A non-invasive end-to-end intelligent assistance system for breast ultrasound

2026-05-20 · Source: Machine learning : nature.com subject feeds · Field: Health & Wellbeing — Medical Devices & Health Technology, Clinical Care & Medical Practice, Health & Medical Research · Depth: Expert, medium

Summary

An end-to-end breast intelligent recognition device (BIRD) has been developed and tested to address generalization challenges in breast ultrasound diagnosis. BIRD achieved an internal test set accuracy of 0.837 (95% confidence interval: 0.827–0.846), significantly improving radiologists' accuracy in two reader studies (*P* < 0.05). The system was applied in breast cancer screening for 6,817 individuals across 107 hospitals, demonstrating high consistency with clinical assessments (Cohen's kappa: 0.702 (95% confidence interval: 0.628-0.777)). Additionally, pathological and molecular subtype models, trained with data from five hospitals, showed satisfactory classification performance. These findings confirm BIRD's capability to enhance diagnostic accuracy, aid screening, and characterize breast lesions, supporting its clinical adoption.

Key takeaway

For healthcare administrators evaluating AI diagnostic tools, BIRD presents a validated solution for breast ultrasound. Its demonstrated accuracy of 0.837 and high consistency (Cohen's kappa: 0.702) across 107 hospitals suggest it can significantly enhance diagnostic efficiency and screening programs. Consider piloting BIRD or similar end-to-end AI systems to improve patient outcomes and streamline breast health practices in your institution.

Key insights

BIRD significantly improves breast ultrasound diagnostic accuracy and screening consistency across diverse real-world clinical settings.

Principles

• AI can overcome generalization challenges in medical imaging.
• Real-world deployment validates AI diagnostic tools.
• Pathological and molecular models enhance lesion characterization.

Method

The BIRD system was developed and tested end-to-end across diverse institution/population datasets, then applied in breast cancer screening for 6,817 individuals across 107 hospitals.

In practice

• Integrate BIRD for improved diagnostic accuracy.
• Utilize BIRD for large-scale breast cancer screening.
• Employ AI for pathological and molecular subtyping.

Topics

• Breast Ultrasound
• Artificial Intelligence
• Medical Diagnostics
• Breast Cancer Screening
• Clinical Validation
• Image Classification

Best for: Computer Vision Engineer, AI Scientist, Research Scientist, Domain Expert

Related on AIssential

• How AI Ultrasound Systems Are Improving Real-Time Diagnostic Accuracy in Clinical Settings — AI ultrasound systems enhance diagnostic accuracy and expand access by automating measurements, guiding image acquisition, and accelerating interpretation.
• A Chain-of-thought Reasoning Breast Ultrasound Dataset Covering All Histopathology Categories — BUS-CoT is a large, richly annotated breast ultrasound dataset enabling AI development for chain-of-thought reasoning across all histopathology types.
• Improving breast cancer screening workflows with machine learning — Google Research, in partnership with NHS organizations, conducted the Artificial Intelligence in Mammography Screening (AIMS) study, publishing two companion studies in Nature Cancer to evaluate an AI…
• Female-RHINO: A Real-Time Scanner-Integrated Framework for Automated Quantitative Uterine MRI Analysis and Structured Reporting — Female-RHINO provides real-time, AI-driven quantitative uterine MRI analysis and structured reporting integrated directly into the scanner workflow.
• Implementing trust in non-small cell lung cancer diagnosis with a conformalized uncertainty-aware AI framework — TRUECAM enhances pathology AI trustworthiness by integrating uncertainty quantification, data filtering, and error rate control.
• DreamReg: Belief-Driven World Model for 2D-3D Ultrasound Registration — DreamReg uses a belief-driven world model and internal imagination for robust 2D-3D ultrasound registration, overcoming real-time challenges.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by Machine learning : nature.com subject feeds.