What is RDC 657/2022 and when did it take effect?

Published March 24, 2022, effective March 1, 2023. Establishes comprehensive requirements for Software as Medical Device (SaMD) including AI/ML systems, cybersecurity, and clinical evaluation standards.

What types of software are covered by RDC 657/2022?

Covers software for medical purposes: diagnosis, prevention, monitoring, treatment, alleviation of disease. Includes standalone apps, AI/ML algorithms, cloud software, mobile health apps with medical claims.

How does RDC 657/2022 define Software as Medical Device (SaMD)?

SaMD is software intended for medical purposes that performs functions without being part of hardware medical device. Must have medical purpose: diagnosis, prevention, monitoring, treatment, or alleviation.

How is SaMD classified under RDC 657/2022?

Classification based on IMDRF framework: healthcare situation (critical/serious/non-serious) and healthcare decision (inform/drive/diagnose). Higher risk situations and autonomous decisions result in Class III-IV.

How are AI/ML medical devices specifically classified?

AI/ML devices typically Class III-IV due to complexity. Continuous learning algorithms require special controls. Classification considers transparency, explainability, and potential for bias.

What are the software validation requirements under RDC 657/2022?

Validation includes V&V activities, unit/integration/system testing, clinical evaluation. IEC 62304 compliance expected. Documentation must demonstrate safety, effectiveness, and performance throughout lifecycle.

What cybersecurity measures are required by RDC 657/2022?

Comprehensive cybersecurity including risk assessment, secure design, vulnerability management, incident response, regular updates. Security by design principles mandatory throughout product lifecycle.

What clinical evaluation is required for SaMD?

Clinical evaluation demonstrates safety and performance through clinical data, literature review, or clinical investigations. Higher risk SaMD typically requires prospective clinical studies.

What transparency requirements apply to AI/ML medical devices?

Algorithm transparency includes training data documentation, model architecture description, performance metrics, bias assessment, explainability features. Black-box algorithms require additional validation.

How are continuous learning algorithms regulated?

Continuous learning requires predetermined change protocol, performance boundaries, update controls, version management. Significant changes may require new submissions. Real-world monitoring mandatory.

What are the requirements for addressing bias and fairness in AI/ML devices?

Bias assessment mandatory including demographic analysis, performance across subgroups, mitigation strategies. Brazilian population representation required in training data or justification for differences.

What documentation is required for SaMD submission under RDC 657/2022?

Technical dossier includes software description, validation reports, cybersecurity documentation, clinical evaluation, IEC 62304 compliance evidence, AI/ML specific documentation if applicable.

What are the typical processing timelines for SaMD registration?

Class I-II SaMD: 30-60 days for notification. Class III-IV SaMD: 120-180 days for marketing authorization. AI/ML devices may require additional 30-60 days for technical review.

What are the post-market surveillance requirements for SaMD?

Continuous monitoring of performance, cybersecurity vigilance, adverse event reporting, periodic safety updates, version control management. AI/ML devices require real-world performance monitoring.

What transition provisions exist for products already on the market?

Products registered before March 1, 2023 had 6-month transition period. New requirements fully enforced September 1, 2023. Significant updates to existing products must comply immediately.

How does RDC 657/2022 align with international SaMD standards?

Aligns with IMDRF SaMD framework, FDA software guidance, EU MDR/IVDR requirements. IEC 62304 compliance recognized. Some Brazil-specific requirements for labeling and Portuguese documentation.

RDC 657/2022 FAQ - Updated October 2025

Brazil Brazil Software as Medical Device (SaMD) Regulation Guide. Expert guidance by Pure Global regulatory specialists

ANVISA SaMD Regulation

Effective March 1, 2023

AI/ML Medical Devices

Cybersecurity Requirements

Official Document

Access the official document:Download RDC 657/2022 (Portuguese)

Navigate Brazil's SaMD Regulatory Landscape with Pure Global

Your trusted partner for Software as Medical Device registration in Brazil and 30+ global markets

Pure Global offers comprehensive regulatory consulting solutions for SaMD and AI/ML medical devices, combining local Brazilian expertise with advanced AI tools to streamline your market access under RDC 657/2022.

SaMD Classification

Expert IMDRF framework application per RDC 657/2022

AI/ML Validation

Algorithm transparency and bias assessment

Cybersecurity

Complete security documentation and SBOM

Learn more about Pure Global's Brazil SaMD regulatory services →

Scope

SaMD

Software as Medical Device

AI/ML

Covered

Machine Learning Devices

Standards

IEC 62304

Software Lifecycle

Cybersecurity

Required

Security by Design

Frequently Asked Questions

On This Page

Overview & Scope

Classification & Risk Assessment

Technical Requirements

AI/ML Specific Requirements

Regulatory Pathway & Submission

Implementation Support

Quick Resources

All Brazil Regulations How to Choose a Brazil Registration Holder Free Brazil Medical Device Classifier Brazil Medical Device Database Brazil Market Insights & Analytics Classification Game Pure Global Brazil Services

Key Dates & Milestones

PublishedMar 24, 2022

EffectiveMar 1, 2023

DJ Fang

MedTech Regulatory Expert

Need help with 30+ markets registration?

Pricing

Free consultation

Note: This FAQ guide provides general information about regulatory requirements and should not be considered as legal advice. Regulations are subject to change. Always consult with qualified regulatory professionals for specific guidance.

Last updated: October 2025 | Based on RDC 657/2022 official documentation