Improving brain tumor segmentation with anatomical prior-informed pre-training

Kang Wang; Zeyang Li; Haoran Wang; Siyu Liu; Mingyuan Pan; Manning Wang; Shuo Wang; Zhijian Song

doi:10.3389/fmed.2023.1211800

Improving brain tumor segmentation with anatomical prior-informed pre-training

Front Med (Lausanne). 2023 Sep 13:10:1211800. doi: 10.3389/fmed.2023.1211800. eCollection 2023.

Authors

Kang Wang^#^{1

2}, Zeyang Li^#³, Haoran Wang^{1

2}, Siyu Liu^{1

2}, Mingyuan Pan⁴, Manning Wang^{1

2}, Shuo Wang^{1

2}, Zhijian Song^{1

2}

Affiliations

¹ Digital Medical Research Center, School of Basic Medical Sciences, Fudan University, Shanghai, China.
² Shanghai Key Lab of Medical Image Computing and Computer Assisted Intervention, Fudan University, Shanghai, China.
³ Department of Neurosurgery, Zhongshan Hospital, Fudan University, Shanghai, China.
⁴ Radiation Oncology Center, Huashan Hospital, Fudan University, Shanghai, China.

^# Contributed equally.

Abstract

Introduction: Precise delineation of glioblastoma in multi-parameter magnetic resonance images is pivotal for neurosurgery and subsequent treatment monitoring. Transformer models have shown promise in brain tumor segmentation, but their efficacy heavily depends on a substantial amount of annotated data. To address the scarcity of annotated data and improve model robustness, self-supervised learning methods using masked autoencoders have been devised. Nevertheless, these methods have not incorporated the anatomical priors of brain structures.

Methods: This study proposed an anatomical prior-informed masking strategy to enhance the pre-training of masked autoencoders, which combines data-driven reconstruction with anatomical knowledge. We investigate the likelihood of tumor presence in various brain structures, and this information is then utilized to guide the masking procedure.

Results: Compared with random masking, our method enables the pre-training to concentrate on regions that are more pertinent to downstream segmentation. Experiments conducted on the BraTS21 dataset demonstrate that our proposed method surpasses the performance of state-of-the-art self-supervised learning techniques. It enhances brain tumor segmentation in terms of both accuracy and data efficiency.

Discussion: Tailored mechanisms designed to extract valuable information from extensive data could enhance computational efficiency and performance, resulting in increased precision. It's still promising to integrate anatomical priors and vision approaches.

Keywords: anatomical priors; brain tumor segmentation; magnetic resonance image; masked autoencoder; self-supervised learning; transformer.

Grants and funding

This work was supported by National Natural Science Foundation of China under Grant 82072021 and SW was supported by Shanghai Sailing Programs of Shanghai Municipal Science and Technology Committee (22YF1409300).