Butterfly iQ/iQ+ Ultrasound System

The Butterfly iQ/iQ+ Ultrasound System is a portable diagnostic tool that now incorporates an 'Auto B-line Counter,' a software-based image analysis algorithm, to automate the detection and quantification of B-lines in lung ultrasound examinations. This system, with expanded indications for lung imaging and new Spectral Pulsed Wave Doppler capabilities, aims to provide objective, quantifiable data for various clinical applications, supported by performance data demonstrating non-inferiority to manual clinician assessment.

The Butterfly iQ/Butterfly iQ+ Ultrasound System is a hand-held, general-purpose diagnostic imaging tool designed for use by trained healthcare professionals. It comprises a single transducer that connects to a compatible commercial mobile device running the Butterfly iQ/iQ+ mobile application, enabling visualization and measurement of anatomical structures and fluids in both adult and pediatric patients. Key to this iteration is the introduction of the 'Auto B-line Counter,' a software feature utilizing an image analysis algorithm to automatically detect, count, and display B-lines in lung ultrasound images, specifically within a given rib space. The system's indications for use have been expanded to include 'Lung' imaging, and its operational modes now feature 'Spectral Pulsed Wave Doppler' alongside existing B-mode, M-mode, Color Doppler, and Power Doppler. The performance of the Auto B-line Counter was rigorously assessed through analytical and clinical validation. Analytical testing, using a large, diverse dataset of 6000 de-identified cines from 253 sites, demonstrated that the algorithm's performance in B-line detection and segmentation is non-inferior to clinician annotators, utilizing metrics such as the Intraclass Correlation Coefficient (ICC) and Dice Coefficient Score (DSC). This performance was consistent across various demographic subgroups and patients with common clinical confounders. Further clinical performance evaluation on 99 subjects confirmed the algorithm's non-inferiority to clinician ground truth for B-line counting and its generalizability across different patient ages, genders, and BMIs. Non-clinical testing also verified that the new software algorithm does not introduce additional risks to the ultrasound system.

AI Regulatory Concerns

Powered by

Jev reads the full FDA letter and returns calibrated concern probabilities per regulatory category. These are AI estimates, not regulatory advice.

• Q1. How does the Auto B-line Counter algorithm differentiate true B-lines from other similar-appearing lung artifacts (e.g., A-lines, Z-lines, or reverberation artifacts not originating from the pleura)?; What was the inter-observer variability (e.g., ICC) among the expert annotators who established the 'ground truth' for B-line counts and locations, and how does this compare to the algorithm's agreement with the consensus ground truth?; While the Auto B-line Counter does not mark images for specific pathologies, were there any analyses conducted on its performance (e.g., B-line counts) across different underlying lung conditions known to produce B-lines (e.g., pulmonary edema vs. interstitial pneumonia vs. ARDS) within the validation dataset?; What is the typical processing time for the Auto B-line Counter to analyze a lung cine and provide the B-line count and overlay, and how might this impact real-time clinical workflow?; Beyond the listed confounders (CHF, diabetes, etc.), were there specific ultrasound image quality factors (e.g., probe-skin contact, rib shadowing, patient movement) systematically evaluated for their impact on the Auto B-line Counter's accuracy?