can work with other video sources. Cloud ecosystem (Touch Surgery Cloud) for storing and analyzing cases; ties into Medtronic’s larger digital surgery strategy.
Proximie	Live video streaming from existing OR cameras or scopes (typically an external camera or the laparoscopic feed) plus two-way audio; cloud recording of sessions	Augmented reality overlays (surgeon can draw on screen); recordings can be later analyzed (currently more manual, though AI partnerships likely)	Collaboration: Remote surgical guidance, telementoring (experts virtually “scrub in” to assist). Content management: Library of recorded surgeries for review or training. Some ability to edit and annotate videos after capture.	Vendor-neutral and network-based – connects any OR with internet to any remote expert. Integrates with hospital scheduling and PACS to pull in imaging for discussion. Data from sessions could be used by third-party AI tools (open ecosystem approach).
Apella	Multiple panoramic cameras in OR and other ambient sensors; taps into OR audio and possibly device signals; data processed locally and sent to secure cloud	Computer vision for event detection (patient in room, incision made, etc.); AI/ML for OR workflow forecasting (when will case end, turnover status); essentially an “ambient intelligence” system for OR ops	Operational: Real-time OR status tracking, automated notifications (e.g. text pages when room is ready), identifying delays in workflow. Quality: Ensuring accurate timestamps for all OR activities, improving billing and resource use. Reports on throughput, causes of delays, etc.	Deployed as facility infrastructure – needs installation of cameras and integration with hospital IT. Sold to large hospital systems aiming for efficiency gains. Works alongside EHR, not surgeon-specific but OR-wide.
VitVio	Fixed camera and IoT sensor setup per OR; integrates with hospital’s info systems; focuses on computer vision from fixed angles and possibly audio	Computer vision for surgical phase tracking and tool usage in real time; rule-based alerts for protocol deviations; AI-driven post-case reports highlighting inefficiencies or safety issues	Operational & Training: Real-time dashboard for OR coordination (remaining time, next step alerts), post-op debrief reports for surgical teams to learn and improve. Tool usage analytics to reduce waste.	Early-stage deployment in pilot hospitals; aims to be vendor-neutral and additive to any OR. Likely offered as a hospital platform service. Emphasizes autonomy and non-intrusiveness, complementing existing workflows.

Looking at this comparison, SurgiCorder’s distinctive angle is its wearable, surgeon-centric nature and its use of an LLM-based AI for understanding the surgical scene from the surgeon’s viewpoint. Unlike the other platforms which rely on fixed room cameras or internal scope video, SurgiCorder™ follows the surgeon’s eyes. This makes it especially useful for surgeries that aren’t fully captured by a single camera (e.g. open surgeries where the critical view is what the surgeon sees directly). It essentially personalizes the digital OR concept to the surgeon level, whereas platforms like Caresyntax or Apella operate at the room or system level, and Touch Surgery/Proximie operate at the procedure video level. SurgiCorder™ could potentially integrate with those systems – for example, a Proximie session could incorporate a SurgiCorder™ feed to show the remote expert exactly what the surgeon is seeing, or SurgiCorder’s data could feed into Caresyntax’s analytics for more granular data.

Strategic Position: If SurgiCorder™ succeeds, it could carve a niche alongside these players by offering a lightweight, portable solution. This might be attractive to resource-constrained settings or as a complement in high-tech ORs (even in a fully wired “digital OR”, a visiting surgeon could wear SurgiCorder™ glasses to ensure their technique is recorded and analyzed, independent of the room’s equipment). For large medtech companies, acquiring or partnering with SurgiCorder™ could fill gaps: e.g., an endoscope company might use it to gather data on open surgeries that their products don’t currently cover, or a robotics company might use the glasses to record the surgeons’ interactions with consoles and the OR staff for broader context.

