Domain Mapping and Deep Learning from Multiple MRI Clinical Datasets for Prediction of Molecular Subtypes in Low Grade Gliomas

Abstract

Brain tumors, such as low grade gliomas (LGG), are molecularly classified which require the surgical collection of tissue samples. The pre-surgical or non-operative identification of LGG molecular type could improve patient counseling and treatment decisions. However, radiographic approaches to LGG molecular classification are currently lacking, as clinicians are unable to reliably predict LGG molecular type using magnetic resonance imaging (MRI) studies. Machine learning approaches may improve the prediction of LGG molecular classification through MRI, however, the development of these techniques requires large annotated data sets. Merging clinical data from different hospitals to increase case numbers is needed, but the use of different scanners and settings can affect the results and simply combining them into a large dataset often have a significant negative impact on performance. This calls for efficient domain adaption methods. Despite some previous studies on domain adaptations, mapping MR images from different datasets to a common domain without affecting subtitle molecular-biomarker information has not been reported yet. In this paper, we propose an effective domain adaptation method based on Cycle Generative Adversarial Network (CycleGAN). The dataset is further enlarged by augmenting more MRIs using another GAN approach. Further, to tackle the issue of brain tumor segmentation that requires time and anatomical expertise to put exact boundary around the tumor, we have used a tight bounding box as a strategy. Finally, an efficient deep feature learning method, multi-stream convolutional autoencoder (CAE) and feature fusion, is proposed for the prediction of molecular subtypes (1p/19q-codeletion and IDH mutation). The experiments were conducted on a total of 161 patients consisting of FLAIR and T1 weighted with contrast enhanced (T1ce) MRIs from two different institutions in the USA and France. The proposed scheme is shown to achieve the test accuracy of 74 . 81 % on 1p/19q codeletion and 81 . 19 % on IDH mutation, with marked improvement over the results obtained without domain mapping. This approach is also shown to have comparable performance to several state-of-the-art methods.

Preoperative Radiomics Analysis of 1p/19q Status in WHO Grade II Gliomas

“…Although the sample size was limited, it allowed the results to be more consistent. Our model showed a similar performance to the machine-learning and deep-learning models mentioned above, with an overall AUC of 0.8079 and an accuracy of 0.758 ( 34 , 35 , 38 , 39 ). We specifically analyzed the predictive model in the subgroup analysis of gliomas with mutant IDH, which excluded the influence of gliomas with wild-type IDH, and found a similar performance value.…”

“…Since the predictive performance of these models was restrained by the small sample size, data augmentation was introduced to enlarge the size of the training set. Based on Cycle Generative Adversarial Network, multi-stream convolutional autoencoder and feature fusion are proposed for the prediction of 1p/19q co-deletion, which displayed an accuracy of 78.41% in low-grade gliomas ( 38 ). After adding 30-fold augmented data, another study improved the accuracy of the convolutional neural networks model from 78.3% to 87.7% ( 39 ).…”

Transfer Learning in Magnetic Resonance Brain Imaging: A Systematic Review

“…We found 7 asymmetric approaches that transformed target domain features into source domain features via generative adversarial networks [ 56 , 57 ], Bregman divergence minimization [ 58 ], probabilistic models [ 59 ], median [ 60 ], and nearest neighbors [ 61 ]. Contrarily, Qin et al [ 62 ] transformed source domain features into target domain features via dictionary-based interpolation to optimize a model on the target domain.…”

Hybridized Deep Neural Network Using Adaptive Rain Optimizer Algorithm for Multi-Grade Brain Tumor Classification of MRI Images

“…To advance molecular facts of cancerous brain substances, MR Spectroscopy (MRS) can be used as additional technique. This diversity of acquirements brands MRI a multimodal device for tumor oncology [9,10] .…”