AI/ML Articles & Videos

Why digital health companies keep dying before and after FDA authorization, the evidence-reimbursement trap strangling the sector, and a post-market-first framework for breaking the cycle.

AI-assisted medical device development is FDA-compliant when done right. The obstacle isn’t regulation. It’s the messy codebase you’ve been ignoring.

Comprehensive analysis of 2025 cardiology AI/ML clearances, market trends, capital events, and regulatory insights.

44 clearances. 18 in 2025 alone. The dental AI market has reached its tipping point. This report dissects the Pearl/Overjet duopoly controlling 34% of the market, the 172-day FDA timeline you need to beat, and the $2.5M validation cost barrier. Discover how the "predicate network" creates shortcuts for fast followers, why clinical validation is the biggest bottleneck, and why the window to establish a foundational regulatory moat is rapidly closing for new entrants.

A data-driven look at 2025’s AI/ML 510(k) clearances: volume, review times, SaMD share, leading specialties, top manufacturers, and early PCCP adoption.

Deep dive into FDA 510(k) K252366, a2z‑Unified‑Triage, a radiology triage and notification device. Explains why the triage indication is likely the best way to get foundation models FDA cleared and . A practical regulatory pattern: constrained shell, generalist core, workflow value, diagnosis out of scope for teams shipping safely.

Breakthrough Device Designation (BDD) is a free, fast FDA program that signals serious unmet need and potential meaningful improvement. For AI/ML SaMD teams, it de-risks pivotal study strategy through earlier, higher-touch FDA engagement (including TAP), faster decision cycles, and clearer evidence expectations. This article explains what BDD is, why teams fail to get it, and answers common questions so companies can decide when and how to apply.

Setting acceptance criteria incorrectly (too high or too low) can delay your FDA submission by weeks or trigger rejection. This guide analyzed 784 AI/ML medical device clearances to provide concrete, data-driven thresholds. You'll find specific starting points: AUC of 0.9, Sensitivity/Specificity of 0.8, Dice of 0.8, Kappa of 0.85, and Bland-Altman LOA within ±10% of mean physiological values. Learn which statistical methods FDA accepts most often, how to justify your criteria using special controls and predicates, and what to do when your device fails testing.

This article shows how to achieve an AI/ML SaMD 510(k) submission in ~3 months by running a compact, well-powered MRMC study while continuously accruing standalone evidence through washout. It clarifies roles, sizing, units, and endpoints, with device examples, flowcharts, and a checklist to prevent rework and delays.

Here we analyze over 200 FDA 510(k) and De Novo summaries to tease out best practices for AI/ML study designs and adjudication strategies specifically.

A practical guide for engineers new to medical imaging AI, highlighting common pitfalls—like coordinate-system errors and orientation mix-ups—and providing simple checks to avoid them.Ask ChatGPT

Ideas on how to get foundation models and generative AI FDA cleared.

Recent advancements in radiology have led to the development of foundation models integrating vision and language capabilities, transforming medical imaging analysis

The U.S. Senate HELP Committee confirmed what we've been publishing for three years. An analysis of the regulatory signal, what it means for AI device companies, and why the companies that move first get to write the rules.

Yujans commentary on the new Non-invasive blood pressure measuring device FDA guidance document in response to WHOOP and others.

A deep dive into K232613 (CT Cardiomegaly)

6 FDA clearances in 10 days. The AI pace is accelerating. Coverage includes neuro, dental, diabetes, and ADHD. Key wins: Nobel Biocare’s implant SDK and Lumosity’s digital therapeutic.

Fifteen AI/ML software 510(k) clearances landed in nine days in mid-December 2025. Here’s what got cleared, where it fits clinically, and what it signals for regulated AI momentum.

Analysis of K252086 (DTX Studio Assist), a strategic 510(k) clearance for a 'headless' SDK. This submission minimizes regulatory burden by offloading UI requirements, employs a dual-predicate strategy to combine CADe with new segmentation and bone level measurement features, and efficiently reuses prior MRMC data (8.7% AUC increase). It establishes a strong template for clearing AI as an enabling technology component.

A client called out the “consultant thing”: hiding behind “ideally” instead of recommending a real plan. This piece explains why ideal talk wastes time, how to anchor on constraints, and how to keep defensibility high while choosing a speed vs spend setpoint for AI medical device work without losing time.

This article explains Bland–Altman analysis for evaluating agreement between two continuous measurement methods, especially in FDA-facing validation. It shows how bias and 95% limits of agreement reveal whether methods are interchangeable, why correlation and AUC can mislead, how to spot proportional bias, and includes practical interpretation tips, examples, and reusable Python code.

This article dissects a real AI/ML SaMD 510(k) submission (Body Check CT Cardiomegaly, K232613) with downloadable examples. After wasting a year over-engineering my first submission, I’ll share practical insights to prevent this from happening to you!

Command Line as a Medical Device (CLaMD) represents a streamlined approach to FDA clearance for medical software. By utilizing command-line interfaces instead of graphical user interfaces, developers can significantly simplify the regulatory process. This approach reduces UI hazard analysis requirements and minimizes human factors testing since interaction occurs through established systems like PACS or EHR.
CLaMD is particularly effective for AI algorithms, which can run automatically, output standardized formats, and operate server-side. This allows clinicians to continue using their familiar workstations while the AI works invisibly in the background. Numerous FDA-cleared devices have successfully implemented this pattern, demonstrating its viability in medical device submissions.

Kris Huang led a discussion on key image processing techniques in medical imaging, focusing on CT and MRI scans. Topics included histogram equalization, morphological operations, and perceptually uniform color maps. The conversation covered their role in visualization, segmentation, usability, accessibility, and color perception in radiology.

