结合胃镜超声和白光内镜可以更准确地识别胃肠道间质瘤.但是现有的多模态方法往往仅关注于图像特征,忽略了诊断文本信息中所包含的语义信息对于精确理解和诊断医学图像的重要性.为此,本文提出一种新的基于文本引导下的多模态医学图像分...结合胃镜超声和白光内镜可以更准确地识别胃肠道间质瘤.但是现有的多模态方法往往仅关注于图像特征,忽略了诊断文本信息中所包含的语义信息对于精确理解和诊断医学图像的重要性.为此,本文提出一种新的基于文本引导下的多模态医学图像分析算法框架(Text-guided Multi-modal Medical image analysis framework,TMM-Net).TMM-Net使用多阶段的诊断文本来引导模型学习,以提取图像中的关键诊断信息特征,然后通过交叉模态注意力机制促进多模态特征之间的交互.值得注意的是,TMM-Net通过预测病变属性来模拟临床诊断过程,从而增强了可解释性.验证实验在两个中心包含10 025个模态数据对的数据集上进行.结果表明,该方法相比目前最优的GISTs诊断方法精度提升7.7%,同时获得了最高的(Area Under the Curve,AUC)值:0.927,其可解释性可以更好地适合临床需求.展开更多
The clinical outcome of patients receiving liver trans-plantation could be significantly affected by biliary complications, including strictures, leaks, stones and bilomas; early diagnosis and treatment of these con-d...The clinical outcome of patients receiving liver trans-plantation could be significantly affected by biliary complications, including strictures, leaks, stones and bilomas; early diagnosis and treatment of these con-ditions lead to markedly reduction in morbidity and mortality. Therapeutic gold standard is represented by conservative approaches, both endoscopic and percutaneous, based on the type of biliary reconstruction, the local availability of the procedures and specific expertise. In patients with previous transplantation, the difficult biliary access and the possible presence of concomitant complications(mainly strictures) further restrict the efficacy of the endoscopic and percutaneous treatments; on the other hand, surgery should generally be avoided because of the even increased morbidity and mortality due to technical and clinical issues. Here we review the most common biliary complications occurring after liver transplantation and discuss available treatment options including future perspectives such as endoscopic ultrasound-guided biliary access in patients with Roux-en-Y choledocho-jejunostomy or extracorporeal shock wave lithotripsy for difficult stones.展开更多
文摘结合胃镜超声和白光内镜可以更准确地识别胃肠道间质瘤.但是现有的多模态方法往往仅关注于图像特征,忽略了诊断文本信息中所包含的语义信息对于精确理解和诊断医学图像的重要性.为此,本文提出一种新的基于文本引导下的多模态医学图像分析算法框架(Text-guided Multi-modal Medical image analysis framework,TMM-Net).TMM-Net使用多阶段的诊断文本来引导模型学习,以提取图像中的关键诊断信息特征,然后通过交叉模态注意力机制促进多模态特征之间的交互.值得注意的是,TMM-Net通过预测病变属性来模拟临床诊断过程,从而增强了可解释性.验证实验在两个中心包含10 025个模态数据对的数据集上进行.结果表明,该方法相比目前最优的GISTs诊断方法精度提升7.7%,同时获得了最高的(Area Under the Curve,AUC)值:0.927,其可解释性可以更好地适合临床需求.
文摘The clinical outcome of patients receiving liver trans-plantation could be significantly affected by biliary complications, including strictures, leaks, stones and bilomas; early diagnosis and treatment of these con-ditions lead to markedly reduction in morbidity and mortality. Therapeutic gold standard is represented by conservative approaches, both endoscopic and percutaneous, based on the type of biliary reconstruction, the local availability of the procedures and specific expertise. In patients with previous transplantation, the difficult biliary access and the possible presence of concomitant complications(mainly strictures) further restrict the efficacy of the endoscopic and percutaneous treatments; on the other hand, surgery should generally be avoided because of the even increased morbidity and mortality due to technical and clinical issues. Here we review the most common biliary complications occurring after liver transplantation and discuss available treatment options including future perspectives such as endoscopic ultrasound-guided biliary access in patients with Roux-en-Y choledocho-jejunostomy or extracorporeal shock wave lithotripsy for difficult stones.
