For most of its clinical life, intravascular imaging in the coronary tree has been evaluated as a report card. Did the operator get the stent expanded properly? Is there malapposition, edge dissection, or under-expansion? The big randomized trials of the last few years — RENOVATE-COMPLEX-PCI, ILUMIEN IV, OCTIVUS, OCCUPI, IVUS-ACS, DKCRUSH VIII — were largely structured around that question. Does adding imaging guidance to PCI improve outcomes versus angiography alone? The answer came back yes, repeatedly, across complex lesion subsets.

The 2025 ACC/AHA/ACEP/NAEMSP/SCAI Acute Coronary Syndromes Guideline elevated imaging guidance to a Class I recommendation. The 2024 ESC Chronic Coronary Syndromes Guideline did the same through its own pathway.

That question is now substantially settled. EuroPCR 2026, which closed in Paris last week, showed the field quietly moving to two newer questions that matter more for how imaging actually gets used. One is about interpretation. The other is about decisions.

The interpretation problem nobody priced in

A modest survey study presented at EuroPCR put a number on something experienced operators have always known intuitively. Van der Zande and colleagues surveyed 43 clinicians, 41 of them interventional cardiologists, on whether AI assistance changed their confidence in OCT-guided decision-making during PCI. AI support raised decision confidence, and the largest gains accrued to less-experienced OCT users. The interpretation gap between expert and non-expert readers narrowed with AI in the loop.

This matters because the intravascular imaging adoption debate has been almost entirely a reimbursement debate. The standard story: U.S. intracoronary imaging penetration sits at 15 to 20 percent, Japan sits near 85 percent, and the gap is a payment-architecture problem. That story is true but incomplete. There is a second barrier that gets almost no airtime — the cognitive and time cost of reading dense cross-sectional image stacks accurately, under pressure, in a live case.

A high-volume operator at an academic center reads OCT and IVUS fluently because of rep count. A mid-volume operator in a community lab running a full day of cases does not accumulate the same volume of reads, and the marginal interpretation burden of a pullback is real even when the procedure is fully reimbursed. The imaging is only as valuable as the operator’s ability to act on it correctly in the moment. If that interpretation step is slow or uncertain, the imaging gets skipped regardless of coverage.

If AI image interpretation reliably closes that gap — and the Van der Zande signal is early but directionally clear — then the adoption curve may bend on usability before it bends on reimbursement. That is a different lever than the one most of the field has been pulling. It is also a lever that is moving fast, because AI-assisted image characterization is improving on a software cadence, not a clinical-trial cadence.

The decision problem: imaging as a triage tool

The second shift showed up in the calcium-modification programming at EuroPCR, and it is more structural.

Suzanne Baron presented the IVUS substudy of the ShortCUT trial. The parent trial, presented at TCT 2025, had randomized 413 patients with calcified coronary disease across 21 U.S. sites to either intravascular lithotripsy or cutting balloon for lesion preparation before stenting, all under IVUS guidance. The headline result was noninferiority — cutting balloons came in noninferior to IVL on the primary endpoint of minimal stent area at the site of maximal calcification, at lower cost. The reasonable read at the time was “reach for the cutting balloon first.”

The IVUS substudy complicated that. When the population was stratified by calcium morphology and by whether rotational atherectomy was planned upfront, the equivalence broke apart. In more complex calcium — 360-degree arc, nodular morphology — without planned atherectomy, IVL achieved a greater minimal stent area than the cutting balloon.

After upfront rotational atherectomy, the cutting balloon matched IVL at lower procedural cost. The right tool depended on the calcium pattern and the preparation sequence, and the thing that revealed the calcium pattern was the imaging.

That sat inside a broader EuroPCR session comparing rotational atherectomy, orbital atherectomy, and lithotripsy head-to-head, plus the ICARE-OFDI randomized comparison of IVL against rotational atherectomy. The collective signal across all of it: the calcium-modification toolbox is now broad enough that the operative question is no longer “which device is best in general.” It is “which device fits this specific calcium in this specific vessel, given what I’ve already done to it.”

