PACS routing decisions differ in where rules run, which messages trigger a transfer and how the router connects to an archive. This article compares documented workflows from aycan, Laurel Bridge Compass, Dicom Systems Unifier and RamSoft Gateway Router [1][2][6][8]. We organize the comparison around routing rules, protocol scope, archive topology and the workflow responsibilities described in those documents. Standards and architecture references provide context for reading the named products; the groups used here are an editorial framework [3][5].

The first decision point is what “routing” is supposed to mean

Aycan describes its DICOM router as a distribution system for sharing DICOM data across disparate clinical and enterprise imaging systems, with flexible routing and HL7-triggered actions [1]. The same page states that routing can be configured for time of day, specialty, and source, and that rules are applied at the instance level [1]. The page also documents auto-routing based on image-level DICOM tags and rules scheduled for specific times or days [1]. That is a router-centered definition: studies arrive, rules fire, and data moves. The same page also mentions prefetching and distribution triggers tied to HL7 engine behavior, which puts some workflow logic around the transfer step without turning the product into a full departmental workflow standard [1].

Laurel Bridge’s Compass is described differently. Its page calls the product a routing workflow manager that automates ingestion, distribution, tracking, and access to imaging studies across PACS, VNA, and clinical IT environments [2]. The document also lists routing or multiplexing DICOM store jobs by any DICOM tag, integrating events and actions with HL7 communications, hot-folder import, DICOM structured report parsing, and support for both traditional DICOM DIMSE and DICOMweb messaging protocols, including conversion between the two [2]. That is a broader control surface than basic route rules alone, at least in the way the vendor documents the product.

Dicom Systems’ Unifier page likewise frames the product as a DICOM router and imaging workflow engine, with metadata-driven routing based on patient attributes, modality, destination availability, and workflow logic [6]. It adds load balancing, analytics, and workflow customization using LUA scripting, which suggests a platform for traffic control and workflow handling rather than a single-purpose send function [6]. The comparison point here is not feature volume for its own sake; it is whether the site wants a routing point, a workflow layer, or both [1][2][6].

Protocol support is the cleanest technical divider in the documents

Laurel Bridge is the clearest source on protocol breadth. Its Compass page explicitly says it supports traditional DICOM DIMSE and DICOMweb messaging protocols and can convert between them [2]. It also links routing with HL7 communications and secure transfer, which means the workflow can cross older modality patterns and newer web-based imaging paths [2]. For organizations with mixed estates, that matters more than a generic promise to route images, because the protocol boundary is where many integrations fail.

Aycan’s page describes DICOM transport and image-handling functions. It describes DICOM connections in the base configuration, automatic DICOM tag modification, compression options, and on-the-fly transcoding for legacy modalities [1]. The document also says it supports JPEG 2000 and lossless, lossy, and JPEG2000 compression, with automatic switching to uncompressed transfer if a device cannot handle compressed data [1]. That is useful for mixed modality fleets, and DICOMweb support or protocol conversion for a particular deployment should be confirmed with the vendor.

RamSoft’s product page describes rules-based DICOM routing, automated DICOM tag scripting, smart pre-fetching, multithreading, flexible storage, and vendor-neutral archiving. It also describes Modality Worklist support and links to separate DICOM conformance details [8]. A buyer should review that separate document and demonstrate the proposed transfer between the actual systems to be connected; the product page alone does not establish compatibility for a particular installation.

Concrete evaluation example: a teleradiology intake test

A useful demonstration suggestion would be to stage the same incoming study through three scenarios: a classic DICOM modality feed, an HL7-triggered prefetch or route event, and a cross-enterprise send to a second archive or read location. The point would be to observe whether the product is reacting only to study arrival, whether it can use message-driven triggers, and whether it can route or transform to more than one destination under the site’s rules [1][2][6][8].

For a disciplined test, the committee could record four things only: which DICOM tags or HL7 events the workflow depends on, whether the system can handle DIMSE only or DIMSE plus DICOMweb, whether the route can be directed to multiple destinations, and whether the product sits inside a larger PACS/VNA stack or outside it [1][2][5][6][8]. That does not prove clinical value; it does show which control model each product actually exposes [5].

The workflow question is really about where the data model lives

The IHE Scheduled Workflow profile gives the reference model. It describes a flow that integrates ordering, scheduling, imaging acquisition, storage, and viewing, and it ties HL7-based registration, order placing, and scheduling systems to DICOM-based modalities and PACS through Modality Worklist, Storage Commitment, and Modality Performed Procedure Step [7]. In other words, SWF is not a router product; it is the workflow grammar that explains what a router is trying to preserve or trigger [7].