Of course, SurgiCorder™ also faces competition: if a hospital already has Caresyntax or Touch Surgery installed, they might rely on those systems’ growing capabilities rather than adding surgeon glasses. SurgiCorder™ will need to demonstrate that the first-person perspective plus LLM intelligence yields unique benefits – such as even more accurate documentation or novel insights on technique. Its success may also hinge on user experience (surgeons won’t adopt if it’s cumbersome) and AI accuracy (trust is key; if the AI routinely mis-identifies surgical steps, clinicians will ignore it). Early adopter testimonials and published results (like “using SurgiCorder™ led to X% reduction in documentation time, or improved a certain performance metric by Y”) will help validate the platform in the eyes of the medical community.

Conclusion

The insights from Joe Mullings at SAGES 2025 highlight a paradigm shift in surgery: data – especially video data – is becoming as valuable as the surgical tools in our hands. High-fidelity video and audio from the OR, when paired with advanced AI, can unlock improvements in efficiency, safety, training, and documentation that were previously unattainable. Mullings’ assertion that video is the “highest value of data” for surgical AI (linkedin.com) is borne out by the flurry of activity in digital surgery – from AI algorithms that dissect surgical videos frame by frame, to entire platforms built around capturing and analyzing every second of an operation. The ability to “autonomously track & analyze” what happens in the OR (linkedin.com) transforms surgical care from an art that happens behind closed doors into a process that can be measured, learned from, and continually optimized.

In this landscape, the SurgiCorder™ concept represents a cutting-edge application of those insights. By leveraging wearable AR glasses and on-the-fly AI analysis (via NVIDIA and LLaVA), SurgiCorder™ aims to put the power of video data directly in the surgeon’s workflow – essentially bringing an AI assistant into the OR without disrupting the flow of surgery. Our analysis shows that such a system is technically on the horizon and could synergize with other digital OR technologies. It stands to augment both the surgeon’s experience (through real-time support and easier documentation) and the hospital’s objectives (through data that improves operations and outcomes).

As digital and physical technologies converge in the operating room, we can expect an ecosystem where surgical video is the central substrate on which AI-driven applications run. Platforms like Caresyntax, Touch Surgery, Proximie, Apella, VitVio – and new entrants like SurgiCorder™ – will likely co-evolve, sometimes competing and sometimes integrating, to deliver the vision of an AI-powered OR. In that future, every surgery could benefit from an extra set of “AI eyes and ears,” leading to safer procedures, more efficient use of OR time, and better training for surgical teams. What’s clear from the SAGES 2025 debrief is that this future is rapidly approaching, and stakeholders across the medtech spectrum are gearing up for it. Video-driven AI in the OR is no longer a speculative idea – it’s the next great frontier of surgical innovation, and those who harness it effectively will shape the next era of healthcare.

Sources:

Joe Mullings, SAGES 2025 Debrief – Digital Surgery Insights (LinkedIn post, Mar 2025) linkedin.com
VitVio Press Release – AI Platform Tracks Tool Usage & OR Efficiency in Real Time (BusinessWire, Jan 2025) businesswire.com
Bill Siwicki, AI Computer Vision Enables Big OR Efficiency Gains (Healthcare IT News, Feb 2025) healthcareitnews.com
Medtronic Digital Surgery – Touch Surgery Performance Insights Launch (SurgicalRoboticsTech, Apr 2024) surgicalroboticstechnology.com
American College of Surgeons Bulletin – Video-Based Technologies in the OR (Nov 2023) facs.org
Matt Krueger (Caresyntax), comment on SAGES 2025 debrief (LinkedIn, Mar 2025) linkedin.com
NVIDIA Developer Forum – Live LLaVA on Jetson Orin (Edge AI Demo) (Feb 2024) forums.developer.nvidia.com
Apella – Ambient Intelligent OR system (Houston Methodist case study) healthcareitnews.com
OR Black Box by Surgical Safety Technologies – Teodor Grantcharov (Product Info via ACS) facs.org
Caresyntax – InfluenceOR and Digital OR platform (ACS Bulletin, 2023) facs.org