Department of Medical Ultrasonics, Division of Interventional Ultrasound, The First Affiliated Hospital of Sun Yat-sen University, Guangzhou, Guangdong, China.
Institute for Brain and Cognitive Sciences, BRNist, Tsinghua University, Beijing, China.
Liver Int. 2022 Jun;42(6):1423-1431. doi: 10.1111/liv.15254. Epub 2022 May 5.
Radiological prediction of microvascular invasion (MVI) of hepatocellular carcinoma (HCC) is essential but few models were clinically implemented because of limited interpretability and generalizability.
Based on 2096 patients in three independent HCC cohorts, we established and validated an MVI predicting model. First, we used data from the primary cohort to train a 3D-ResNet network for MVI prediction and then optimised the model with "expert-inspired training" for model construction. Second, we implemented the model to the other two cohorts using three implementation strategies, the original model implementation, data sharing model implementation and skeleton sharing model implementation, the latter two of which used part of the cohorts' data for fine-tuning. The areas under the receiver operating characteristic curve (AUCs) were calculated to compare the performances of different models.
For the MVI predicting model, the AUC of the expert-inspired model was 0.83 (95% CI: 0.77-0.88) compared to 0.54 (95% CI: 0.46-0.62) of model before expert-inspiring. Taking this model as an original model, AUC on the second cohort was 0.76 (95% CI: 0.67-0.84). The AUC was improved to 0.83 (95% CI: 0.77-0.90) with the data-sharing model, and further improved to 0.85 (95% CI: 0.79-0.92) with the skeleton sharing model. The trend that the skeleton sharing model had an advantage in performance was similar in the third cohort.
We established an expert-inspired model with better predictive performance and interpretability than the traditional constructed model. Skeleton sharing process is superior to data sharing and direct model implementation in model implementation.
肝癌（HCC）微血管侵犯（MVI）的影像学预测至关重要，但由于可解释性和通用性有限，很少有模型在临床上得到实施。
基于三个独立 HCC 队列中的 2096 名患者，我们建立并验证了一种 MVI 预测模型。首先，我们使用主要队列的数据来训练 3D-ResNet 网络进行 MVI 预测，然后使用“专家启发式训练”对模型进行优化以构建模型。其次，我们使用三种实施策略将模型应用于另外两个队列，即原始模型实施、数据共享模型实施和骨架共享模型实施，后两者使用部分队列的数据进行微调。计算受试者工作特征曲线下的面积（AUC）以比较不同模型的性能。
对于 MVI 预测模型，专家启发式模型的 AUC 为 0.83（95%CI：0.77-0.88），而专家启发前模型为 0.54（95%CI：0.46-0.62）。以该模型作为原始模型，第二个队列的 AUC 为 0.76（95%CI：0.67-0.84）。使用数据共享模型，AUC 提高到 0.83（95%CI：0.77-0.90），使用骨架共享模型进一步提高到 0.85（95%CI：0.79-0.92）。骨架共享模型在性能上具有优势的趋势在第三个队列中相似。
我们建立了一个具有更好预测性能和可解释性的专家启发模型，优于传统构建的模型。在模型实施中，骨架共享过程优于数据共享和直接模型实施。
