Proscia: We Asked AI What’s Next in Digital Pathology for 2026. Then We Checked It.

The latest from the Proscia blog

With the end of 2025 in sight, it’s only natural to ask what’s next, especially as digital pathology keeps advancing. Predictions are everywhere, and a lot of them are painfully obvious: adoption will accelerate, AI in pathology will keep improving, interoperability will remain a top priority. True, but not particularly useful.

So we tried something different. We asked AI to generate bold, realistic 2026 predictions for digital pathology, AI, and precision medicine, with one constraint: no filler trends, no generic adoption narratives. We wanted signals that point to where the industry is actually headed for laboratories and life sciences organizations.

And because the best AI work still needs a human at the helm, I’m adding context and conviction to what surfaced. Here’s what rose to the top, and my reality check on what we’ll actually see once the New Year rolls around.

1. Prediction: Success in digital pathology and AI will depend on proven impact, not adoption itself.

Digitization is becoming assumed. The new standard is impact. Labs, biopharma teams, and health systems are going to be much less patient with “cool tech” that doesn’t change outcomes or efficiency in a measurable way. The winners will be the groups that can point to real gains: reduced variability, faster trial readouts, better patient stratification, or clearer links between tissue-based biomarkers and response.

Nathan’s reality check: Agreed. Improved clinical metrics and economic return on investment (ROI) remain supreme, and we now have enough evidence to provide that they are achievable. For example, laboratories that have gone digital are laying the foundation for generating millions in high-margin revenue through collaboration with life sciences organizations. 

2. Prediction: Pathology images will increasingly be treated as a shared digital asset across life sciences and diagnostics.

This convergence is farther along than people think. Whole slide images (WSIs) are no longer “used up” after a read. They’re being reused to discover predictive and prognostic biomarkers, train AI models, set clinical trial endpoints, and guide diagnostic decision-making. In 2026, the industry leans harder into that reality. Pathology images start to live on a continuous path from diagnosis to discovery to development, and the value of a slide is measured by how many decisions it can support over time.

Nathan’s reality check: I agree directionally, but I’d argue that AI glosses over the practical challenges that need to be overcome. First, life sciences teams and diagnostic laboratories need more pathways for engagement. Proscia Aperture is one way forward, but the broader point is that the biopharma-lab ecosystem has to make that partnership routine. Second, images only become truly valuable as a cross-continuum asset when they are systematically linked to molecular profiling, clinical context, and patient outcomes. 2026 feels like a real inflection year for both challenges, with the tooling and the market pressure finally lining up.

3. Prediction: Trial matching moves upstream to the moment of diagnosis.

Eligibility criteria are getting tighter as therapies get more targeted. That makes downstream trial screening too slow and too manual to keep up. The move upstream is about timing and scale. When trial signals can be surfaced right at diagnosis, based on tissue and context, enrollment aligns with the routine care workflow, rather than a separate scramble. 2026 is when that workflow starts to feel necessary, not experimental.

Nathan’s reality check: I’ll add that the key dependency is on volume. You need enough cases reviewed digitally to surface a meaningful pool of candidates. The upside is that this same digital foundation is what enables AI to plug into matching and make patient surfacing earlier and more consistent. 

4. Prediction: New CPT and billing codes will grow, but consistent reimbursement will still lag behind and negatively impact adoption.

Coding and reimbursement never move in lockstep. We’ll see more formal ways to document digital pathology services and AI-enabled services, because the system needs language for what’s happening in practice. But stable, broad reimbursement tends to follow only after value is undeniable and utilization is easy to track. So 2026 likely brings better coding clarity, while payment remains uneven and use-case dependent.

Nathan’s reality check: This is where I’ll have to disagree. Adoption is accelerating because labs are seeing real clinical, operational, and economic ROI that doesn’t depend on direct reimbursement. Where reimbursement starts to matter much more is on the AI side, especially for diagnostic and prognostic algorithms that introduce new billable value and need a clear path to payment. In other words, 2026 may still be uneven on dollars, but the market won’t wait for perfect reimbursement to keep digitizing. The pressure will be on proving and codifying AI-driven clinical impact, because that’s where sustainable, scaled payment will be decided.

5. Prediction: Digital pathology becomes an enterprise IT deployment, not a departmental tool.

By 2026, digital pathology will be rolled out less like a standalone lab upgrade and more like core enterprise infrastructure. The scale of whole slide imaging and AI, plus the need for multi-site access, security, governance, and rapid updates, is pushing many organizations toward cloud-based deployment as the practical default. In other words, pathology digitization stops living inside one department’s budget and technical stack and starts living inside the enterprise architecture where it can grow across networks and use cases.

Nathan’s reality check: What AI provides here is a clear, matter-of-fact explanation of what’s already unfolding. But more than that, the key will be striking the right balance among stakeholders: integrating pathology into an organization’s broader enterprise imaging strategy, while still ensuring that the primary users and beneficiaries of digitization get solutions tailored to the critical use cases, workflows, and behaviors that they know best. For this reason, I also predict that we’ll see a lot more collaboration between IT and pathologists next year. 

6. Prediction: AI will be routinely built into companion diagnostics, not tested as a one-off.

The momentum here is already visible in the way new precision therapies are being developed. As biomarkers become more nuanced, the idea of validating an algorithm as a late-stage add-on starts to feel outdated. In 2026, the real shift is operational: AI becomes part of companion diagnostic (CDx) program design from the beginning, with clearer expectations for evidence, performance monitoring, and clinical reliability across diverse populations.

Nathan’s reality check: Spot on. Biopharma’s investment in precision medicine is unprecedented, and the therapies coming out of that wave will only deliver if the diagnostics evolve with them. AI is uniquely capable of translating the rich pathology signal embedded in pathology images into insights about likely therapeutic response and patient outcomes. We already have a glimpse of where this is headed: in AstraZeneca’s TROPION-Lung01 Phase III work, a computational pathology–based TROP2 quantitative continuous score (QCS) predicted clinical outcomes for patients treated with Dato-DXd and is now being advanced as a companion diagnostic.

If there’s a single takeaway from these predictions, it’s that digital pathology is entering its outcomes era. The groundwork is being laid now for images to carry value far beyond a single diagnosis, and for AI to be judged by what it changes, not that it exists.

As you plan for 2026, prioritize what turns digitization into durable advantage: enterprise-wide deployments, repeatable collaboration across biopharma and diagnostics, and measurable impact. That’s where the real momentum is headed.

SOURCE: Proscia