Track papers across labs, topics, and trends.
Search papers, labs, and topics across Lattice.
This study addresses the complex challenge of lesion segmentation in moderate to severe traumatic brain injury (msTBI) using a deep learning approach based on the nnU-Net framework. By implementing an adaptive intensity normalization strategy specifically for brain parenchyma, the researchers effectively reduced inter-subject variability and artifacts, leading to competitive performance on the AIMS-TBI 2025 Challenge leaderboard. The model achieved an Overall Dice Coefficient of 0.6305, with a notable Dice score of 0.4805 for lesion segmentation, highlighting both the challenges of detecting heterogeneous lesions and the model's strong specificity for non-lesion tissue.
Achieving a Dice score of 0.4805 for lesion segmentation reveals the significant challenges posed by heterogeneous lesions in msTBI, while maintaining high specificity for non-lesion identification.
The segmentation of lesions in Moderate to Severe Traumatic Brain Injury (msTBI) from T1-weighted MRI presents a significant clinical challenge due to the profound heterogeneity of lesion characteristics in terms of size, shape, and location. To address this, the AIMS-TBI 2025 Challenge was organized to promote the development of robust and accurate segmentation algorithms. In this paper, we present our deep learning-based solution. Our methodology employs the nnU-Net framework with an adaptive intensity normalization strategy confined to the brain parenchyma, effectively reducing inter-subject variability and mitigating artifacts from non-brain structures. Upon final evaluation on the held-out test set, our method demonstrated highly competitive performance on the official leaderboard, achieving an Overall Dice Coefficient of 0.6305. The model obtained a Dice score of 0.4805 for lesion segmentation and 0.9324 for non-lesion tissue. While the lesion Dice reflects the difficulty of detecting highly heterogeneous lesions, the high non-lesion Dice primarily indicates the model's strong ability to correctly identify non-lesion voxels, demonstrating good specificity in differentiating lesion from non-lesion regions. These results demonstrate that incorporating anatomically constrained normalization within the nnU-Net pipeline is a powerful and effective strategy for tackling the complexities of msTBI lesion segmentation.