AI News Today.

Diagnostics

University of Notre Dame Australia (PANACEA-HF program) · Us2.ai

Vendor-reportedHealthcareAustraliaPilotPublic sectorComputer vision

Vendor-reported. The customer is named and the numbers are quoted from the source page, but the account comes from the vendor. No independent confirmation.

Reported by us2.ai (vendor self-report). Checked against the source page on 2026-09-06. 10 of 10 figures below appear on that page word for word.

The problem

Heart failure most often goes undetected until a patient is acutely hospitalized, a costly and avoidable outcome. In primary care, even among high-risk individuals, the syndrome remains chronically missed.

What was deployed

Vendor
Us2.ai
Products
Us2.ai
Technique
Computer vision
Build or buy
Bought off the shelf
Deployment
SaaS
Data used
echocardiographic images, ESC-defined signs and symptoms, 12-lead ECG findings, NT-proBNP levels
Scale
multi-center surveillance study across metropolitan and rural-remote primary care clinics in Australia; 300 formally screened cases to date (150 men, 150 women, mean age ~71); full results from 700+ patients expected August 2026

What changed

Each row is quoted from the source. Figures we could not find on the page in those words are marked — they are kept, not deleted, so you can judge them.

zeroImaging failure ratesfrom 16%from the first 50 cases to the final 50 (within 300 formally screened cases)

“Imaging failure rates dropped from 16% in the first 50 cases to zero in the final 50”

Quoted word for word from the source

68%Rate of complete AI-PoCUS reports generating more than 30 cardiac indicesfrom 26%from the first 50 cases to the final 50 (within 300 formally screened cases)

“the rate of complete AI-PoCUS reports generating more than 30 cardiac indices rose from 26% to 68%”

Quoted word for word from the source

6.30-fold more likelyLikelihood of full report generationby cases 251 to 300 versus the first 50

“Full report generation was 6.30-fold more likely by cases 251 to 300 versus the first 50 (p<0.001)”

Quoted word for word from the source

100%Sensitivity of GP-led referral on clinical groundson an initial cohort of 100 patients

“GP-led referral on clinical grounds achieved 100% sensitivity”

Quoted word for word from the source

88.8%Specificity of GP-led referral on clinical groundson an initial cohort of 100 patients

“but only 88.8% specificity”

Quoted word for word from the source

nineUnnecessary specialist referrals by GP-led referralon an initial cohort of 100 patients

“generating nine unnecessary specialist referrals”

Quoted word for word from the source

96.7%Sensitivity of the clinical algorithmon an initial cohort of 100 patients

“The algorithm achieved 96.7% sensitivity”

Quoted word for word from the source

98.8%Specificity of the clinical algorithmon an initial cohort of 100 patients

“and 98.8% specificity”

Quoted word for word from the source

just oneFalse-positive referrals by the clinical algorithmon an initial cohort of 100 patients

“with just one false-positive referral”

Quoted word for word from the source

77.3Positive likelihood ratio of the clinical algorithmon an initial cohort of 100 patients

“and a positive likelihood ratio of 77.3 (95% CI 11.0 to 543)”

Quoted word for word from the source

Sources

Similar deployments

HappyPath · Oodle

30 minutes — Render integration setup time

Vendor-reportedTechnology & software · Diagnostics · LLM · In production

MUSC Health · Notable

110000 — monthly digital registrations completed

Vendor-reportedHealthcare · Process automation · Classical ML · Scaled · 2022