Platform & software · Cloud based PACS

Cloud based PACS: cloud PACS systems, vendors and pricing compared

The short answer

A cloud based PACS stores and serves imaging studies from a vendor-operated data center rather than from servers in your building, so radiologists reach the archive over the internet instead of over the hospital LAN. Cloud is a hosting decision, not a regulatory category: the FDA clears the software function under product code LLZ (21 CFR 892.2050), and of the 2,285 clearances ever granted under that code, only 7 carry the word cloud in the device name. That means you cannot tell whether a cloud PACS has been through FDA review by reading its marketing, and several widely promoted cloud PACS products hold no 510(k) under their own name at all. Look the vendor up by applicant name in the 510(k) database before you shortlist.

The phrase cloud PACS gets attached to at least four different products. One is a traditional PACS someone installed on a rented virtual machine. One is a vendor-managed instance of conventional PACS software. One is genuinely cloud-native, written to run multi-tenant from the start. And one is not a PACS at all: it is a cloud archive sitting behind the on-premise PACS you already own. All four are sold with the same word, and they do not cost the same, migrate the same, or fail the same way.

The question that separates them fastest is not technical. It is regulatory. Image management and processing software is a Class II device in the United States, cleared under product code LLZ, and the clearance attaches to the software function rather than to where the servers sit. So the useful test is simple: does this vendor appear as the applicant on a 510(k) under LLZ, and for which product? Some cloud PACS vendors have a long clearance history. Some have none. The table further down this page lists real records, with K numbers you can verify yourself.

Radiological.ai is not a PACS and does not replace one. It works alongside whatever archive you run, cloud or on-premise, over standard DICOM and HL7 interfaces. It flags suspected findings for a second look, pushes urgent studies up the worklist and drafts the structured report into your template. The radiologist reviews, edits and signs every study.

Last updated September 2026

The Reading Station

Assistant ready

Worklist

SERIES 1 · AX

SLICE 24/64

SAMPLE STUDY

NOT FOR DIAGNOSTIC USE

W 80 · L 40

Structured report

Draft

Run the assistant to draft this report for review.

You review & sign

Illustrative sample · not a real patient study, not a diagnosis

Drafted in · you review & sign Worklist re-prioritized

Decision support for qualified clinicians. Radiological.ai does not provide a diagnosis and is not a substitute for professional judgment.

Run the assistant

Flag · prioritize · draft · you review and sign

X-RAY CT MRI BUILT WITH RADIOLOGISTS

Decision support not a diagnosis

You review & sign

Cloud is one axis of the buying decision and it is rarely the one that decides the project. If the open question is which PACS to buy at your study volume rather than where it runs, that is PACS software and PACS system vendors compared by study volume. If the real problem is where studies live long term and how they survive a vendor change, that is enterprise imaging and vendor neutral archive software, where the archive is treated as its own layer. The operational half, meaning scheduling, worklists and charge capture, sits in radiology information system software.

On the display side, whether a browser-based viewer is cleared for primary interpretation is a separate question from whether the archive is in the cloud, and it is covered in DICOM viewer software and which viewers hold an FDA clearance. For how an assistant like this one connects to whatever you run, see radiology PACS integration. For the wider picture of what imaging software costs, drawn from public federal contract awards, see our radiology software pricing research, and for a shortlist aimed specifically at outpatient sites, the best cloud PACS software for imaging centers.

Side by side

The four things vendors mean by cloud PACS

These are four different purchases with different cost curves, different exit costs and different failure modes. Establish which column a proposal actually sits in before you compare it to anything else, because a hosted instance of conventional PACS software and a multi-tenant cloud-native platform are not the same product with different branding.

