Source: Johnson & Johnson — 2026-08-17
Summary
Johnson & Johnson received U.S. 510(k) FDA clearance for MONARCH QUEST 3, the fourth software update to its robotic bronchoscopy platform in 18 months, used for lung nodule biopsy and lung cancer diagnosis. The update adds AI-powered nodule segmentation that auto-generates nodule boundaries with a single click during planning, improved registration and navigation accuracy to correct for anatomical shift mid-procedure, a 3D Compass overlay for scope-tip orientation, and broader compatibility with existing CBCT imaging systems.
Key Takeaways
- AI nodule segmentation replaces manual boundary-drawing during pre-procedural planning with a single-click, automatically generated outline — a concrete time-and-consistency win in a step that's traditionally manual and reader-dependent.
- The navigation/registration improvements specifically target CT-to-body divergence — the well-known problem where a patient's lung anatomy shifts between the pre-procedure CT scan and the actual live procedure, which is one of the main sources of missed or off-target biopsies in bronchoscopy.
- The 3D Compass overlay is a UI-layer fix aimed at reducing the disorientation clinicians report between joystick controls and the patient's actual internal anatomy while steering the scope.
- Broader CBCT (cone-beam CT) compatibility means hospitals can use imaging equipment they already own to manage the CT-to-body divergence problem, rather than needing to buy new imaging hardware alongside the robotic platform.
Reel Script
Hook (18s, ~40 words): A patient's lung doesn't stay perfectly still between the CT scan that maps it and the actual biopsy procedure minutes later. That gap is where robotic lung biopsies go wrong — and J&J just got FDA clearance for an AI fix aimed directly at it.
Core Concept (80s, ~185 words): Here's the clinical problem this solves. Robotic bronchoscopy — using a thin, steerable robotic scope to navigate deep into the lung's airways and biopsy a suspicious nodule — starts with a CT scan taken beforehand to build a 3D map. The robot then navigates that map during the actual procedure. The problem is called CT-to-body divergence: lungs aren't rigid, static objects. Breathing, positioning on the table, even the scope's own presence shifts the tissue slightly from where the pre-procedure CT said it would be. Navigate off that stale map without correction, and you can miss the nodule you're trying to biopsy by a meaningful margin — which either means a failed diagnosis or a second procedure. MONARCH QUEST 3's core upgrade is software that improves registration and navigation accuracy specifically to correct for this live shift, rather than assuming the CT map is still accurate once the procedure actually starts.
Hands-On (65s, ~150 words): Three concrete features stack together here. First, AI nodule segmentation: instead of a clinician manually tracing the nodule's boundary on the CT scan during planning — a task with real reader-to-reader variability — the system auto-generates that boundary with a single click. Second, a 3D Compass overlay layered onto the live view, designed to make the mapping between joystick movement and where the scope tip actually is in 3D space more intuitive — reducing the disorientation that's a real, reported source of navigation error in robotic bronchoscopy. Third, expanded compatibility with CBCT — cone-beam CT — imaging systems hospitals already have installed, which lets clinicians re-image and re-register mid-procedure to correct for that anatomical shift, without needing to purchase new imaging hardware alongside the robot itself.
Takeaway (22s, ~50 words): This is a good example of AI in medicine doing unglamorous, high-value work — not a diagnostic breakthrough, but tightening the accuracy of a mechanical process where a few millimeters of drift is the actual difference between a successful biopsy and a missed one.