The New Era of AI-MDLC (AI Medical Device Life Cycle)

MedTech Leading Voice - Issue 166

📝 From the Editor

A few things I want to note:

1. Today is my wedding anniversary; help us celebrate.
2. Voting for MLV100 2026 has closed. More than 4,500 votes were cast; results will be announced soon.
3. I'll be in Boston this week for IMSC26. If you're there, come say hello.

— Sean Smith

🧠 Featured Post

From SDLC to AI-MDLC: Why AI Devices Don't Fit the Old QA Playbook

Kiran Kumar Prabu's point is that you can't run an AI device through a QA process built for software that stops changing. Conventional software gets written, tested against fixed boundaries, and locked. An AI model keeps learning from data and drifts after it ships.

His diagram tracks what that does to the lifecycle. Validation stops being "does the code execute" and becomes "does the model still perform on independent clinical data." Failure shows up as data drift and model degradation rather than crashes and memory leaks. The regulatory anchors move toward GMLP, QMSR, and a Predetermined Change Control Plan, so the model can update without a fresh submission every time. Worth reading if you're scoping an AI submission. View the post >>

📌 This Week's Top Posts (In no special order)

• The distinction between deterministic logic and true AI in MedTech devices and its regulatory impact, contrasted by Lavanya Ramnath.
• Saira Ali discussed priorities for modernizing Canada’s drug and medical device regulatory framework in her recent article.
• Kathryn B. shared FDA updates on partial recognition of ISO 10993-1:2025 and key implementation exceptions for medical device manufacturers.
• A regulatory checklist for medical device clinical investigation plans, detailed by Sumeet Tarke.
• Evangeline Loh underlined updates to Mexico’s medical device labeling standard and its alignment with EU MDR principles.
• Shweta Thakkar highlighted the growing importance of UDI terminology for traceability, compliance, and patient safety in medical devices.
• Simplified medical device labeling compliance in India was announced by Nutan Quality System Consultant.
• FDA’s latest SaMD guidance and evolving expectations for software medical devices were explored by Renuka Dixit.
• Joel Kent outlined FDA updates to recognized medical device standards and evolving regulatory expectations.
• Governed AI applications within CAPA workflows and the importance of deterministic validation and traceability in quality systems, broken down by Martin King.
• Tandem Health published findings from Sweden’s Medical Products Agency classifying AI medical scribes as Class IIa medical devices under EU MDR.
• Mario Gabrielli Cossellu shared key upcoming EUDAMED module requirements becoming mandatory for EU medical device market access in 2026.
• Stephanie Grassmann raised open questions on Implementing Regulation 2026/977 and its impact on MDR/IVDR re-certification, transitional timelines, and enforcement.

🗓️ Upcoming Events & Webinars

• June 1, 2026 | Boxborough, MA (and Virtual) — IEC 60601-1 4th Edition Seminar,an Overview and Q&A with Members of the Standards Team. Free in-person/virtual seminar hosted by Leo Eisner and TÜV Rheinland.
• June 2–5, 2026 | Boston — IMSC26 International MedTech Safety Conference. Use MLV codes to register: MLV$200 (general, in-person or virtual), Start$250 (startup in-person), StartEd$50 (startup education day).
• 🔥 NEW: June 5, 2026 | 11:00 am ET — Why MDR Technical Files Keep Breaking Down, featuring Nicolas Gehring and Leon Kobinger of CertHub.
• June 12, 2026 | 11:00 AM ET — In Silico Evidence and the Future of Clinical Evaluation, featuring Surash Surash, Clinical Reviewer at TÜV SÜD.

👀 In Case You Missed It

Last Thursday's May 28 webinar on From Observation to Action: What FDA Expects After a 483, featuring Michelle Lott and Tianna Benson from leanRAQA, drew more than 400 registrants. Key topics included:

• What an FDA Form 483 means and how it differs from a Warning Letter
• Why the 15-business-day response window matters
• How to restate the FDA observation clearly and identify what requirement was missed
• What containment means and how to document immediate risk control
• How to investigate root cause and determine whether the issue is isolated or systemic
• How to assess impact across products, lots, suppliers, CAPA trends, training records, and patient risk
• How to build a CAPA plan with evidence, due dates, ownership, and realistic commitments
• How to organize objective evidence for FDA follow-up and prepare for reinspection readiness

The replay is now available >>

💥 Weekly 510(k) Roundup

May 25 Issue - from Marcus Engineering, LLC

70 clearances this week, including 12 first-time clearances (🚀)

Featured first-time clearances:

• 🚀 WiserDoc (AISTETH) — adds AI to the stethoscope to detect heart and lung problems earlier.
• 🚀 Pulmera Inc. — software-based C-beam imaging that upgrades existing C-arm systems already in operating rooms.
• 🚀 Neko Health — combines imaging, sensors, and measurements in one visit for a full-body health check.

📋 Regulatory Roundup

May 25, 2026 | By Martin King

Regulators across the FDA, EMA, MHRA, and other authorities continue advancing lifecycle-based oversight, with stronger emphasis on AI classification, digital infrastructure like EUDAMED, and post-market control. This week's updates span EU AI Act high-risk guidance, HTA and Joint Clinical Assessment implementation, the FDA's TEMPO Digital Health Pilot and expanded real-world evidence, and accelerating SaMD and AI change-management frameworks across Brazil, Canada, China, South Africa, and Malaysia.

Explore the update >>

🔥 Postscript

Last Friday's In Focus session, What ChatGPT Doesn't Know About Sterilization, featuring Manuel Grandy, was a huge hit.

Watch the replay >>

Sean 🐝