February 1, 2026 — Copenhagen, Denmark — Radiobotics is excited to announce the release of RBfracture™ version 2.6, which focuses on improving transparency and confidence when assessing negative trauma X-ray examinations.
With this release, Radiobotics becomes the first AI-driven fracture detection provider to place explicit emphasis on the interpretation of negative exams. This is an area with significant potential to support efficiency in trauma radiology workflows.
Introducing Enhanced Transparency for Negative Findings
RBfracture already communicates positive findings using two confidence levels: A Solid-line bounding box for high-confidence findings and a dashed-line bounding box for lower-confidence findings.
In RBfracture version 2.6, this concept is extended to exams without positive findings.
The exam-wide negative classifications are:
- AI Negative
- AI Negative (very high confidence)
AI Negative (very high confidence) indicates no trauma-related findings with a false negative rate of 0.7%1.
Evidence-Based Development
The feature is supported by results from a retrospective study, Retrospective Evaluation of an AI System to Classify Negative Musculoskeletal Trauma Radiographs, which demonstrates RBfracture’s ability to reliably classify negative MSK trauma exams while maintaining patient safety.
For hospitals facing overcapacity and resource constraints, RBfracture 2.6 offers additional possibilities to explore more efficient patient workflows. These capabilities may support the evaluation of optimised patient routing and prioritisation strategies, within the hospital’s existing clinical governance framework.
Clinical Adoption in High-Volume Emergency Care
Södersjukhuset in Stockholm, Scandinavia’s largest emergency department, is already using RBfracture to support more efficient patient routing in a high-volume trauma setting.
“RBfracture 2.6 represents a meaningful step toward greater transparency in AI-supported trauma diagnostics,” said Peter Ulvskjold, CEO of Radiobotics. “By clearly distinguishing negative exams of very high confidence, we aim to give clinicians better insight into how the AI reaches its conclusions, without compromising safety.”
“RBfracture 2.6 represents a meaningful step toward greater transparency in AI-supported trauma diagnostics,” said Peter Ulvskjold, CEO of Radiobotics. “By clearly distinguishing negative exams of very high confidence, we aim to give clinicians better insight into how the AI reaches its conclusions, without compromising safety.”
References
1 Steffen Czolbe, Jonas Christophersen, Rikke Bachmann, Janitha Mudannayake, Andreas Nexmann, Fabrizio Perria, Nikolaus Knauer, Pavel Lisouski, Michael Lundemann, Retrospective Evaluation of an AI System to Classify Negative Musculoskeletal Trauma Radiographs, medRxiv 2025.11.27.25341142; doi: https://doi.org/10.1101/2025.11.27.25341142
