Transductive Transfer Learning for Domain Adaptation in Brain Magnetic Resonance Image Segmentation

Abstract

Segmentation of brain images from Magnetic Resonance Images (MRI) is an indispensable step in clinical practice. Morphological changes of sub-cortical brain structures and quantification of brain lesions are considered biomarkers of neurological and neurodegenerative disorders and used for diagnosis, treatment planning, and monitoring disease progression. In recent years, deep learning methods showed an outstanding performance in medical image segmentation. However, these methods suffer from generalisability problem due to inter-centre and inter-scanner variabilities of the MRI images. The main objective of the study is to develop an automated deep learning segmentation approach that is accurate and robust to the variabilities in scanner and acquisition protocols. In this paper, we propose a transductive transfer learning approach for domain adaptation to reduce the domain-shift effect in brain MRI segmentation. The transductive scenario assumes that there are sets of images from two different domains: (1) source—images with manually annotated labels; and (2) target—images without expert annotations. Then, the network is jointly optimised integrating both source and target images into the transductive training process to segment the regions of interest and to minimise the domain-shift effect. We proposed to use a histogram loss in the feature level to carry out the latter optimisation problem. In order to demonstrate the benefit of the proposed approach, the method has been tested in two different brain MRI image segmentation problems using multi-centre and multi-scanner databases for: (1) sub-cortical brain structure segmentation; and (2) white matter hyperintensities segmentation. The experiments showed that the segmentation performance of a pre-trained model could be significantly improved by up to 10%. For the first segmentation problem it was possible to achieve a maximum improvement from 0.680 to 0.799 in average Dice Similarity Coefficient (DSC) metric and for the second problem the average DSC improved from 0.504 to 0.602. Moreover, the improvements after domain adaptation were on par or showed better performance compared to the commonly used traditional unsupervised segmentation methods (FIRST and LST), also achieving faster execution time. Taking this into account, this work presents one more step toward the practical implementation of deep learning algorithms into the clinical routine.

Pseudo-Label Assisted Nnu-Net (PLAn) Enables Automatic Segmentation of 7T MRI From a Single Acquisition

“…While one prior study has utilized a similar approach for cardiac cine MRI, (38) to our knowledge, this method has not previously been explored for brain MRI segmentation. This is somewhat surprising, since other types of domain adaptation, such as adaptation between different sites (36,39) or contrasts,(28) have been considered. Another area for further exploration may be transfer learning across different neurological diseases.…”

“…Another area for further exploration may be transfer learning across different neurological diseases. While this has been explored in a few prior studies, (28,39) it is typically not decoupled from other domain adaptation tasks (e.g., changes in scanner or contrast), meaning it is yet underexplored.…”

On the usability of synthetic data for improving the robustness of deep learning-based segmentation of cardiac magnetic resonance images

“…We compare the performance of the model adapted to vendor C using a synthesis pipeline to a model that is trained in a similar fashion, where instead of generated synthetic data, we utilize a histogram standardization approach to mimic the average intensity distribution of vendor C (model Hist-C Aug). Histogram standardization has become a common approach to tackle the domain shift appearing in medical images of the same modality, but acquired from different vendors and centers (Kushibar et al, 2021;Campello et al, 2021). For this purpose, we utilize a landmark-based histogram standardization approach proposed in Nyúl et al (2000).…”