AI Diagnostics Clear IVDR: What Two CE Marks Tell Us About SaMD

Two AI-driven diagnostics companies—Spotlight Medical and BrightHeart—secured IVDR CE marks within hours of each other this week. One addresses breast cancer staging in pathology labs. The other tackles prenatal ultrasound interpretation. Different clinical domains, different technical architectures, yet both crossed the same regulatory threshold under one of Europe's most demanding frameworks. For software as a medical device (SaMD) teams navigating IVDR, these aren't just corporate milestones. They're data points in a pattern that's clarifying what regulators expect when AI meets in vitro diagnostics—and what your clinical evidence strategy needs to deliver now.

What Spotlight Medical and BrightHeart Actually Cleared

Spotlight Medical's myStage Dx is an AI platform designed to assist pathologists in breast cancer staging. It analyses digitised pathology slides, integrating with existing laboratory workflows to support oncology centres and digital pathology environments. The IVDR CE mark positions it for commercial deployment across the EU, targeting a sector where diagnostic precision directly influences treatment pathways and patient outcomes.

BrightHeart's B-Right AI Platform focuses on prenatal ultrasound. It's built to assist clinicians in interpreting fetal imaging, a domain where operator variability and diagnostic confidence intersect with profound clinical consequences. The company is debuting the platform at the Fetal Medicine Foundation World Congress in Vienna, signalling intent to scale across European maternal-fetal medicine networks.

Both devices share a regulatory profile: AI-driven, software-only, intended to support clinical decision-making in high-stakes diagnostic environments. Both required IVDR compliance—meaning they navigated performance evaluation, clinical evidence generation, risk management, and quality management system alignment under Regulation (EU) 2017/746. And both are entering markets where regulatory expectations for SaMD continue to tighten, particularly around transparency, post-market surveillance, and algorithmic robustness.

The IVDR Pathway for AI Diagnostics: What Changed and Why It Matters

IVDR came into full effect in May 2022, replacing the In Vitro Diagnostic Directive (IVDD) with a framework that demands more rigorous clinical and performance evidence, especially for higher-risk devices. AI-based diagnostics often fall into Class C or even Class D categories, depending on their intended use and the consequences of incorrect results. That classification triggers requirements for notified body review, comprehensive technical documentation, and ongoing post-market clinical follow-up.

For manufacturers, this means your clinical evidence can't rest on algorithm validation alone. You need real-world performance data, demonstrable clinical benefit, and traceability from design inputs through post-market surveillance. Spotlight Medical and BrightHeart's approvals suggest both companies built clinical evaluation plans that aligned analytical performance (sensitivity, specificity, accuracy) with clinical utility in representative populations—and that they satisfied notified bodies on data quality, risk mitigation, and post-market commitments.

This is where many SaMD teams stumble. IVDR doesn't just ask 'does your algorithm work?' It asks 'does it improve clinical outcomes, and how will you prove that over time?' If your evidence generation plan treats post-market surveillance as an afterthought, or if your clinical evaluation report lacks depth on intended user context and real-world variability, you're not ready for notified body scrutiny.

What This Means for SaMD Teams Across Borders

The timing of these approvals is not coincidental. As the FDA and MHRA deepen their liaison on SaMD oversight, manufacturers are facing a reality where regulatory expectations are converging—not harmonising, but converging in substance. The clinical rigour demanded by IVDR increasingly mirrors what FDA expects under Software as a Medical Device guidance, especially for AI/ML-enabled devices. And the MHRA's own sandbox initiatives and risk-based frameworks are pulling in the same direction.

For manufacturers planning multi-market strategies, this means your EU submission isn't a siloed exercise. The clinical evidence you generate for IVDR—especially around algorithmic transparency, training data provenance, and performance across demographic subgroups—will inform your FDA premarket submission. Your post-market surveillance infrastructure will shape your vigilance obligations in the UK. And your quality management system, if built right, becomes the unifying architecture across all three regimes.

BrightHeart's parallel pursuit of FDA clearance is instructive here. The company isn't treating its CE mark as a finish line. It's treating it as a foundation for a global regulatory strategy where clinical evidence, risk management, and quality systems scale across jurisdictions. For startups and scale-ups, this is the model: build once, adapt intelligently, and avoid the trap of jurisdiction-specific firefighting.

