4D Deep Learning for Multiple Sclerosis Lesion Activity Segmentation
This work addresses lesion activity detection in multiple sclerosis patients, offering an incremental improvement over existing methods.
The paper tackled the problem of multiple sclerosis lesion activity segmentation by extending deep learning to full 4D using a history of MRI volumes, resulting in a lesion-wise true positive rate of 0.84 at a false positive rate of 0.19, outperforming previous approaches.
Multiple sclerosis lesion activity segmentation is the task of detecting new and enlarging lesions that appeared between a baseline and a follow-up brain MRI scan. While deep learning methods for single-scan lesion segmentation are common, deep learning approaches for lesion activity have only been proposed recently. Here, a two-path architecture processes two 3D MRI volumes from two time points. In this work, we investigate whether extending this problem to full 4D deep learning using a history of MRI volumes and thus an extended baseline can improve performance. For this purpose, we design a recurrent multi-encoder-decoder architecture for processing 4D data. We find that adding more temporal information is beneficial and our proposed architecture outperforms previous approaches with a lesion-wise true positive rate of 0.84 at a lesion-wise false positive rate of 0.19.