# Assessment of Point of Care Lung Ultrasound in the Ambulatory Heart Failure Setting

## TL;DR

This study shows that brief training allows non-experts to perform high-quality lung ultrasounds for heart failure patients, with automated tools improving consistency and detecting early congestion.

## Contribution

Demonstrates the feasibility of automated B-line counting in improving novice provider performance and detecting mild congestion in ambulatory heart failure.

## Key findings

- Novice providers achieved 88% good-quality lung ultrasounds after 2 hours of training.
- Automated B-line counting showed better consistency than manual counts by experts.
- Lung ultrasound detected mild volume overload not captured by clinical exams.

## Abstract

Limited studies have evaluated lung ultrasound (LUS) in ambulatory heart failure (HF). A six‐zone LUS assesses B‐lines, a marker of congestion. The Butterfly IQ+ probe features an automated B‐line counter (ABLC), eliminating manual counting. We evaluated LUS quality by novice HF providers after training, compared expert manual counts to ABLC, and explored associations between LUS and clinical HF metrics.

Three novice providers underwent 2 h of didactics and 30 proctored exams. Image quality was independently reviewed by two LUS experts. B‐lines were counted manually by experts and ABLC. We assessed associations between LUS and four clinical metrics: provider‐assessed volume status, > 30% NT‐proBNP increase, > 5 lb weight gain, and PAD above goal (CardioMEMS).

Seventy‐five subjects were enrolled. Overall, LUS quality was excellent, with 88% good quality. Surprisingly, agreement between expert B‐line counts was moderate (Gwet's AC1: 0.49, 95% CI: 0.27 to 0.71) while the accuracy of experts as compared to ABLC was modest (Expert 1 = 61.2%, Expert 2 = 40.3%). Experts correctly identified 93% of positive studies but only 19% of negative studies versus ABLC. Provider volume assessments substantially agreed with LUS (Gwet's AC1: 0.76, 95% CI: 0.61 to 0.91), but providers identified only half of positive LUS cases, suggesting utility in detecting mild volume overload. Only volume overload correlated with positive LUS. Only 25% of subjects had a CardioMEMS.

Novice providers can perform high‐quality LUS after brief training. ABLC reduces B‐line counting variability. LUS detects mild pulmonary congestion undetectable by clinical exam, potentially preventing worsening in HF patients.