That question can only be answered by characterizing the calcium before preparation. Arc, depth, nodularity, and length drive the modality choice. A 270-to-360-degree superficial calcium arc behaves differently from a deep calcium sheet, which behaves differently from an eruptive nodular calcium. IVUS and OCT each see these features with different strengths — OCT resolves superficial calcium thickness and crack propagation better, IVUS penetrates deeper and reads the full arc in heavily attenuating lesions. The imaging is becoming the triage step that routes the patient to the right preparation pathway.

Why this reframes the imaging category

For most of the last decade, the commercial and clinical narrative around coronary intravascular imaging has been about penetration. What share of PCI procedures use imaging, and how do you get that share up. The implicit model treated imaging as an add-on layer — valuable, guideline-supported, but optional and reimbursement-gated.

The two shifts at EuroPCR point at a different model. If imaging becomes the tool that selects the calcium-modification strategy, it stops being an optional verification layer and starts being a decision node in the procedural workflow. You cannot rationally choose between lithotripsy, cutting balloon, and atherectomy on a heavily calcified lesion without characterizing the calcium first. The angiogram does not show you calcium arc and depth.

The imaging does. That makes imaging a precondition for optimal device selection rather than a nice-to-have confirmation of a decision already made.

And if AI assistance lowers the interpretation cost of doing that characterization, the two shifts compound. Cheaper, faster, more confident image reads make it more practical to image before lesion prep, which makes the modality selection better, which makes the imaging more clinically load-bearing, which strengthens the case for routine use.

That is a more durable adoption thesis than the penetration-versus-reimbursement framing. Reimbursement gates an optional add-on. Workflow necessity does not get gated the same way — when a step is required to do the next step correctly, it gets done. The calcium-modification decision is becoming a place where imaging is hard to skip.

What to watch

Three things. First, whether the AI-interpretation tooling — from the imaging vendors and from independent software players — matures fast enough and gets regulatory clearance broadly enough to actually move the interpretation-burden barrier at scale. The Van der Zande study is a confidence survey, not an outcomes trial; the outcomes evidence will take longer.

Second, whether the calcium-modification-selection use case gets formalized into guideline language or appropriate-use criteria. Right now it is operator judgment informed by trial subgroups. If a society document codifies “characterize calcium morphology before selecting a modification strategy,” that is the moment imaging-for-triage becomes standard rather than sophisticated.

Third, whether the imaging vendors reorient their own positioning from “imaging improves outcomes” toward “imaging selects the right tool.” The clinical evidence already supports the first claim. The second claim is where the workflow becomes sticky.

EuroPCR 2026 did not produce a single blockbuster imaging trial. What it produced was subtler and arguably more important — two signals that the role of intravascular imaging in the cath lab is shifting from after-the-fact report card to real-time decision tool. That shift, if it holds, changes the adoption math more than another outcomes trial would.

Citations

1.    Van der Zande J, et al. Impact of AI assistance on clinician confidence in OCT-guided decisions across experience groups. Abstract A101837Jv. EuroPCR 2026, Paris, May 19-22, 2026.

2.    Baron S. Intravascular lithotripsy vs. cutting balloon by calcium type: IVUS substudy of the ShortCUT trial. Presented at EuroPCR 2026, Paris, May 19, 2026.

3.    ShortCUT trial (parent). IVL vs. cutting balloon angioplasty in calcified CAD; 413 patients, 21 U.S. sites; primary endpoint minimal stent area by IVUS at maximal calcification. Presented at TCT 2025, San Francisco, October 2025.

4.    Honton B, Motreff P, Mallet J-S, et al. Intravascular lithotripsy in comparison to rotational atherectomy for calcified lesions: the ICARE-OFDI randomized trial. EuroIntervention. 2026, Epub ahead of print.

5.    IVUS-ACS trial. IVUS- versus angiography-guided PCI in acute coronary syndromes (48% 1-year TVF reduction with IVUS). Lancet 2024.

6.    DKCRUSH VIII. IVUS-guided versus angiography-guided PCI in complex bifurcation lesions, DK crush technique. Presented at ACC 2026, New Orleans, March 30, 2026.

7.    Rao SV, et al. 2025 ACC/AHA/ACEP/NAEMSP/SCAI Guideline for the Management of Patients With Acute Coronary Syndromes. J Am Coll Cardiol. 2025;85:2135-2237.

8.    Vrints C, et al. 2024 ESC Guidelines for the management of chronic coronary syndromes. Eur Heart J. 2024;45:3415-3537.