Breast ultrasound diagnosis improved with concept-informed AI system

A Concept-informed Interpretable Artificial Intelligence Framework for Breast Ultrasound Diagnosis

Computer Vision and Pattern Recognition

Summary

Breast ultrasound scans are important for spotting cancer, but computers often only look at images without understanding medical concepts doctors use. The authors created a large collection of breast ultrasound images with detailed medical notes and built an AI system that mimics how doctors think by using this knowledge. This helps the AI explain its decisions better, catch risky cases more accurately, and assist doctors in making better diagnoses. Tests show the AI works well across different hospitals and helps both new and experienced doctors improve sensitivity in finding cancer.

What this means in practice

  • For hospital radiology teams: Use the AI to provide interpretable risk assessments alongside ultrasound images, improving diagnosis accuracy and consistency across clinics.
  • For medical device developers: Integrate the framework into ultrasound machines or software to offer enhanced diagnostic support with concept-informed explanations for clinicians.$Commercial implications: Enables creation of AI-enhanced ultrasound diagnostic tools that can be sold to hospitals for improved breast cancer detection.

Authors

Shengzhou Zhong, Ziying Wang, Liang Li, Lihua Yan, Shuochun Chen, Yudie Zou, Jifan Zhang, Wenhao Li, Huiling Zhang, Jiayin Lu, Ziting Xu, Qianjin Feng, Zhenwei Shi, Yan Chen, Yingjia Li, Yu Zhang, Zhenyuan Ning

Abstract

Most existing artificial intelligence systems for breast ultrasound rely primarily on image-driven prediction, with limited incorporation of the knowledge from the Breast Imaging Reporting and Data System (BI-RADS) and the intermediate assessment processes underlying radiologists' workflow, thereby constraining their interpretability and alignment with clinical practice. We curated, to our knowledge, the largest multicenter collection of breast ultrasound images paired with structured BI-RADS descriptors, comprising 9,941 patients across 20 independent datasets. Across 14 testing datasets, IBUS demonstrated consistent generalization performance across different clinical centers and imaging conditions. Incorporating BI-RADS descriptors enabled fine-grained risk stratification, while cross-checking predicted BI-RADS grades against malignancy probabilities identified internally discordant predictions that may warrant additional clinical review. In a reader study involving eight radiologists, IBUS assistance increased diagnostic sensitivity among both junior and senior readers, with the potential of guiding radiologists to make targeted revision in risk stratification. Overall, IBUS exhibits its potential as a second-stage decision-support system that complements radiologists by providing additional risk estimates and interpretable evidence for diagnostically challenging cases. The demo video of IBUS platform is publicly available at https://github.com/zsz1997/IBUS.