人工智能在乳腺癌超声筛查中的应用进展

Artificial intelligence advancements in breast cancer ultrasound screening

超声是乳腺癌筛查的一线检查方法,BI-RADS的应用使超声对乳腺疾病的诊断具有相对一致性,但仍不同程度受操作者主观因素影响。随着计算机技术的发展和大数据时代的到来,人工智能乳腺超声从基于二维静态图像分析、到动态捕获病灶和关键帧分析、到全自动乳腺容积扫查和多模态研究,基于深度学习建立的计算机辅助诊断模型性能不断提升和完善。人工智能超声的应用可辅助超声医师提高诊断的准确性和一致性,在乳腺癌的筛查和诊疗过程中具有重要价值。

Ultrasound is the first-line screening method for breast cancer.The application of BI-RADS makes the ultrasound diagnosis of breast diseases relatively consistent,but it is still affected by subjective factors of operators.With the advancement of computer technology and the arrival of the era of big data,the performance of computer-aided diagnostic models based on deep learning has been continuously improved from two-dimensional static image analysis to dynamic capture of lesions and keyframe analysis to automatic breast volume scanning and multi-modal research.The use of artificial intelligence in ultrasound can help sonographers increase the consistency and accuracy of diagnosis,and play an important role in the process of breast cancer screening,diagnosis and treatment.