AIM: To find occult metastases during hepatectomy in patients with colorectal cancer liver metastases (CRCLM), contrast-enhanced intraoperative ultrasonography (CE-IOUS) was performed using a new microbubble agent, so...AIM: To find occult metastases during hepatectomy in patients with colorectal cancer liver metastases (CRCLM), contrast-enhanced intraoperative ultrasonography (CE-IOUS) was performed using a new microbubble agent, sonazoid, which provides a parenchyma-specific contrast image based on its accumulation in the Kupffer cells. METHODS: Eight patients with CRCLM underwent CE- IOUS using sonazoid before hepatectomy. The liver was investigated during a late Kupffer-phase imaging, which is a valuable characteristic of sonazoid. RESULTS: CE-IOUS using sonazoid provided the early vascular- and sinusoidal-phase images for 10 min followed by the late Kupffer-phase image up to 30 min after the injection of sonazoid. IOUS did not provide new findings of metastatic lesion in the 8 patients. However, during the late Kupffer-phase image of sonazoid, a metastatic lesion was newly found in two of the 8 patients. These newly detected lesions were removed by an additional hepatectomy and histopathologically diagnosed as a metastasis. CONCLUSION: CE-IOUS using sonazoid can allow surgeons to investigate the whole liver with enough time and to find new metastases intraoperatively.展开更多
目的:评估基于注射用全氟丁烷微球[商品名示卓安(Sonazoid)]超声造影Kupffer期的深度学习模型预测肝细胞癌(hepatocellular carcinoma,HCC)微血管侵犯(microvascular invasion,MVI)的效能,并将其与影像组学模型及临床模型进行比较。方法...目的:评估基于注射用全氟丁烷微球[商品名示卓安(Sonazoid)]超声造影Kupffer期的深度学习模型预测肝细胞癌(hepatocellular carcinoma,HCC)微血管侵犯(microvascular invasion,MVI)的效能,并将其与影像组学模型及临床模型进行比较。方法:回顾并纳入2020年7月—2022年9月于广西医科大学第一附属医院接受Sonazoid超声造影检查的146例原发性HCC患者,以7∶3随机划分为训练集102例和验证集44例。基于肿瘤感兴趣区,使用ResNet101模型通过迁移学习提取深度学习特征,使用PyRadiomics提取影像组学特征。采用Mann-Whitney U检验、最小绝对收缩和选择算子(least absolute shrinkage and selection operator,LASSO)算法进行特征降维。LASSO回归用于构建深度学习模型和影像组学模型,同时还基于临床特征构建一个临床模型。采用受试者工作特征曲线的曲线下面积(area under the curve,AUC)、灵敏度、特异度和准确度评估模型的诊断效能。DeLong检验用于比较模型间的诊断效能。结果:在训练集中,深度学习模型、影像组学模型、临床模型的AUC(95%CI)分别为0.931(0.880~0.981)、0.823(0.744~0.903)、0.719(0.614~0.824)。在验证集中,深度学习模型、影像组学模型、临床模型的AUC(95%CI)分别为0.895(0.757~1.000)、0.711(0.514~0.909)、0.606(0.390~0.822)。DeLong检验表明在训练集和验证集中,深度学习模型的诊断效能均优于影像组学模型及临床模型(P<0.05)。单因素及多因素logistic回归分析示甲胎蛋白和巴塞罗那临床肝癌分期可作为HCC患者MVI的独立预测因子(P<0.01)。结论:基于Sonazoid超声造影Kupffer期的深度学习模型在预测HCC患者MVI方面表现出优异的性能,有望成为预测MVI的无创影像学生物标志物。展开更多
文摘AIM: To find occult metastases during hepatectomy in patients with colorectal cancer liver metastases (CRCLM), contrast-enhanced intraoperative ultrasonography (CE-IOUS) was performed using a new microbubble agent, sonazoid, which provides a parenchyma-specific contrast image based on its accumulation in the Kupffer cells. METHODS: Eight patients with CRCLM underwent CE- IOUS using sonazoid before hepatectomy. The liver was investigated during a late Kupffer-phase imaging, which is a valuable characteristic of sonazoid. RESULTS: CE-IOUS using sonazoid provided the early vascular- and sinusoidal-phase images for 10 min followed by the late Kupffer-phase image up to 30 min after the injection of sonazoid. IOUS did not provide new findings of metastatic lesion in the 8 patients. However, during the late Kupffer-phase image of sonazoid, a metastatic lesion was newly found in two of the 8 patients. These newly detected lesions were removed by an additional hepatectomy and histopathologically diagnosed as a metastasis. CONCLUSION: CE-IOUS using sonazoid can allow surgeons to investigate the whole liver with enough time and to find new metastases intraoperatively.
文摘目的:评估基于注射用全氟丁烷微球[商品名示卓安(Sonazoid)]超声造影Kupffer期的深度学习模型预测肝细胞癌(hepatocellular carcinoma,HCC)微血管侵犯(microvascular invasion,MVI)的效能,并将其与影像组学模型及临床模型进行比较。方法:回顾并纳入2020年7月—2022年9月于广西医科大学第一附属医院接受Sonazoid超声造影检查的146例原发性HCC患者,以7∶3随机划分为训练集102例和验证集44例。基于肿瘤感兴趣区,使用ResNet101模型通过迁移学习提取深度学习特征,使用PyRadiomics提取影像组学特征。采用Mann-Whitney U检验、最小绝对收缩和选择算子(least absolute shrinkage and selection operator,LASSO)算法进行特征降维。LASSO回归用于构建深度学习模型和影像组学模型,同时还基于临床特征构建一个临床模型。采用受试者工作特征曲线的曲线下面积(area under the curve,AUC)、灵敏度、特异度和准确度评估模型的诊断效能。DeLong检验用于比较模型间的诊断效能。结果:在训练集中,深度学习模型、影像组学模型、临床模型的AUC(95%CI)分别为0.931(0.880~0.981)、0.823(0.744~0.903)、0.719(0.614~0.824)。在验证集中,深度学习模型、影像组学模型、临床模型的AUC(95%CI)分别为0.895(0.757~1.000)、0.711(0.514~0.909)、0.606(0.390~0.822)。DeLong检验表明在训练集和验证集中,深度学习模型的诊断效能均优于影像组学模型及临床模型(P<0.05)。单因素及多因素logistic回归分析示甲胎蛋白和巴塞罗那临床肝癌分期可作为HCC患者MVI的独立预测因子(P<0.01)。结论:基于Sonazoid超声造影Kupffer期的深度学习模型在预测HCC患者MVI方面表现出优异的性能,有望成为预测MVI的无创影像学生物标志物。
