FDA's Robotic Surgical Device Draft Guidance: What Canadian Device Makers Should Say Before 24 November

On 25 September 2026, the United States Food and Drug Administration released a draft guidance on premarket submissions for robotically assisted surgical devices. It is a detailed, roughly 50-page document, and it tells device makers what a submission for one of these systems should now contain. For Canadian companies that sell, or plan to sell, a surgical robot into the United States, this is not background reading. It is the standard your next submission will be measured against.

The draft is open for public comment until 24 November 2026 under docket FDA-2026-N-9505. That matters. A draft guidance is the one moment when a manufacturer can shape the expectations it will later have to meet, and Canadian makers who build surgical robots or their instruments and accessories have a direct commercial interest in the final wording. This article explains what the guidance asks for, how it affects Canadian device firms, and the practical steps to take now, whether you plan to comment, prepare a submission, or both.

What the Draft Guidance Covers

What it is: a draft FDA guidance that describes the information a premarket submission for a robotically assisted surgical device should include. Why it matters: it signals how the FDA will review these submissions, so it shapes testing plans and timelines. What to do: map your evidence plan to the guidance now, and submit a comment before 24 November 2026 if any expectation would be difficult or disproportionate for your device.

The FDA defines a robotically assisted surgical device, or RASD, as a teleoperated, software-controlled system that integrates robotic technologies to help a qualified practitioner position and control surgical instruments. These systems support open, minimally invasive, and endoluminal procedures. The draft is careful about its edges. It does not cover pre-operative planning systems or stereotaxic systems, and it does not cover robots that perform significant parts of a procedure autonomously, although the FDA notes that some of its recommendations may still be useful for those devices. If your system includes automated functions, read the scope section closely, because where your device sits on the teleoperation-to-autonomy spectrum changes which recommendations apply.

The guidance is organized around the content of a premarket submission. It addresses the device description, the pathway and predicate comparison, accessories and compatible devices, nonclinical bench and reliability testing, clinical and in vivo evaluation, human factors, cybersecurity and software, electrical safety and electromagnetic compatibility, optical radiation safety, sterility, reprocessing, shelf life, packaging, pyrogenicity, biocompatibility, labelling, and training. It also discusses clinical endpoints for different procedure types and a pathway for an umbrella marketing authorization that can cover more complex procedures within the same category. That breadth is the point. A surgical robot is a system, not a single instrument, and the FDA expects the submission to treat it as one.

Why It Matters for Canadian Device Makers

Direct answer: a Canadian company does not escape this guidance by being based in Canada. Any manufacturer that files a premarket submission with the FDA for a RASD, wherever it is headquartered, will be reviewed against these expectations. And the same surgical robot is almost always a Class III or Class IV medical device in Canada, licensed under the Medical Devices Regulations. So most Canadian makers are building one evidence package for two demanding regulators at once.

That dual reality is an opportunity, not just a burden. Bench testing, reliability data, human factors validation, cybersecurity documentation, and biocompatibility files built to FDA expectations map closely to what Health Canada asks for in a Class III or IV licence application and to the ISO 13485 quality system that underpins both. A Canadian maker who plans the evidence once, with both jurisdictions in view, avoids running two disconnected programs and repeating expensive tests. A maker who treats the US submission as an afterthought usually discovers the gaps late, when a test has to be redesigned and a launch date slips.

There is also the comment itself. Draft guidances exist to be improved. If a proposed reliability expectation, a clinical endpoint, or a cybersecurity test would be disproportionate for your device, or ambiguous enough to invite inconsistent review, the comment period is the time to say so, with evidence. Comments are public and are read. A well-reasoned submission from a Canadian innovator carries weight, and it signals to the FDA that your company understands the technology at a system level.

What a Premarket Submission Must Now Contain

The guidance is specific about submission content. Four elements deserve early attention because they shape everything else.

Device description

The FDA wants a clear description of all major functional components, organized from a whole-system overview down to subassemblies and individual components. For a teleoperated robot that means the surgeon console, the patient-side manipulator, the instruments and end effectors, the vision and display system, and the software and control architecture, each described and then shown in how it connects to the rest. Sponsors may also submit videos of the device in use, especially of the user interface and any novel features. Plan to record clean, annotated footage early, not the week before you file.

Classification and substantial equivalence

For a 510(k), the FDA asks you to identify the classification regulation number and product code of your device, and to compare it to a predicate. Use words like identical or same only where features genuinely are, and explain every technological difference and why it does not raise new questions of safety or effectiveness. Where a recognized consensus standard applies, a Declaration of Conformity can streamline the review. Not every surgical robot has a clean predicate. A genuinely novel system may belong in the De Novo pathway, and the most complex or highest-risk systems may require a premarket approval application. Decide the pathway deliberately, because it drives the evidence you need.

