| task_categories: | |
| - feature-extraction | |
| This is the official dataset(HCC) of "Interpretable 2.5D Network by Hierarchical Attention and Consistency Learning for 3D MRI Classification" . | |
| **Detailed Information** | |
| HCC: which is a private dataset collected from the West China Hospital, | |
| and is a prospective trial aim to validate the effectiveness of MRI images for | |
| predicting Microvascular Invasion(MVI) in patients with Hepatocellular Carci- | |
| noma. This dataset totally consists of 246 patients, including 110 patients with | |
| MVI and 136 ones without MVI. | |
| **Only HBP Modality** | |
| HBP is an MRI sequence based on hepatobiliary-specific contrast agents, which is the most advantageous sequence for detecting and evaluating intrahepatic lesions. | |
| Therefore, in our current work, we only use HBP modality and have currently anonymized and disclosed only the HBP modality. | |
| **How to use data?** | |
| 1. Download 'img.zip' | |
| 2. read '{patient_number}_{label}.nii.gz' and generate labels | |
| **Future** | |
| The MRI images of the other modalities(such as PVP, AP and so on) will be continuously updated... | |