What you are comparing	Self-hosted in your own cloud tenancy	Vendor-hosted single tenant	Cloud-native multi-tenant	Cloud archive behind an on-premise PACS
What it actually is	Conventional PACS software you install on rented infrastructure such as AWS or Azure	The same conventional software, run for you as a dedicated instance by the vendor	Software written from the start to serve many organizations from shared infrastructure	Your existing PACS keeps reading; long-term storage moves off-site
Who patches and updates it	You do, including the operating system and the database	The vendor, usually on their release schedule	The vendor, continuously, with everyone on one version	The archive vendor for storage; you still own the PACS
Typical commercial shape	Perpetual or subscription license plus your own infrastructure bill	Annual subscription per site or per study volume band	Per-study or per-user subscription, lowest entry cost	Storage priced per study or per terabyte, added to existing PACS costs
Where the exit cost hides	Low software lock-in, high operational burden	Migration is a vendor project and priced as one	Data export terms matter more than anything else in the contract	Lowest disruption, because the reading system does not change
Best fit	Health systems with a real cloud engineering team	Groups replacing an aging on-premise PACS without adding staff	Imaging centers and new sites starting without legacy infrastructure	Anyone whose PACS works but whose storage is full
Question to ask first	Who is on call at 2am when the VM stops responding?	What is the contractual restore time after an outage?	What format is my data in when I leave, and what does export cost?	Does retrieval of a 10-year-old prior add delay to the read?

Compiled September 2026. Clearance statements are from the openFDA 510(k) API, product code LLZ, queried September 2, 2026. Contract figures are from USAspending.gov prime award records and describe what federal buyers paid, not vendor list prices.

Primary source

Cloud and web based PACS vendors with a verified FDA 510(k)

Every row is a real record in the FDA 510(k) database, pulled through the openFDA API on September 2, 2026 under product code LLZ, which is image management and processing software under 21 CFR 892.2050. The K number is the most recent clearance held by that applicant; the count is every LLZ clearance under that applicant name. Look any of them up at accessdata.fda.gov. This is not a ranking and not a shortlist. It answers one narrow question most buyers never ask, which is whether the company selling you imaging software has been through review under its own name.

Applicant on the submission	Most recent LLZ product	K number	Decision date	LLZ clearances held
Fujifilm Medical Systems U.S.A.	ASPIRE Bellus II	K171463	2017-09-29	13
Sectra Imtec AB	Sectra Workstation IDS7	K081469	2008-06-12	12
Novarad Corporation	OpenSight	K172418	2018-09-21	6
Merge Healthcare Incorporated	Merge Universal Viewer (MUV)	K250301	2025-04-14	6
Intelerad Medical Systems	IntelePACS	K192176	2020-04-02	5
eRAD, Inc.	eRAD PACS	K241223	2024-10-31	5
RamSoft, Inc.	OmegaAI Image Viewer	K222476	2022-11-17	5
Change Healthcare	Change Healthcare imaging software	See accessdata.fda.gov	Multiple	4
DICOM Grid dba Ambra Health	Ambra PACS	K231360	2023-06-07	2
Visage Imaging GmbH	Visage Ease Pro	K142196	2015-04-28	2
Konica Minolta Healthcare Americas	EXA	K203743	2021-12-10	1

Why it works

What your group gets with cloud based PACS

Cloud is a hosting choice, and the FDA does not clear hosting

Product code LLZ covers image management and processing software regardless of where it runs. Across all 2,285 LLZ clearances ever granted, only 7 name the word cloud in the device, and those include products as different as a dental imaging package and a 3D segmentation service. So a vendor page that says cloud-based and FDA-cleared in the same paragraph is describing two unrelated properties. Verify the clearance by applicant name and product, not by the hosting description, and be aware that some well-marketed cloud PACS products return no LLZ record under their own company name at all.

The exit terms matter more than the monthly price

On-premise PACS locks you in with hardware and a migration project. Multi-tenant cloud locks you in with data gravity and export terms, which are cheaper to ignore at signature and far more expensive later. Ask three things in writing before you sign: what format studies are returned in, whether the export includes presentation state and report objects or only the pixel data, and what the export costs per study. Migration is commonly quoted at roughly $0.05 to $0.50 per study, so an archive of 200,000 studies is a five-figure line item that belongs in the business case rather than in a surprise a few years out.

Read latency is the clinical risk, not uptime

Vendor uptime numbers describe whether the service answered, not whether a radiologist could work. The measure that matters is time to first image on a large study over your actual connection, and time to retrieve a decade-old prior that has been moved to cheaper storage tiers. Ask for both, test them during the evaluation on your worst-connected site rather than at headquarters, and put the retrieval time for archived priors into the contract. A cloud PACS that serves new studies quickly and takes two minutes to return a comparison is a workflow problem that no service level agreement about uptime will surface.