Accessories and compatible devices

Instruments, energy sources, staplers, and other compatible devices must be described, and their interoperability with the robot demonstrated. If an accessory already holds a marketing authorization, include its submission number. This is where many robotic programs underestimate the work, because every instrument that connects to the system is part of the safety story, and compatibility has to be shown rather than asserted.

Supporting evidence and the umbrella pathway

The draft describes clinical endpoints tied to specific procedures and sets out an umbrella approach, where a sponsor can seek authorization for a category of procedures and then add more complex procedures within that category on a defined basis. For a maker expanding an existing system into new procedures, that pathway can be the difference between one coordinated program and a series of standalone submissions. Read it against your product roadmap, not just your current indication.

The Evidence the Guidance Expects, by Domain

The table below summarizes the main evidence domains the draft addresses and what the FDA expects in each. It is a planning tool, not a substitute for the guidance itself.

Testing Expectations: Bench, Reliability, Clinical and Human Factors

Direct answer: the guidance expects layered evidence. Bench testing proves the machine does what it claims, reliability testing proves it keeps doing it over its service life, clinical and in vivo data prove it works in a real operating environment, and human factors testing proves people can use it safely.

Bench performance should cover teleoperation accuracy and latency, instrument motion and position control, the behaviour of the system controls, and the quality and timing of the visualization and display. Reliability testing has to show the system holds its integrity across its stated service life and never loses control of an instrument, with acceptance criteria defined in advance and justified. Clinical and in vivo evaluation should demonstrate the device functions reliably during actual procedures, with endpoints chosen to match each target procedure. Human factors work is not a formality for a surgical robot: the FDA expects testing that surfaces use-related hazards for the whole operating-room team. If you have not mapped your human factors submission category and planned validation around real users, start there.

Cybersecurity and Software

A surgical robot is software before it is anything else, and the guidance treats it that way. Software verification and validation, including performance testing, is central. If the device meets the FDA definition of a cyber device, the submission should follow the FDA cybersecurity guidance: identify the assets, threats, and vulnerabilities, assess the risk, describe the mitigations, and run a cybersecurity assessment that includes penetration testing across every network element that communicates with the system. If your robot includes machine-learning-enabled functions, expect the FDA to look for a predetermined change control plan, and remember that Health Canada has its own pre-market expectations for machine-learning-enabled devices. Build the software, cybersecurity, and change-control documentation as one coherent file rather than three.

Training and Labelling

The FDA asks sponsors to include a training plan in the submission, and to show that the training is comprehensive and effective. Crucially, training must reach more than the surgeon. Bedside assistants, scrub nurses, and other operating-room staff interact with the system, and the guidance expects their training to be addressed. Labelling is listed as a covered topic, and it should be consistent with the intended users, the validated procedures, and the human factors findings. Treat the training plan as evidence to be designed and tested, not a manual written at the end.

The Canadian Angle: Class III and IV Licensing and MDSAP

Direct answer: in Canada, a robotically assisted surgical device is a high-risk device. It needs a Medical Device Licence from Health Canada, and the manufacturer needs ISO 13485 certification issued under the Medical Device Single Audit Program (MDSAP). The FDA evidence package is most of what a strong Canadian application needs, if it is planned that way.

Health Canada licenses devices by risk class, and a surgical robot that controls instruments inside a patient falls in the highest classes. A Class III or IV application draws on the same safety and effectiveness evidence the FDA now describes: design and manufacturing information, risk management, bench and clinical data, software and cybersecurity documentation, and biocompatibility. The FDA has moved its quality system requirement onto ISO 13485 through the Quality Management System Regulation, which brings US and Canadian quality expectations closer together. Biocompatibility is another shared file: the FDA's partial recognition of ISO 10993-1:2025 affects how you document patient-contacting instruments in both markets. Plan the evidence once, and the Canadian licence and the US submission reinforce each other.

Two practical reminders for exporters. First, selling into the United States also means FDA establishment registration and device listing, with annual user fees and a US agent. Second, if your pathway is a premarket approval application, the submission now runs through the PMA eSTAR electronic template, so build your evidence to fit that structure from the start.

Compliance Checklist

Use this checklist to gauge readiness for an FDA submission and a parallel Canadian licence application.

• Confirm your device meets the FDA definition of a RASD, and note where any automated functions sit relative to the scope.
• Choose your US pathway deliberately: 510(k) with a defensible predicate, De Novo, or PMA.
• Build a whole-system device description, from overview to subassemblies, with annotated videos of the user interface and novel features.
• Define bench and reliability acceptance criteria in advance for teleoperation, controls, kinematics, visualization and service-life integrity.
• Plan clinical and in vivo evaluation with endpoints matched to each target procedure, and consider the umbrella pathway for future procedures.
• Run human factors validation that includes bedside assistants and scrub nurses, not only surgeons.
• Prepare software verification and validation, a threat and vulnerability assessment, mitigations, and penetration testing across all connected network elements.
• Document accessories and compatible devices and demonstrate interoperability, with submission numbers where they are already authorized.
• Design a training plan for the whole operating-room team and evidence its effectiveness.
• Align the package with a Canadian Class III or IV Medical Device Licence application and MDSAP ISO 13485 certification.
• Decide whether to submit a public comment by 24 November 2026, and draft it with evidence.

