Department of Radiology, Jiangmen Central Hospital, 23 Beijie Haibang Street, Jiangmen, People's Republic of China.
Zunyi Medical University, 1 Xiaoyuan Road, Zunyi, People's Republic of China.
Abdom Radiol (NY). 2024 May;49(5):1397-1410. doi: 10.1007/s00261-024-04202-1. Epub 2024 Mar 3.
To investigate the value of a multimodal deep learning (MDL) model based on computed tomography (CT) and magnetic resonance imaging (MRI) for predicting microvascular invasion (MVI) in hepatocellular carcinoma (HCC).
A total of 287 patients with HCC from our institution and 58 patients from another individual institution were included. Among these, 119 patients with only CT data and 116 patients with only MRI data were selected for single-modality deep learning model development, after which select parameters were migrated for MDL model development with transfer learning (TL). In addition, 110 patients with simultaneous CT and MRI data were divided into a training cohort (n = 66) and a validation cohort (n = 44). We input the features extracted from DenseNet121 into an extreme learning machine (ELM) classifier to construct a classification model.
The area under the curve (AUC) of the MDL model was 0.844, which was superior to that of the single-phase CT (AUC = 0.706-0.776, P < 0.05), single-sequence MRI (AUC = 0.706-0.717, P < 0.05), single-modality DL model (AUC = 0.722, AUC = 0.731; P < 0.05), clinical (AUC = 0.648, P < 0.05), but not to that of the delay phase (DP) and in-phase (IP) MRI and portal venous phase (PVP) CT models. The MDL model achieved better performance than models described above (P < 0.05). When combined with clinical features, the AUC of the MDL model increased from 0.844 to 0.871. A nomogram, combining deep learning signatures (DLS) and clinical indicators for MDL models, demonstrated a greater overall net gain than the MDL models (P < 0.05).
The MDL model is a valuable noninvasive technique for preoperatively predicting MVI in HCC.
研究基于计算机断层扫描（CT）和磁共振成像（MRI）的多模态深度学习（MDL）模型预测肝细胞癌（HCC）微血管侵犯（MVI）的价值。
本研究纳入了来自我们机构的 287 例 HCC 患者和另一机构的 58 例患者。其中，119 例患者仅具有 CT 数据，116 例患者仅具有 MRI 数据，用于开发单模态深度学习模型，然后使用迁移学习（TL）迁移选择参数以开发 MDL 模型。此外，110 例同时具有 CT 和 MRI 数据的患者被分为训练队列（n=66）和验证队列（n=44）。我们将从 DenseNet121 中提取的特征输入到极端学习机（ELM）分类器中，以构建分类模型。
MDL 模型的曲线下面积（AUC）为 0.844，优于单期 CT（AUC=0.706-0.776，P<0.05）、单序列 MRI（AUC=0.706-0.717，P<0.05）、单模态 DL 模型（AUC=0.722，AUC=0.731；P<0.05）和临床模型（AUC=0.648，P<0.05），但不如延迟期（DP）和同相位（IP）MRI 以及门静脉期（PVP）CT 模型。MDL 模型的性能优于上述模型（P<0.05）。当与临床特征相结合时，MDL 模型的 AUC 从 0.844 增加到 0.871。结合深度学习特征（DLS）和 MDL 模型的临床指标的列线图，与 MDL 模型相比具有更大的总体净收益（P<0.05）。
MDL 模型是一种有价值的术前预测 HCC 微血管侵犯的非侵入性技术。
