SIIM 2026 AI Landscape

The SIIM 2026 Annual Meeting sits at the intersection of imaging informatics, enterprise imaging, interoperability, and clinical AI deployment. Compared with general radiology congresses, its programme is more operational by default, which makes it a useful signal for where AI is showing up in day-to-day imaging infrastructure rather than only in research demos.

This analysis focuses on 72 radiology-related sessions, including 49 sessions with an AI focus and 39 research abstracts. It follows how AI appears across session formats, learning topics, imaging specialties, themes, and speaker roles¹.

The resulting signal is clear: SIIM 2026 is not just about AI in a generic sense. It is concentrated around workflow, standards, enterprise imaging, digital pathology, and education, with a strong abstract pipeline and a speaker mix anchored by academic and technical operators.

The more important pattern is not only the size of the AI footprint, but where that footprint appears. AI is showing up inside the operational layers of imaging: workflow design, enterprise imaging, standards, education, governance, and data infrastructure. That suggests the conversation is moving beyond standalone model performance toward the practical architecture needed to deploy, monitor, and integrate AI across imaging organizations.

Key signals

• AI at SIIM is becoming infrastructure, not just content. 49 of 72 radiology-related sessions focus on AI, representing 68.1% of the total.
• The meeting is closer to the adoption layer than the invention layer. Education Session is the main AI delivery format with 30 AI sessions.
• AI is spreading sideways into informatics. Beyond the dedicated AI topic, Productivity & Workflow is the largest cross-cutting learning topic, with 50 sessions, including 38 AI-focused sessions.
• The agenda is implementation-first, not modality-first. Workflow / productivity is the broadest theme, with 56 sessions.
• The research pipeline is still academic, but the problems are practical. The 39 research abstracts are mostly led by Trainee / early career presenters in Academic / health system settings.

Where AI appears in the programme

AI is distributed unevenly across SIIM formats. Education Sessions dominate absolute volume, which suggests AI has moved into mainstream professional education rather than remaining isolated in technical or research-only tracks. Applied Informatics Abstracts and sessions grouped under the Research Abstracts format are fully AI-focused in this analysis, showing that the innovation pipeline is still active, but the larger signal is that AI is now being taught, operationalized, and discussed across practical meeting formats.

Learning topics

The learning-topic distribution suggests that SIIM's AI conversation is tightly linked to implementation infrastructure. Productivity & Workflow, Enterprise Imaging, and Standards all show meaningful AI activity, which points to a field asking how AI can be embedded into imaging systems, data flows, and organizational workflows. That is different from a modality-first AI agenda, where the focus would be dominated by individual detection or classification tasks.

Themes point to deployment maturity

The thematic distribution is more deployment-oriented than modality-oriented. Workflow / productivity, Education / training, and Clinical AI deployment / SaMD / regulation all sit near the top of the stack, reinforcing the view that SIIM is tracking how AI gets operationalized, governed, and taught.

Foundation-model content is highly AI-concentrated, with 16 of 17 sessions focused on AI (94.1%). That concentration suggests foundation models are still being treated as a distinct technical frontier rather than a fully absorbed part of routine imaging operations. In contrast, workflow and enterprise-imaging themes indicate areas where AI is already being pulled into broader operational discussions.

This is also where SIIM differs from broader radiology meetings. A modality-driven conference would usually show AI clustered around anatomy, disease categories, and model performance. Here, the stronger signal is the connective tissue around AI: deployment, standards, education, workflow, and enterprise imaging. That makes the meeting especially useful for tracking the implementation layer of the radiology AI market.

Speaker composition is operator-heavy

Speaker composition also reflects the meeting’s practical orientation. Academic and health-system affiliations account for 71.6% of all speaker appearances, while technical, trainee, and executive/product roles make up the largest role clusters.

The speaker mix matters because AI adoption in imaging is no longer shaped only by model developers. It increasingly depends on people who manage clinical workflows, imaging infrastructure, governance, procurement, integration, and education. The presence of technical, trainee, academic, physician, and executive/product roles in AI-associated sessions suggests that SIIM is capturing a multi-stakeholder implementation conversation rather than a narrow research discussion.

Research abstracts remain an academic pipeline

The abstract layer is still strongly academic. 82.0% of abstract presenters come from academic or health-system institutions, and Trainee / early career accounts for the largest presenter role segment.

Topic-wise, the abstract set leans toward education, foundation models, workflow, and data governance rather than a pure modality-by-modality distribution. That suggests the early research pipeline is not only producing new models, but also working on the surrounding systems needed for evaluation, adoption, and safe deployment.

Bottom line

SIIM 2026 points to an imaging AI market that is becoming more operational, not less technical. The strongest signals sit around workflow, enterprise imaging, standards, education, deployment, and governance, while the abstract pipeline still leans heavily academic and infrastructure-focused.

The practical takeaway is that AI at SIIM is being treated less as a standalone model category and more as part of the working architecture of imaging operations. For vendors, that means model performance alone is not enough. For health systems, it means AI adoption increasingly depends on integration, monitoring, workflow fit, and organizational readiness.

In that sense, SIIM is a useful signal for the next phase of radiology AI: not whether AI exists in imaging, but whether it can be implemented, governed, and sustained inside real imaging environments.

Get more reports like this

RadAI Slice publishes structured analysis of imaging AI across conference programmes, product activity, research, and vendor positioning. Subscribe to RadAI Slice for structured coverage and data-backed summaries.