Reporting Redefined
Start from the image, not a blank page.
Sirona's reporter ingests DICOM pixel data, so AI can work from the images. The reporter doesn't interpret images itself: the radiologist reviews, edits, and accepts everything that enters the report.
How Sirona is Different
You can't accelerate reporting without pixel data
Reporting platforms have spent two decades optimizing text workflows — better speech recognition, smarter templates, faster macros. But the images themselves have always been invisible to the reporter. That's a ceiling. AI can now auto-label anatomy, surface findings, propose measurements, and draft impressions for a radiologist to review — but only if it can work from the images. Sirona's reporter ingests DICOM pixel data natively, even when deployed as a standalone reporting solution without the viewer. Every AI acceleration that matters in the next decade starts with pixel access.
What the reporter does today — and what pixel access enables next
When the reporter can see the images, everything about reporting changes — from anatomy labeling to impression drafting to the long-term trajectory toward fully AI-generated reports.
Spine Suite
Automatic spine labeling from the images, surfaced in the report for you to review and accept.
Explore Spine Suite
Pixel-Word Link
Linking a finding in the report to the region of the image it describes is what pixel access makes possible. With Anatomic Navigator, a disc level in the report jumps the viewer to that level.
Explore Reporter
AI-Drafted Impressions
Auto-Impressions drafts the impression from your dictated findings in seconds, inside the same platform that holds the images and priors — with style tuning to match how you write.
Explore Auto-Impressions
Quality Assist (Coming Soon)
Designed to check report quality in the background while you dictate, surfacing potential issues as suggestions rather than interruptions.
Explore Quality Assist
Priors Summary
AI condenses the patient's prior reports into a short synopsis, so the relevant history is in front of you when you start the exam.
Explore Platform AI
Foundation Model Trajectory
Today: AI drafts impressions, summarizes priors, and labels the spine. Tomorrow: foundation models that draft complete reports from images alone. A reporting platform positioned for this future is one that already ingests pixels.
Multimodal AIThe pixel-powered argument
Speed Ceiling
Text-only reporting has hit its ceiling
Macros, templates, and better speech recognition can only take you so far. The step change comes when AI can work from the images while the radiologist reads — labeling anatomy, proposing measurements, linking findings to the pixels they describe. Each of those requires pixel access.
Anatomic labeling can't exist without pixels
Auto-Impressions sits in the same platform as the images and priors
Linking a finding to its image region starts with pixels
Every next-generation acceleration starts with pixels
Architecture
Why this is hard to retrofit
Traditional reporters are HL7-in, HL7-out systems — text architecture, not image architecture. And analyzing pixels for clinical use brings a product into scope as a regulated medical device, which requires design controls, verification, and validation traced back to the product's inception. That is hard to retrofit onto a shipped product, which is why text-architecture reporters typically rely on a separate system for anything involving pixels.
A text-only data model has no place for pixels
Design controls require a design history back to inception
Retrofitting design controls onto a shipped product is a rebuild
Pixel analysis done elsewhere stays outside the report
AI-Native Gateway
Pixel-native is the entry condition for AI-native
Multimodal AI — the thing that will reshape radiology over the next decade — analyzes images, text, and audio in one pass. A reporting platform without pixel access can't host it. Choosing a text-only reporter today means changing reporters again when that AI arrives.
Multimodal foundation models require pixels
Agentic workflows will coordinate across image and text
Speech-to-Action already drives the viewer and the reporter together
The AI roadmap starts at pixel access
Performance
What pixel-powered reporting delivers today
Sirona's reporter ingests DICOM pixel data natively, in production. Today it drafts impressions from your dictated findings and summarizes prior reports, with the radiologist reviewing and accepting everything that enters the report. Spine Suite shows where pixel access leads: labels drawn from the images, surfaced in the report.
Impressions drafted from your findings, for you to edit and accept
Auto-Impressions adapt to your reporting style
Prior reports summarized before you start
Built under FDA design controls
510(k)
clearance for the Sirona Advanced Imaging Suite — October 2025
DICOM
pixel data ingested natively by the reporter — not just HL7 text
Drafted
impressions from your findings, for you to edit and accept
What pixel-powered reporting feels like
10-Minute Demo: Why Unified Data is Necessary for Reporting's Future
Sirona's Chief Technology Innovation Officer, Dr. Mark Longo, demonstrates why Sirona's unified platform is necessary for the future of reporting AI. If AI is going to transform radiology reporting, that process starts with pixels and ends with action — as Dr. Longo illustrates with a demo of Sirona's production platform.
“We built our AI on Sirona because it's the only infrastructure where the model, the images, and the report can live in one place.”
Rustin Rassoli
Founder & CEO, Epsilon Health
FAQs
What makes a reporter 'pixel-powered'?
Why can't a text-based reporter do this?
Does this work as a standalone reporter?
Is Sirona an FDA-regulated medical device?
How does this pair with the Platform AI features?
What's the long-term roadmap?