Common Mistakes

• Treating the robot as one device rather than a system. Consoles, manipulators, instruments, vision and software each generate evidence, and a submission that describes only the headline unit looks incomplete.
• Leaving human factors to the end. Use-related hazards in surgery are high-stakes, and validation that excludes bedside staff misses real risks the FDA expects you to find.
• Underestimating cybersecurity. A connected surgical system needs a full threat model and penetration testing, not a one-line statement that the device is secure.
• Forcing a weak predicate. Calling features identical when they are only similar invites additional questions. A De Novo or PMA may be the faster route for a novel system.
• Running the US and Canadian programs separately. Duplicated testing and misaligned documentation waste money and time when one planned package could serve both.
• Skipping the comment period. If an expectation is disproportionate for your device, silence now becomes a fixed requirement later.

Frequently Asked Questions

What is a robotically assisted surgical device under the FDA draft guidance?

It is a teleoperated, software-controlled system that helps a qualified practitioner position and control surgical instruments during open, minimally invasive, or endoluminal procedures. The draft does not cover pre-operative planning systems, stereotaxic systems, or fully autonomous robots, although some of its recommendations may still apply to autonomous devices.

When is the comment deadline and what is the docket number?

Comments are due 24 November 2026 under docket FDA-2026-N-9505 on regulations.gov. The draft was published in the Federal Register on 25 September 2026 (91 FR 60967).

Does the guidance apply to Canadian manufacturers?

Yes. Any manufacturer filing a premarket submission with the FDA for a robotic surgical device, regardless of where it is based, will be reviewed against these expectations. Canadian makers also need a Health Canada Medical Device Licence for the same device, so a single evidence plan usually serves both markets.

Which US submission pathway applies to a surgical robot?

It depends on the device. A system with a suitable predicate may use a 510(k). A novel system with no predicate may fit the De Novo pathway. The highest-risk systems may require a premarket approval application. The draft guidance helps you identify the classification regulation number and product code and build the comparison.

How is a surgical robot classified in Canada?

A robotic surgical device is a high-risk device under the Medical Devices Regulations and typically falls in Class III or Class IV. It requires a Medical Device Licence and ISO 13485 certification issued under the Medical Device Single Audit Program.

What cybersecurity evidence does the FDA expect?

If the device is a cyber device, the submission should identify assets, threats, and vulnerabilities, assess risk, describe mitigations, and include a cybersecurity assessment with penetration testing across all network elements that communicate with the system, alongside software verification and validation.

Should we submit a comment?

If any proposed expectation would be disproportionate, ambiguous, or technically difficult for your device, yes. A clear, evidence-based comment filed before 24 November 2026 can shape the final guidance you will be held to.

How MFLRC Can Help

MF License and Regulatory Consultants works with device companies on exactly this kind of dual-market program. We run gap assessments that map your current evidence against both the FDA draft guidance and Health Canada Class III and IV expectations, so you test once and file twice. We build and strengthen the ISO 13485 quality system behind an MDSAP certificate, draft and review the SOPs that support design controls, risk management, software and cybersecurity, and prepare submission-ready documentation for a 510(k), De Novo, PMA, or a Canadian Medical Device Licence.

We also help teams decide whether and how to comment, and we support validation planning, human factors strategy, and mock FDA and Health Canada reviews so there are no surprises when the real submission lands. Every engagement is led by a senior consultant, not a checklist.

Building or exporting a robotic surgical device? Talk to MFLRC about a single evidence plan that satisfies the FDA and Health Canada together.

Conclusion

The FDA's draft guidance is the clearest signal yet of how robotic surgical devices will be reviewed in the United States. It asks for whole-system evidence: bench and reliability testing, clinical and in vivo data, human factors validation for the full operating-room team, serious cybersecurity, and training that is designed and proven rather than assumed. For Canadian makers, the same evidence largely builds the Class III or IV licence application at home. The smart move is to plan the package once, with both regulators in view, and to use the comment window before 24 November 2026 to shape the standard rather than inherit it.

Sources and References

• Federal Register: Robotically-Assisted Surgical Devices - Premarket Submissions; Draft Guidance (91 FR 60967, 25 September 2026)
• Regulations.gov: Docket FDA-2026-N-9505 (submit comments by 24 November 2026)
• FDA: Computer-Assisted Surgical Systems (robotically assisted surgical devices)
• Health Canada: Medical device licence application information
• ISO 13485:2016 Medical devices - Quality management systems