The document details the process and significance of Clinical Performance Assessments (CPAs) for medical devices, focusing on their role in regulatory approval. It highlights the complexities of study design, including methods to reduce variability and establish reliable ground truths. Statistical tools like FROC curves are emphasized for evaluating device impact on reader performance. Collaboration with regulatory bodies is stressed as essential for aligning study protocols with approval criteria.

Discussion covered key takeaways from the RSNA conference, balancing safety with speed, and setting risk thresholds. Topics included statistical power, U.S. data needs, AI/ML sample sizing, and using the PCCP tool to de-risk regulatory pathways. Collaboration between regulatory and engineering teams was emphasized for managing software changes.

This is a recording of the talk Yujan gave at the RSNA AI Theater

The discussion explores generative AI (GenAI) in medical devices, focusing on FDA's considerations for regulation, risks, and the lifecycle management of such technologies. Key points include the complexity of training data in foundation models, the risks of hallucinations in outputs, and the evolving role of post-market surveillance. Insights highlight challenges in balancing innovation with regulatory oversight and safety.

Dr. Elena Sizikova's presentation highlights the role of synthetic data in addressing data scarcity and privacy challenges in medical imaging AI. Synthetic data offers benefits like customizable variability, reduced risks, and improved bias control, particularly in fields like histopathology and pediatric imaging. She contrasts generative and knowledge-based models, emphasizing the potential of hybrid datasets for enhancing AI performance. The talk also stresses the importance of robust evaluation frameworks to validate synthetic data's effectiveness.

Yujan and David discussed the topic of FDA's performance in regulating AI medical devices, focusing on the ideal regulatory situation, the impact of AI on risk analysis, and the regulatory pathways for adding AI tools to existing software medical devices. They also explored the potential of continuous learning models in medical devices, the adoption of AI and ML in the US healthcare market, and the cybersecurity threats facing medical devices. The conversation ended with a discussion on the regulatory implications of continuous learning algorithms and the potential for a letter to file for a SAMD cleared by FDA for surgical planning.

A 10x talk exploring the basics of prompt engineering, followed by an open-ended discussion.

The talk covers device changes and their regulatory implications using FDA guidance. It focuses on two key documents for deciding when a 510(k) submission is needed for hardware and software changes. A "letter to file" process is used for post-market evaluations, while the guidance also helps in pre-market strategies for anticipating design changes that may require a 510(k).

The article outlines a method for documenting AI/ML algorithms for FDA pre-submissions. It emphasizes clear algorithm descriptions to avoid delays. Key steps include visual runtime descriptions, dataset analysis, non-ML testing, annotation details, AI/ML test plans, and performance metrics. It stresses reducing annotation costs and leveraging existing data while advising comprehensive documentation for effective FDA evaluation.

With predetermined change control plans (PCCPs), FDA has given manufacturers a great new regulatory tool. In this article, we discuss PCCP best practices and questions based on our experiences with FDA and our thorough research of the FDA databases.

Tips and FAQs on how to get a foundation model FDA 510(k) or De Novo cleared.

What is the difference between user needs and requirements?

Learn how to avoid invalidating your test dataset while still allowing yourself some flexibility to re-use the test data.

This video discusses the potential avenues for getting generative AI applications cleared through FDA.

This article outlines the process of developing an AI/ML algorithm from scratch and getting it FDA cleared. It covers the four phases of the process (Explore, Develop, Validate, and Document) and discusses the costs, time, and data requirements involved. It also provides advice on regulatory strategy, data annotation, and algorithm prototyping. If you're interested in developing a medical device involving AI/ML.

This article provides a questionnaire for those looking to get their AI algorithm FDA cleared. It covers topics such as validation studies, non-ML and ML software items, and acceptance criteria for the device. Reading this article can help uncover gaps that may need to be addressed before FDA submission.

This article outlines the challenges of bringing radiology AI/ML to market, from clinical workflow to IT installation. It provides insights and lessons learned for those interested in developing and deploying AI models in radiology.

FDA's Draft PCCP (Predetermined Change Control Plan) guidance should streamline model changes AI/ML-enabled devices, allowing for pre-approved modifications without the need for additional marketing submissions.

This webinar covers the four phases of the regulatory process: Explore, Develop, Validate, and Document. It also discusses the costs, time, and data requirements involved in the process. Additionally, it provides advice on regulatory strategy, data annotation, and algorithm prototyping.

This FDA guidance applies to AI/ML-enabled radiology software that assists in the detection of disease. Many of the principles and approaches, however, apply beyond this limited scope.

This guidance is describes FDA’s thoughts on evaluating the performance of CADe radiological devices. Designing standalone and clinical performance testing studies for AI/ML-enabled SaMD can be tricky. You want to do enough testing to ensure the device is safe and effective, yet you also need to avoid going bankrupt on the way!

This document outlines the technical performance assessment required for premarket submissions of quantitative imaging devices to the FDA. It provides guidance on the evaluation of the safety and effectiveness of these devices, including the methodology for quantitative imaging and the appropriate use of imaging biomarkers. Medical device manufacturers may be interested in reading this document to ensure their products meet the required standards and to understand the regulatory expectations for premarket submissions.

Introduces standardized terms and definitions to foster a consistent understanding across the industry for ML-enabled devices.

This article outlines Innolitics' process for developing machine learning algorithms for FDA-cleared medical devices. It covers documentation requirements, common pitfalls (overfitting, data leakage, augmentation errors), and a systematic workflow using four key reports: Input Verification, Data Augmentation QA, Model Performance, and Model Comparison—ensuring safe, effective, and compliant AI development.