Output gap vs legacy
+30%
AI PACS Buyers Guide
Imaging volumes doubled. The number of radiologists didn't.
Cumulative extra wRVUs / rad
+16,229
New Lantern
Legacy PACS (+1% / year)
Conservative planning model for directional budgeting only; not a guarantee of realized outcomes.
/
BACKGROUND
What is an all-in-one AI PACS?
A PACS (Picture Archiving and Communication System) is the software radiologists use to view, manage, and report on medical images. A traditional PACS was built to store and display images, nothing more. An AI PACS does all of that, but also integrates AI models directly into the reading workflow. Instead of bolting AI onto a legacy platform, an AI PACS is purpose-built to surface AI findings, auto-fill reports, and manage worklists intelligently, in real time, as you read.
The key difference is architecture. A true cloud PACS is built cloud-native from the ground up, with zero installs and continuous updates, working across every site and every modality from a single web-based workspace. At New Lantern, the AI worklist, advanced viewer, and AI-powered reporting all live in one unified platform. One workspace, no on-premise servers to maintain, and updates that ship without taking the reading room offline.
Replaces systems like
Philips
Intelerad
Sectra
Fujifilm
Merge
GE
Powerscribe
Fluency
/
COMPARISON
AI PACS vs traditional PACS plus AI.
Reporting
Dictation into separate reporting software. Tab through templates, dictate dates, measurements, scores, and more.
AI drafts the report as you read. On eligible measurement-driven exams, Curie's OCR auto-fills structured findings in seconds. Dictate only positive findings, let AI put it into the right place. and generate an impression that sounds like you.
Worklist
Highly complex, or completely static list. Cases sorted by modality or time. No prioritization logic. Logic that's impossible to decode.
Intelligent worklist surfaces cases by STAT priority, modality, and assignment. Zero manual prioritization. AI agents to help build, manage, and contextualize study routing and load balancing.
Infrastructure
On-premise servers required for any expansion. Heavy IT requirements for expansion and new sites. Site-specific installs.
Cloud-native. Zero installs. Any device, any location. Read from the browser.
EMR / EHR Integration
Custom integrations required. Often expensive and fragile.
Use AI to chat with the patient chart. Integrate with EPIC, and get patient context instantly without switching systems.
AI Models
Require multiple integrations across the worklist, reporter, dictation, or PACS viewer.
AI built into every step. Prior surfacing, auto-detection, auto-filling, and structured output natively integrated. One integration point for CAD tools that works across the entire platform.
/
CAPABILITIES
The core capabilities of a modern AI PACS.
Cloud PACS architecture
A true cloud PACS, built from the ground up rather than bolted onto on-premise servers. No client installs, no per-site hardware. Access from any device, anywhere.
AI-assisted reporting
Curie's OCR auto-fill pulls measurements directly from handwritten or digital tech sheets for calcium scoring, DEXA, and specific ultrasound studies like carotid, renal artery, and thyroid. Structured reports in seconds. Review and sign.
Dictation engine
Trained on hundreds of thousands of real radiologists' speech patterns and voices, so it understands radiology terminology and individual dictation style from day one.
Intelligent worklist
Cases surface in the right order by STAT priority, modality, and assignment. Query across all sites. Zero manual prioritization.
Native EMR/EHR integrations
Native HL7 FHIR and DICOM integrations included out of the box. No expensive custom middleware required.
Specialty modality support
CT, MRI, X-ray, ultrasound, mammography, and nuclear medicine. Full multi-modality support, with third-party AI findings from partners like Gleamer and Riverain surfaced directly in the viewer.
Cloud PACS your way: on-premise, cloud-hosted, or hybrid.
On-Premise
Legacy infrastructure, still supported
New Lantern can connect with on-premise infrastructure. You get the AI PACS experience without requiring an immediate full migration.
No full migration
DICOM compatible
Cloud-Hosted → Recommended
Fully managed cloud. Zero servers.
Everything runs in the cloud. No hardware to buy, manage, or upgrade. No IT overhead. Just access from any browser.
Zero installs
Any device
Instant updates
Hybrid
Split across cloud and site.
Some workloads stay on-site while others shift to the cloud. Ideal for practices with existing infrastructure investments they aren’t ready to sunset.
Flexible migration
Best of both
/
TIMELINE
Onboard in a quarter, not a year.
A legacy PACS replacement takes 12 to 18 months and costs hundreds of thousands in IT resources. New Lantern onboards most practices in under 90 days.
01
Discovery and data audit
We map your existing DICOM infrastructure, modality list, and reporting workflow. No RFPs, no consultants.
02
PACS and EMR migration options
Historical imaging and prior studies are imported. Multiple migration integration options are available, with no downtime for your reading room.
03
Training and go-live
Radiologists are live in one session, with the first hospital often reading in a few weeks. No lengthy change-management program.
04
Continuous improvement
Updates ship automatically. AI models improve over time from your site’s data. Customer success check-ins every quarter.
/
DUE DILIGENCE
What to expect from a serious vendor.
→
A true cloud PACS architecture, cloud-native rather than a legacy server with a web viewer bolted on top.
→
AI models trained on real radiology reads and dictation, not general-purpose LLMs pointed at medical images.
→
Full multi-modality support that includes the hard studies, tomosynthesis mammography, PET/CT, and nuclear medicine, not just CT and MRI.
→
Structured reporting output, not free-text narratives passed through NLP after the fact
→
Native HL7, FHIR, and DICOM integration included in the platform, not a third-party middleware charge.
→
Image loading fast enough for high-volume reads on large studies, not throttled by on-premise hardware or bandwidth.
→
An intelligent worklist you control, with distribution rules by subspecialty, site, and modality and cherry-picking prevention built in.
→
An open platform that runs third-party detection AI such as Gleamer and Riverain inside the same viewer, not a closed system.
EVALUATION TOOLS
A buyer's checklist for evaluating an AI PACS.
0 / 8Questions answered
Want to see how New Lantern answers every question on this list? Request a live, un-scripted demo.
Request a DemoWe spent months shadowing radiologists before writing a line of product code, and the same pattern showed up everywhere: hours lost dictating routine measurements like calcium scoring, instead of interpreting images or catching pathology.
That's not a staffing problem. It's an architecture problem. Too many radiology platforms are stitched together through acquisitions: a worklist from one vendor, a viewer from another, AI bolted on top of all of it. None of it was built as a single system for the radiologist actually reading the case.
New Lantern was built by Silicon Valley engineers and radiologists who were tired of watching their colleagues burn out on software that actively got in the way of good medicine. We didn’t patch the problem. We built an entirely new workspace from scratch.
Keep exploring.
New to AI PACS
New Lantern AI Reporting with Curie
How Curie auto-fills structured reports directly from your existing dictation workflow.
For Practice Admins
One dashboard. Every site. No surprises.
Volume metrics, TAT reporting, and RVU tracking across every site in one workspace.
For IT Teams
Zero servers. Zero installs. Minimal maintenance overhead.
Cloud-native means no servers to patch and no per-site hardware to maintain, so your team isn't carrying imaging infrastructure on top of everything else.