That matters because the vendor documents sit at different layers of that grammar. Aycan says routing rules can depend on time of day, specialty, source, and instance-level conditions, and that HL7 order messages on study arrival are optional [1]. Laurel Bridge says Compass can parse DICOM structured reports and route them into dictation and transcription systems, and it can integrate events and actions with HL7 communications [2]. Dicom Systems says Unifier uses metadata-driven routing and can run custom workflow logic, while RamSoft describes rules-based routing, smart pre-fetching, and Modality Worklist support [6][8].

The decision point is whether the site wants the imaging path to react mainly to arrival of a study, or whether it also wants the routing layer to participate in upstream order data, downstream reporting, or cross-enterprise exchange. The documentation reviewed for this article support that distinction, but they do not prove that every vendor handles every step in the same way [1][2][6][7][8]. A site that needs a single routing node for distribution will read these products one way; a site that needs workflow control across PACS, VNA, dictation, or web exchange will read them another way [2][6].

Architectures differ more than product labels suggest

The VA/DoD conformance document is useful because it separates architecture from function. It states that VA supports two architectures for digital acquisition modality interfaces, one using a commercial PACS and one using VistA as the PACS, while MHS supports only Architecture A [3]. The document also says the scope is the point of interface between the modality and the PACS, and that modalities should perform the same functions whether they connect directly to VistA Imaging or through a commercial PACS [3].

That makes routing comparisons easier to place. A router installed in an environment like Architecture A has a different job than one installed where the modality speaks directly to a PACS or archive. In one case it may be part of a commercial archive chain; in another it may sit in front of a VistA-based path or a deconstructed PACS model [1][3]. Aycan describes its solutions as vendor neutral and says they integrate into deconstructed PACS environments, and Dicom Systems says its platform supports on-premises, cloud, and hybrid environments [1][6].

RamSoft also places routing inside a larger archive and imaging suite, saying Gateway Router is built on the same SQL platform as PowerServer and combines routing with online VNA functions [8]. Laurel Bridge, by contrast, presents Compass as a routing workflow manager that can sit across PACS, VNA, and clinical IT systems [2]. The architectural comparison is therefore not “router versus router” only; it is “where does the routing layer sit relative to archives, worklists, and downstream consumers” [1][2][8].

Procurement should compare workflows, not just feature lists

The 2025 Journal of Imaging Informatics in Medicine article describes PACS selection as a committee process that starts by identifying pillars, then concepts within those pillars, then rating and weighting them before vendor demonstrations [5]. The excerpt shows that the committee used concepts such as workflow, performance, and user interface, and that standard virtual demonstrations and onsite demonstrations were part of the process [5]. That is useful here because routing products are often chosen inside a broader PACS or imaging platform procurement, not as isolated utility software.

For a demonstration, ask each vendor to identify the workflow problem being evaluated and show the relevant routing configuration. Ask the vendor to demonstrate the relevant routing task in the proposed configuration. If the site wants a workflow engine with load balancing and scripting, Unifier is framed that way [6]. If the site is already inside a larger PACS/VNA stack and wants routing inside that stack, RamSoft’s product page makes that positioning explicit [8].

The article also shows why documentation alone does not settle the issue: the committee’s ranking depended on local needs, presentations, and demonstrations, not on a universal hierarchy [5]. That matters for routing comparisons because the best fit depends on whether a hospital is trying to normalize a single modality path, connect multiple sites, or build an imaging layer around downstream analytics, transcription, or cloud exchange [2][5][6].

Evidence limits are part of the comparison

The documentation reviewed for this article is primarily vendor documentation plus one conformance requirements document, one workflow standard, and one procurement article. That means it supports documented capabilities and architectural framing, but it does not support head-to-head performance claims, market share, or a universal winner [1][2][3][5][6][7][8]. It also does not establish implementation effort, total cost, or whether any one product is better for every radiology environment [1][2][5][6][8].

There is also an important boundary between what the sources show and what they do not show. The documents do not prove that one product can or cannot meet a particular institution’s interface inventory, network design, or reporting stack without configuration work; they only show the vendors’ documented descriptions [1][2][6][8]. The VA/DoD document confirms that interface architecture choices depend on the environment, and the procurement article confirms that local committees use weighted criteria and demonstrations [3][5].

So the clearest comparison is conditional: basic router, workflow manager, workflow engine, or routing inside a broader imaging suite. Those are different procurement questions, and the documents support treating them as such [1][2][6][8].