InVision delivers clinically-validated AI for the most common cardiac imaging exam in the world — streamlining interpretation, standardizing measurements, and surfacing disease that today goes undiagnosed.
Paid on both sides of the house. Outpatient: CPT Category III code 0932T (effective January 1, 2025) covers adjunctive AI analysis supporting heart failure subtype classification — Precision Cardiac Amyloid is in scope when LVEF is 50% or above, and Precision LVEF establishes that threshold in the same pass. Inpatient: CMS has approved a Medicare New Technology Add-on Payment of up to $2,275 per eligible stay, on top of the MS-DRG, effective October 1, 2026.
Echo interpretation consumes over an hour of physician and technician time per exam. Precision LVEF has been shown in a blinded, randomized trial to reduce sonographer and cardiologist effort while improving measurement consistency.
FDA 510(k) clearance for both Precision LVEF and Precision Cardiac Amyloid. Cardiac Amyloid was previously granted Breakthrough Device Designation and enrolled in the FDA's Total Product Lifecycle Advisory Program (TAP).
Vendor-neutral by design: GE, Philips, Siemens, Canon, and handheld POCUS. InVision reads the standard DICOM your lab already produces, so there is no new hardware, no new views, and no acquisition change — it deploys alongside your existing PACS, CVIS, and Epic EHR. Typical implementation is about four weeks from kick-off to go-live.
InVision runs in your data center or in our validated cloud. PHI never has to leave your network, and on-premise no image data is retained once processing completes. Customer data stays in the United States either way — the deployment model is a procurement and security decision, not a product limitation.
Cardiac amyloidosis is rare, under-diagnosed, and now treatable. Our algorithms analyze every routine echo at zero marginal clinician burden — turning your existing imaging volume into a population-health instrument.
LVEF and Cardiac Amyloid are live. Precision Cirrhosis holds Breakthrough Device designation, and Precision Reporting — powered by our Nature-published foundation model — is in active development. A single deployment path that grows across your cardiovascular service line.
Automated evaluation of left ventricular ejection fraction from standard echocardiographic DICOM images. The only AI-Echo product to undergo a blinded, randomized prospective clinical trial — published in Nature, demonstrating time savings and improved clinical precision.
Learn more about Precision LVEFAI Detection for cardiac amyloidosis from apical-4-chamber and parasternal-long-axis echocardiographic views. Developed in collaboration with Alexion, AstraZeneca Rare Disease, to accelerate early detection of a life-threatening and routinely missed condition.
Learn more about Precision Cardiac AmyloidAutomated, structured echocardiographic reports built on two complementary open-source foundation models. EchoPrime (Nature, 2026) — a multi-view vision-language model trained on over 12 million video-report pairs — delivers holistic clinical interpretation across 23 diverse benchmarks of cardiac form and function, validated in five international health systems. EchoNet-Measurements (JACC, 2025) — trained on 877,983 sonographer annotations from 155,000+ studies — automates 18 standard B-mode and Doppler parameters with accuracy comparable to expert sonographers (R² 0.97 — 0.99).
Learn more about Precision ReportingOpportunistic detection for cirrhosis and steatotic liver disease directly from the subcostal view already captured during routine echocardiography. Built on EchoNet-Liver — a deep learning pipeline developed from over 1.5 million echocardiogram videos and validated against paired abdominal ultrasound and MRI across two large academic medical centers.
Learn more about Precision CirrhosisPrecision Valvular Heart Disease. The same research foundation extends to multi-valve assessment. Mitral regurgitation detection was published in Circulation (2024), and automated deep learning phenotyping of tricuspid regurgitation — led by Vrudhula et al. — was published in JAMA Cardiology (2025). Aortic valve work is in active development, with clearance pathways to follow.
Learn more about Precision VHDOur foundation is a decade of academic work at Stanford and Cedars-Sinai — culminating in the first and only prospective, blinded, randomized clinical trial of AI in cardiology. Here's the evidence trail, in the words of the journals that published it.
InVision is founded and led by the people who authored the foundational literature on AI in echocardiography — not by acquirers of someone else's research. That difference shows up in how our models are built, how they're tested, and how they behave at the bedside.
Dr. Ouyang is an attending cardiologist who reads echoes every week — currently at Kaiser Permanente Northern California and previously at Cedars-Sinai. Our models are designed against real workflow, not inferred from one.
Backed by Y Combinator and strategic partners including Alexion (AstraZeneca Rare Disease), with live deployments through partners like HeartLab.
Precision LVEF is trained on more echocardiographic videos than any other FDA-cleared echo AI model, giving it unmatched generalizability across scanners and populations.
Schedule a 30-minute briefing with our clinical and commercial teams. We'll walk through deployment, reimbursement, and the evidence base — tailored to your cardiovascular service line.