What it handles

Flagged, prioritized and drafted for your review

The assistant pre-reads each study, surfaces a region of interest for review, re-prioritizes the worklist, and drafts the structured report in your template. You confirm, edit and sign.

Works with cloud and on-premise PACS alike
Connects over standard DICOM and HL7 interfaces
No change to your archive or storage contract
Flags suspected findings for a second look
Reorders the worklist so urgent studies surface first
Drafts the structured report into your template
Radiologist reviews, edits and signs every study

CLOUD BASED PACS STAT

Region of interest flagged for review

A focal region is surfaced on the sample study for the radiologist to review. The assistant does not characterize it as a diagnosis.

Draft impression

Suspected finding flagged for radiologist review. Correlate clinically and confirm. Draft for review and sign-off.

Illustrative sample · not for diagnostic use You review & sign

Why Radiological.ai

One assistant across the whole read

Not three vendors stitched together. Flag, prioritize and draft in one calm pane, on X-ray, CT and MRI, with the radiologist signing every study.

Flags suspected findings

A second set of eyes surfaces regions of interest for review on every study, so a suspected finding is less likely to slip past late in a shift.

Prioritizes the worklist

Suspected-critical studies move to the top, so urgent reads surface ahead of routine follow-ups across your sites and shifts.

Drafts the report

A structured draft arrives in your template, ready to edit and sign. The draft saves the typing and the measuring, never the judgment.

Good questions

Questions about cloud based PACS

Cloud PACS is a picture archiving and communication system whose storage and application servers run in a vendor-operated data center rather than in your building, reached over the internet instead of the hospital network. It performs the same job as a conventional PACS: it receives studies from the modalities, stores them, and serves them to a viewer for reporting. The difference is who owns and operates the infrastructure, and therefore who is responsible for capacity, patching, backup and disaster recovery.

Neither is better in general; they fail differently. Cloud removes the capital purchase, the refresh cycle and the server room, and shifts the risk to your internet connection and to the vendor's data export terms. On-premise keeps images on a fast local network and keeps you in control of the exit, at the cost of hardware, staff and a refresh every five to seven years. Sites without dedicated IT staff usually do better on cloud; sites with poor or single-path connectivity usually do not.

Almost no vendor in this category publishes list pricing, so any single figure quoted online is someone's estimate. What is public is what government buyers paid. Federal prime award records on USAspending.gov show PACS contracts ranging from roughly $250,000 for a single-site replacement at a small facility up to $11.1 million for a multi-site maintenance award across one VA network and $12.8 million for a national teleradiology PACS program. Price your own project per study and per site, and get migration quoted separately.

A product is not HIPAA compliant on its own; the arrangement is. For a cloud PACS you need a signed business associate agreement with the vendor, encryption in transit and at rest, audit logging you can actually read, and clarity on which subcontractors hold the data and in which country. Ask specifically whether the vendor uses a hyperscale cloud underneath, because that adds a subcontractor to the chain, and confirm that the business associate agreement flows down to them.

Clearance sits with the applicant company and the named product, not with the hosting model. Among vendors selling cloud or web-delivered imaging software, Fujifilm holds 13 clearances under product code LLZ, Sectra 12, Novarad and Merge Healthcare 6 each, Intelerad, eRAD and RamSoft 5 each, Change Healthcare 4, Ambra Health and Visage Imaging 2 each, and Konica Minolta Healthcare Americas 1. Several products marketed as cloud PACS return no LLZ clearance under their own company name, which is worth checking before a shortlist rather than after.

That depends on the viewer, not on the storage. Primary interpretation requires a viewer whose clearance and intended use cover diagnostic display, together with a display that meets your accreditation program's requirements. Plenty of cloud platforms pair a cleared diagnostic viewer for radiologists with a lighter reference viewer for referrers, and the reference one is explicitly not for primary reads. Confirm in writing which of the vendor's viewers is cleared for diagnostic use, because most vendors sell both.

No. It connects over standard DICOM and HL7 interfaces and works with cloud, hosted and on-premise systems alike, so it does not require you to change archives or storage contracts. It is decision support rather than storage: it flags suspected findings for a second look, prioritizes the worklist and drafts the structured report for review. It is not a PACS, it is not a diagnostic viewer, and the responsible radiologist reviews, edits and signs every study.

Read every AI radiology question

Explore more