Quick Answer
Use 510(k) when a device requires premarket notification and can be shown substantially equivalent to a legally marketed predicate. Use De Novo when the device is novel, has no suitable predicate, and can be classified as Class I or Class II with general or special controls. Use PMA when the device is Class III and requires valid scientific evidence showing reasonable assurance of safety and effectiveness.
Key Takeaways
- 510(k), De Novo, and PMA answer different FDA questions. They are not interchangeable filing formats.
- Predicate availability is the central issue for 510(k). Risk and control strategy are central for De Novo. Valid scientific evidence is central for PMA.
- eSTAR is mandatory for 510(k) submissions and De Novo requests unless exempted, but only voluntary for certain PMA application and supplement types.
- The wrong pathway can create months of delay through NSE decisions, declined De Novo requests, refused files, or major deficiencies.
- Pathway selection should happen before test strategy, clinical planning, labeling claims, and submission architecture are locked.
- FDA device pathway selection is one of the highest-leverage decisions in a medical device program. A device team can have strong engineering, good testing, and a clean submission package, but still lose time if the regulatory pathway is wrong.
- This guide compares the three main FDA premarket routes for medical devices: 510(k), De Novo, and PMA. It is designed as the umbrella pathway guide that connects to the detailed 510(k) submission, De Novo submission, and PMA submission articles.
- In this guide, you will learn:
- How FDA device class affects pathway choice
- When predicate strategy matters
- How evidence burden differs by pathway
- How eSTAR requirements differ
- How to avoid common pathway mistakes
FDA Device Classification Comes First
FDA classifies medical devices into Class I, Class II, and Class III based on the controls needed to provide reasonable assurance of safety and effectiveness.
| Device Class | Risk and Controls | Common Premarket Outcome |
|---|---|---|
| Class I | General controls usually sufficient | Often exempt, but some require 510(k) |
| Class II | General controls plus special controls | Often 510(k), sometimes De Novo |
| Class III | General and special controls insufficient | PMA unless a different lawful pathway applies |
The pathway decision usually starts with three questions:
- What is the device's intended use?
- Is there a legally marketed predicate with the same intended use?
- Are general controls or special controls sufficient, or is PMA-level evidence needed?
The answers determine whether the team should pursue 510(k), De Novo, PMA, or another route such as exemption, IDE, HDE, or 513(g) request.
510(k): Substantial Equivalence to a Predicate
A 510(k) submission is appropriate when the device requires premarket notification and can be shown substantially equivalent to a legally marketed predicate device.
FDA evaluates whether:
- The new device has the same intended use as the predicate
- The technological characteristics are the same, or differences do not raise different questions of safety and effectiveness
- Performance data support substantial equivalence
- Labeling, indications, and claims are aligned with the evidence
When 510(k) Fits
| Good 510(k) Fit | Poor 510(k) Fit |
|---|---|
| Valid predicate exists | No legally marketed predicate exists |
| Intended use matches predicate | Intended use is meaningfully different |
| Technology differences are well characterized | Technology raises different safety or effectiveness questions |
| Testing can support substantial equivalence | Independent safety and effectiveness evidence is needed |
The 510(k) pathway is not a shortcut around evidence. It is a predicate-based review. If the predicate comparison is weak, the submission can receive a not substantially equivalent decision or a significant Additional Information request.
De Novo: Novel Low-to-Moderate-Risk Devices
A De Novo submission is appropriate when the device is novel, lacks a legally marketed predicate, and can be classified into Class I or Class II with general controls or general plus special controls.
FDA evaluates whether:
- The device is eligible for De Novo
- The probable benefits and risks are adequately identified
- General controls or special controls can provide reasonable assurance of safety and effectiveness
- The evidence supports the proposed classification
- Proposed controls are sufficient for the device type
When De Novo Fits
| Good De Novo Fit | Poor De Novo Fit |
|---|---|
| No suitable predicate exists | Valid predicate exists and 510(k) is appropriate |
| Device risk is low to moderate | Device is high risk and requires PMA |
| Risks can be controlled with general or special controls | Controls are insufficient to mitigate risk |
| Evidence supports a new classification | Evidence is too limited to support benefit-risk decision |
A granted De Novo creates a new classification regulation and can establish a predicate for future 510(k) submissions. This can be strategically valuable, but it also means the first requester carries the burden of establishing the device type and controls.
PMA: High-Risk Class III Devices
A PMA submission is appropriate when the device is Class III and FDA must determine whether valid scientific evidence provides reasonable assurance of safety and effectiveness.
