Heartflow CT-guided PCI matches IVUS outcomes in 1,105-patient P4 trial

The NEJM-published P4 trial found CT-guided PCI non-inferior to intravascular ultrasound while cutting procedure time by 17 minutes.

A brightly lit, minimalist medical room features a blurred white treatment chair in the foreground, a medical device with hanging tubes on a three-drawer cabinet, and a window with a plant and books in the background.

Heartflow has presented landmark randomised data showing that CT-guided percutaneous coronary intervention planning is clinically equivalent to intravascular ultrasound guidance, the current procedural gold standard, while delivering meaningful gains in efficiency. The findings, reported at Transcatheter Cardiovascular Therapeutics (TCT) 2026 in San Francisco and simultaneously published in the New England Journal of Medicine, are among the most substantive clinical readouts the AI-cardiac imaging company has produced since listing on Nasdaq.

The P4 trial enrolled 1,105 patients with stable coronary artery disease across 23 centres in the United States, Europe, and Japan. Participants were randomised to CT-guided PCI planning, using Heartflow's FFR CT Analysis and its Planner tool, or to intra-procedural IVUS. At 12 months, the primary endpoint of target-vessel failure occurred in 4.1% of the CT-guided arm versus 4.6% in the IVUS arm, meeting the pre-specified non-inferiority threshold.

Efficiency gains and procedural data

Beyond the headline clinical match, the CT arm demonstrated a 27% reduction in median procedure time, at 46 minutes against 63 minutes for IVUS, and a roughly 12% reduction in contrast volume (median 150 mL versus 170 mL). Lower contrast use is clinically significant because contrast-induced nephropathy remains a meaningful complication risk, particularly in patients with pre-existing renal impairment who make up a substantial share of the elective PCI population. The CT-guided cohort also more than doubled the rate of advanced calcium modification, at 9.2% versus 3.8%, which the company attributes to better pre-procedural lesion characterisation.

Principal investigator Carlos Collet, Director of Cardiovascular Imaging, Physiology and Translational Therapeutics at the Cardiovascular Research Foundation, said CT-guided planning "enables interventionalists to anticipate challenges, tailor their strategy, and enter the cath lab better prepared," framing the approach as one that converts pre-procedural intelligence into intra-procedural confidence.

PCI Navigator registry and competitive context

Supporting the P4 data, the NAVIGATE-PCI registry evaluated Heartflow's commercially available PCI Navigator tool in 203 patients across multiple centres. Investigators found that procedural strategies were modified in approximately 70% of cases after physicians reviewed Navigator's integrated 3D anatomy, plaque composition, and physiologic data relative to what angiography alone had suggested. The most common changes were to planned stent length (41% of modifications) and lesion preparation technique (12%).

The commercial read-across matters. Heartflow is positioning PCI Navigator as the only AI-driven PCI planning tool backed by prospectively validated accuracy and now by randomised trial data, a distinction that could be meaningful when the tool enters reimbursement conversations with payers and health systems. The analogy to transcatheter aortic valve replacement planning, where CT guidance is already routine, was invoked by Heartflow's chief medical officer Campbell Rogers and is likely to feature in those discussions.

The broader landscape for AI-assisted cardiac imaging is competitive. A number of companies, including Cleerly, Nanox Cardiology, and platform offerings from established imaging vendors, are pursuing CT-derived coronary analysis, though few have randomised procedural-outcomes data at this scale. For Heartflow, P4 provides a durable clinical anchor at a time when demonstrating real-world impact beyond diagnostic accuracy is increasingly expected by guideline bodies and health technology assessment agencies in Europe and the UK. MHRA and NICE pathways for AI-driven cardiac software remain a work in progress, and data of this type are precisely what those bodies will want to see before recommending routine commissioning.