The Post-Market Reality No One Talks About Enough

Securing a CE mark is not the end of the compliance journey—it's the beginning of a different one. IVDR Article 10 requires manufacturers to actively update their clinical evidence and performance evaluation throughout the device lifecycle. For AI-based IVDs, this is particularly demanding. Your algorithm's performance in controlled validation studies may not reflect real-world deployment, especially if clinical workflows, patient populations, or hardware environments differ from your test conditions.

This is where AI-driven post-market surveillance and adaptive QMS infrastructure separate reactive manufacturers from those scaling sustainably. If your post-market clinical follow-up plan relies on manual data aggregation or episodic review cycles, you're already behind. The expectation now is continuous, near-real-time monitoring of device performance, integrated with design feedback loops that inform algorithm updates, labelling refinements, and risk mitigation actions.

Spotlight Medical and BrightHeart will face this reality as they scale. Every deployment generates new data. Every clinical site introduces variability. And every algorithm update—if it materially affects performance or intended use—triggers a reassessment of conformity. Manufacturers who treat post-market surveillance as a compliance checkbox rather than a strategic capability are the ones who end up facing field safety corrective actions, or worse, market withdrawal.

What Your Team Should Do Next

If you're developing AI-based diagnostics or any SaMD targeting IVDR, these approvals should prompt three immediate actions. First, audit your clinical evidence plan against IVDR Article 56 and the relevant common specifications or harmonised standards. Ask whether your evidence demonstrates not just analytical validity but clinical utility—and whether your performance evaluation addresses real-world variability in user skill, patient demographics, and integration contexts.

Second, map your quality management system against ISO 13485:2016 with explicit attention to software lifecycle controls, risk management per ISO 14971, and IVDR-specific obligations around post-market surveillance and vigilance. If your QMS still treats software as an afterthought, or if your design controls don't account for iterative algorithm development, you're carrying regulatory risk that will surface during notified body review.

Third, build your post-market infrastructure now—before you submit, not after you clear. Define your performance metrics, establish data collection protocols, and integrate clinical feedback mechanisms into your product roadmap. If you're planning algorithm updates or adaptive learning models, your predetermined change control plan needs to be part of your initial submission. Regulators are increasingly scrutinising how manufacturers handle post-market modifications, and the bar is rising.

The Broader Signal: AI in Diagnostics Is No Longer Experimental

These two CE marks arrive amid a growing portfolio of AI-enabled diagnostics securing regulatory clearance across the EU, US, and UK. They reinforce a shift that's already underway: AI in healthcare is transitioning from research novelty to clinical standard of care. Regulators are no longer asking 'should we allow this?' They're asking 'how do we ensure it's safe, effective, and continuously validated?'

For manufacturers, this means the window for minimal viable compliance is closing. Regulators have seen enough poorly documented algorithms, opaque training datasets, and post-market surveillance plans that exist only on paper. The approvals that are landing now—like Spotlight Medical's and BrightHeart's—are coming from teams who invested early in rigorous evidence generation, transparent risk management, and quality systems that scale with the technology.

If your SaMD strategy is still treating regulatory compliance as a final-stage gatekeeping exercise, you're not just risking delays. You're risking irrelevance. The companies clearing regulatory hurdles today are the ones who embedded compliance into product design from day one—and who built post-market infrastructure capable of sustaining that compliance as their devices evolve.

Key Takeaways

• Spotlight Medical and BrightHeart's IVDR CE marks demonstrate that AI diagnostics can meet Europe's most demanding regulatory framework—but only with rigorous clinical evidence, transparent risk management, and robust post-market plans.
• IVDR compliance for SaMD isn't just about algorithm performance; it requires demonstrable clinical utility, real-world validation, and continuous post-market surveillance integrated into product lifecycle management.
• Multi-market SaMD strategies increasingly demand unified clinical evidence and QMS infrastructure that scales across EU, US, and UK regimes—regulatory expectations are converging, even if pathways differ.
• Post-market surveillance is no longer a compliance afterthought—it's a strategic capability that separates sustainable manufacturers from those facing field actions and market withdrawal.

The regulatory bar for AI-driven diagnostics is higher than it's ever been. But it's also clearer. Spotlight Medical and BrightHeart's approvals aren't outliers—they're proof that the pathway exists for manufacturers willing to invest in clinical rigour, transparent documentation, and post-market discipline. If your team is navigating IVDR, FDA SaMD guidance, or MHRA frameworks, the lesson is consistent: build your compliance infrastructure as carefully as you build your algorithm. Because in 2026, regulators are evaluating both with equal scrutiny.