多模态超声和基于超声深度学习的影像组学在预测乳腺癌患者新辅助化疗疗效中的研究进展

Research progress of multimodal ultrasound and imageomics based on ultrasound deep learning in predicting the efficacy of neoadjuvant chemotherapy in breast cancer patients

随着乳腺癌在女性恶性肿瘤的发病率和死亡率位居世界第一,乳腺癌也已经进入多种模式相结合的综合治疗时代。而新辅助化疗(NAC)是综合治疗的重要组成部分,能够增加手术机会、提高保乳率、改善患者预后。超声凭借多样化、安全无辐射及可重复性强等多种优势,已成为了早期预测和评估NAC疗效的一种重要的影像学检查手段。目前主要的多模态超声技术包括常规超声、彩色多普勒超声、超声弹性成像、超声造影技术等,而近年来基于卷积神经网络的深度学习影像组学也成为了研究的热点,可以帮助医生更高效、精确地对乳腺癌患者NAC疗效进行预测。本文基于先前文献对应用于预测乳腺癌患者NAC疗效的多模态超声成像技术和基于超声深度学习的影像组学进行综述。

With the incidence rate and mortality of breast cancer in female malignant tumors ranking first in the world,breast cancer has also entered the era of comprehensive treatment combining multiple modes.Neoadjuvant chemotherapy(NAC)is an important component of comprehensive treatment,which can increase surgical opportunities,improve breast retention rate,and improve patient prognosis.Ultrasound has become an important imaging examination method for early prediction and evaluation of the efficacy of NAC,with various advantages such as diversity,safety,non radiation,and strong repeatability.At present,the main multimodal ultrasound technologies include conventional ultrasound,color Doppler ultrasound,ultrasonic elastog-raphy,contrast-enhanced ultrasound,etc.In recent years,the deep learning imageomics based on convolutional neural network has also become a research hot spot,which can help doctors predict the efficacy of NAC in breast cancer patients more efficiently and accurately.This article reviews multimodal ultrasound imaging technology and imageomics based on ultrasonic deep learning for predicting NAC efficacy in breast cancer patients based on previous literature.