FDA evaluates:
- Clinical and nonclinical evidence
- Device design and performance
- Manufacturing information and controls
- Labeling and intended use
- Risk-benefit profile
- Quality system and inspection information where applicable
When PMA Fits
| Good PMA Fit | Poor PMA Fit |
|---|---|
| Device is Class III | Device is Class II with a valid predicate |
| Independent safety and effectiveness evidence is needed | Substantial equivalence can be shown |
| Clinical evidence is central to approval | Low-to-moderate-risk device can be controlled through De Novo |
| General and special controls are insufficient | Classification supports less burdensome pathway |
PMA is not only more documentation. It is a different evidentiary standard. The application must support FDA approval of the device for its intended use.
Side-by-Side Comparison
| Factor | 510(k) | De Novo | PMA |
|---|---|---|---|
| Main FDA question | Is the device substantially equivalent to a predicate? | Can this novel device be Class I or II? | Is there reasonable assurance of safety and effectiveness? |
| Predicate needed | Yes | No | No |
| Typical risk class | Class II, some Class I or Class III | Class I or II | Class III |
| Evidence burden | Predicate comparison plus performance data | Benefit-risk, risk controls, performance data | Valid scientific evidence, often clinical |
| Outcome | Substantially equivalent order | De Novo grant and new classification | PMA approval order |
| Future lifecycle | New 510(k)s for significant changes when required | Future devices may use De Novo as predicate | PMA supplements and reports |
| eSTAR status | Mandatory unless exempted | Mandatory unless exempted | Voluntary for certain PMA types |
eSTAR Requirements by Pathway
FDA's eSTAR program is now central to device submissions, but the requirements differ by pathway.
| Pathway | eSTAR Status |
|---|---|
| 510(k) | Required as of October 1, 2023 unless exempted |
| De Novo | Required as of October 1, 2025 unless exempted |
| PMA | Voluntary for certain PMA application and supplement types |
| IDE and Pre-Submissions | Voluntary for certain submission types |
This matters because eSTAR changes submission preparation mechanics. It creates structured prompts and required fields, but it does not decide the regulatory pathway or fix weak evidence.
For teams preparing 510(k) or De Novo submissions, eSTAR readiness is now part of filing readiness. For PMA teams, the first question is whether the specific PMA submission type and FDA center support voluntary eSTAR use.
Common Pathway Mistakes
Mistake 1: Forcing 510(k) Without a Predicate
If no valid predicate exists, trying to force a 510(k) can lead to a not substantially equivalent decision. A direct De Novo may be more appropriate when the device is novel and low-to-moderate risk.
Mistake 2: Choosing De Novo for a High-Risk Device
De Novo is not a way to avoid PMA for a high-risk device. If general and special controls cannot provide reasonable assurance of safety and effectiveness, PMA may be required.
Mistake 3: Treating PMA Like a Large 510(k)
PMA is not based on substantial equivalence. The submission must independently support safety and effectiveness for the intended use.
Mistake 4: Waiting Too Long to Align Testing With Pathway
Pathway selection affects bench testing, clinical evidence, software documentation, human factors, cybersecurity, biocompatibility, sterilization, and labeling. Delayed pathway decisions often create rework.
How Assyro Supports FDA Device Pathway Readiness
Assyro supports device teams across the submission readiness workflow: pathway logic, evidence organization, eSTAR completeness, cross-reference consistency, and readiness review before FDA filing.
Medical device regulatory submission software supports the broader workflow, while eSTAR Validation helps teams identify missing content, inconsistent claims, and package readiness issues for eSTAR-based filings.
The practical goal is simple: make sure the chosen pathway, device claims, evidence package, and submission structure all tell the same regulatory story.
510(k) is based on substantial equivalence to a predicate. De Novo is for novel low-to-moderate-risk devices without a predicate. PMA is for Class III devices requiring valid scientific evidence of safety and effectiveness.
References
This guide reflects FDA device pathway information current as of May 2026. Confirm current FDA guidance, classification information, product code guidance, and review division feedback before selecting a pathway.
About the author
Assyro Team
Expert regulatory operations consultants helping pharmaceutical companies navigate complex compliance challenges.
See Assyro in action
Catch eCTD and eSTAR errors before your FDA review cycle.
Book a 20-minute demo this week. We'll validate a sample of your submission live and show you exactly where Assyro catches what your current QC misses.