RESEARCH PREVIEW — NOT FOR CLINICAL USE. EchoReview AI is an investigational workflow tool. Output is a preliminary, non-diagnostic impression intended to support cardiologist review for research and quality-improvement purposes only. Not FDA-cleared. Do not use for patient care decisions. Do not upload identified PHI without an executed BAA.
Clinical Advisory Network

Help build the echo review workflow you actually want.

EchoReview AI is a research-preview platform that pairs rural hospital echo acquisition with remote cardiologist review. We are recruiting board-certified cardiologists and Level III echo readers to shape the protocol, validate the AI triage layer, and co-author the concordance study.

Low-commitment 90-day shadow pilot. ~2 hours/week of research-grade reads. No clinical liability — every AI impression shadows your read, never replaces it.

Research preview. Not FDA-cleared. Not for primary diagnosis.

What advisors get

  • Honorarium per reviewed case (research-grade reads, not billable clinical reads)
  • Co-authorship on the concordance / time-to-read publication
  • Equity consideration for founding clinical advisors
  • Direct line to engineering — your workflow feedback ships
  • Early access to DICOM ingest, structured reporting, and PACS integration

Ready to start reviewing? Sign the Clinical Advisor Agreement online.

Why we built this

Rural and critical-access hospitals often wait 48–96 hours for an echo read. Patients get transferred for studies that turn out to be normal. Cardiologists get faxed PDFs with no images attached. EchoReview AI is the workflow we wished existed during call.

Faster triage

AI flags suspected severe AS, reduced EF, pericardial effusion, and RV strain so urgent reads jump the queue.

Immutable sign-off

Password-re-authenticated final read. Studies lock on signature. Full audit trail. PDF export.

Built for partnerships

Per-organization isolation, audit logs, 2FA, idle timeout, and a compliance page that tells partners exactly what is and isn't in place.

Liability, team, and scope

No team required. Scope-limited. Not the medico-legal record.

We built the pilot so an individual cardiologist can participate safely and within their comfort zone. You choose whether to review cases from your own institution, or to advise on de-identified cases from other sites.

  • Site-based shadow pilot: your signed research report is for concordance/QI only — the official clinical read remains the medico-legal record
  • External advisor model: review de-identified cases with no patient-care liability and no official signature
  • Individual cardiologists can sign the DUA as a single PI; no team or co-investigator is required
  • Scope-limited agreements: opt into pediatric-only, adult congenital-only, or full adult acquired disease — and change it anytime

What you are — and aren't — signing

During the pilot, the EchoReview AI report is a research-grade shadow read. It is never inserted into the EHR or used for patient care. For site-based pilots, your institution's existing signed clinical report stays the legal record. For external advisors, there is no clinical signature at all.

Scope example

"I will review pediatric echocardiograms and adult congenital cases only. Please do not route general adult acquired valve disease or adult TTEs to my queue."

Handling subjectivity

AI makes subjectivity visible, not invisible.

Echocardiography is interpretive. EchoReview AI does not claim to eliminate that — it exposes it. Every output carries a confidence score and severity label, and every cardiologist override is logged.

  • Confidence score (0–1) on every impression so low-confidence cases get more attention
  • Cardiologist can override, edit, or discard the AI impression; signed report is always theirs
  • Immutable audit trail of who signed, when, and what changed
  • Concordance reporting: normal vs. abnormal triage, severity agreement, high-risk miss rate

How calibration works in the pilot

During the 90-day shadow pilot, the same cases can be reviewed by two cardiologists when feasible. AI impressions are compared to both reads, and disagreements are discussed in weekly 30-minute calibration calls. The model prompt is tuned based on your feedback — your data is not used to train it.

Research question we answer together

"Does the AI triage layer correctly identify urgent cases, flag low-confidence reads, and route them to the right cardiologist faster than the current fax/phone workflow?"

What the AI triage layer looks for

The current research-preview model returns a structured impression across the pathologies below. Each is available as an interactive example study in the demo dashboard — click any into the reviewer to see the AI output alongside the image.

Normal baseline (PLAX)Normal
Normal baseline (PSAX)Normal
Normal apical four-chamberNormal
Aortic stenosis (suspected severe)Severe
Mitral regurgitation (color Doppler)Moderate
Mitral valve prolapseMild–Moderate
Tricuspid regurgitationModerate
Pericardial effusionModerate
LV hypertrophy / HCM patternModerate–Severe
Dilated cardiomyopathy patternSevere
Regional wall-motion abnormality (post-MI)Moderate
LV apical thrombus (post-STEMI)Severe
RV dilation with pulmonary hypertensionSevere
Atrial septal defectModerate
Bicuspid aortic valveMild
Suspected endocarditis / vegetationSevere

Coverage list reflects the current research-preview scope; findings are non-diagnostic and always subject to cardiologist review. Additional views (strain, 3D, stress) are on the roadmap and prioritized by advisor feedback.

Pediatric cardiology

A separate scope for congenital heart disease.

Pediatric echocardiography is a distinct discipline. Neonatal anatomy, shunt physiology, and normal reference ranges differ materially from adult acquired disease. We are honest about the boundary.

  • Adult TTE is the current primary focus: acquired valve disease, LV/RV function, pericardial disease
  • Pediatric TTE is a research-preview expansion with 12 demo cases across common congenital lesions
  • Same guardrails apply: AI triage + pediatric cardiologist final read, never AI-only

Pediatric demo cases in the dashboard

Normal neonate / infant
Patent ductus arteriosus (PDA)
Secundum ASD
Perimembranous VSD
Tetralogy of Fallot
Coarctation of the aorta
d-Transposition of the great arteries
Hypoplastic left heart syndrome
Complete AV canal defect
Valvular pulmonary stenosis
Ebstein anomaly
Kawasaki coronary aneurysm surveillance

A pediatric-specific validation track would add z-score integration, congenital-specific endpoints, and a dedicated pediatric cardiologist review queue.

The reading workflow

  1. 01

    Sonographer uploads

    DICOM or image bulk-upload. Patient metadata auto-parses from DICOM tags.

  2. 02

    AI preliminary impression

    Structured triage: suspected finding, severity, confidence, key observations.

  3. 03

    Cardiologist review

    Dual-pane viewer. You write the impression, recommendations, and severity.

  4. 04

    Sign & lock

    Password re-auth, study locks, PDF report exports, addendums tracked.

Who we're looking for

Founding clinical advisors

  • Board-certified cardiologist (NBE Level II or III strongly preferred)
  • Active US medical license in at least one state
  • Experience reading TTE at volume — academic, private practice, or tele-cardiology
  • Interest in rural access, AI-assisted workflows, or quality-improvement research
  • Willing to commit ~2 hours per week during the 90-day shadow pilot
What we're honest about

No clinical use yet

EchoReview AI is a research preview. The AI impression is not FDA-cleared and is never the final read. During the pilot, all reads are research-grade and shadow the on-site clinical workflow. We will not ask you to sign anything that influences patient care until clearance and BAAs are executed. See our compliance page for the full posture.

Apply to advise

Send a short note with your background, NBE level, state license(s), and what you'd want to see in an echo triage tool. We respond to every applicant within 5 business days.

Honoraria, co-authorship, and equity consideration discussed individually. No PHI in the application, please.

Refer a cardiologist

Know a colleague who'd like this?

The fastest way we grow is cardiologist-to-cardiologist. Point us to someone in your network and we'll reach out with a short, respectful intro — always mentioning you.

About you

About your colleague

We route cases to match each cardiologist's declared scope.

We'll mention you in the outreach. No spam — one thoughtful email.