Prediction of lower-grade glioma molecular subtypes using deep learning

Matsui, Yutaka; Maruyama, Takashi; Nitta, Masayuki; Saito, Taiichi; Tsuzuki, Shunsuke; Tamura, Manabu; Kusuda, Kaori; Fukuya, Yasukazu; Asano, Hidetsugu; Kawamata, Takakazu; Masamune, Ken; Muragakı, Yoshihiro

doi:10.1007/s11060-019-03376-9

Cited by 74 publications

(74 citation statements)

References 34 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…It is worth mentioning that these comparisons can only be used just as an indication because they were applied to different datasets with different scan types, MRI modalities and patient’s characteristics. For instance, Matsui et al [ 6 ] used residual network-based deep network that required more modalities of data (FLAIR, T1ce, T1, T2), including PET and CT scans in addition to other side information of patients as numeric data. Zhou et al [ 7 ] used hand-crafted features such as histograms, shape and texture from data that was collected from single institution combined with age information for a random forest classifier.…”

Section: Experimental Resultsmentioning

confidence: 99%

“…These methods may provide solutions for predicting molecular subtype gliomas by automatic feature learning. Matsui et al [6] proposed a residual network-based model using multiple scans from MRI, positron emission tomography (PET) and computed tomography (CT) along with different characteristics of patients as numeric data for predicting three categories of molecular subtype. Liang et al [16] applied 3D DensNets using multi-modal MRIs for IDH mutation prediction.…”

Section: Related Workmentioning

confidence: 99%

“…The molecular information would be of practical value since oligodendrogliomas harbor better prognosis than the other LGG subtypes and also seem to be more sensitive to oncological treatment [3,4]. This molecular information requires a tissue diagnosis, but recently several advanced machine learning techniques have been shown to predict molecular subtypes in gliomas based upon preoperative imaging [5][6][7][8]. Non-invasive diagnostic tools are attractive in identification since it may assist in prognostication and would significantly enhance patient counseling and shared decision making.…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

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

Ali

Berger

et al. 2020

Brain Sciences

View full text Add to dashboard Cite

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.

show abstract

Section: Experimental Resultsmentioning

confidence: 99%

Section: Related Workmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

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

Ali

Berger

et al. 2020

Brain Sciences

View full text Add to dashboard Cite

show abstract

“…Also more recently, researchers have demonstrated achievements of deep learning (DL) in the image segmentation and glioma grades prediction (32)(33)(34)(35)(36)(37). Convolutional neural networks (CNNs) started outperforming other methods on several high-profile image analysis projects.…”

Section: Discussionmentioning

confidence: 99%

Machine Learning-Based Radiomics Predicting Tumor Grades and Expression of Multiple Pathologic Biomarkers in Gliomas

et al. 2020

View full text Add to dashboard Cite

Background: The grading and pathologic biomarkers of glioma has important guiding significance for the individual treatment. In clinical, it is often necessary to obtain tumor samples through invasive operation for pathological diagnosis. The present study aimed to use conventional machine learning algorithms to predict the tumor grades and pathologic biomarkers on magnetic resonance imaging (MRI) data. Methods: The present study retrospectively collected a dataset of 367 glioma patients, who had pathological reports and underwent MRI scans between October 2013 and March 2019. The radiomic features were extracted from enhanced MRI images, and three frequently-used machine-learning models of LC, Support Vector Machine (SVM), and Random Forests (RF) were built for four predictive tasks: (1) glioma grades, (2) Ki67 expression level, (3) GFAP expression level, and (4) S100 expression level in gliomas. Each sub dataset was split into training and testing sets at a ratio of 4:1. The training sets were used for training and tuning models. The testing sets were used for evaluating models. According to the area under curve (AUC) and accuracy, the best classifier was chosen for each task. Results: The RF algorithm was found to be stable and consistently performed better than Logistic Regression and SVM for all the tasks. The RF classifier on glioma grades achieved a predictive performance (AUC: 0.79, accuracy: 0.81). The RF classifier also achieved a predictive performance on the Ki67 expression (AUC: 0.85, accuracy: 0.80). The AUC and accuracy score for the GFAP classifier were 0.72 and 0.81. The AUC and accuracy score for S100 expression levels are 0.60 and 0.91. Conclusion: The machine-learning based radiomics approach can provide a noninvasive method for the prediction of glioma grades and expression levels of multiple pathologic biomarkers, preoperatively, with favorable predictive accuracy and stability.

show abstract

“…18) The preoperative prediction of the genetic classification using deep learning would be combined in future. 19) The log-based estimation could support execution of such resection strategies by providing intraoperative EOR monitoring.…”

Section: Discussionmentioning

confidence: 99%

Reliability of Residual Tumor Estimation Based on Navigation Log

Yamada

Maruyama

Konishi

et al. 2020

Neurol. Med. Chir.(Tokyo)

Self Cite

View full text Add to dashboard Cite

The mass of residual tumors has previously been estimated using time-series records of the position of surgical instruments acquired from neurosurgical navigation systems (navigation log). This method has been shown to be useful for rapid evaluation of residual tumors during resection. However, quantitative analysis of the method's reliability has not been sufficiently reported. The effect of poor log coverage is dominant in previous studies, in that it did not highlight other disturbance factors, such as intraoperative brain shift. We analyzed 25 patients with a high log-acquisition rate that was calculated by dividing the log-available time by the instrument-use time. We estimated the region of resection using the trajectory of surgical instrument that was extracted from the navigation log. We then calculated the residual tumor region and measured its volume as log-estimation residual tumor volume (RTV). We evaluated the correlation between the log-estimation RTV and the RTV in the post-resection magnetic resonance (MR) image. We also evaluated the accuracy of detecting the residual tumor mass using the estimated residual tumor region. The log-estimation RTV and the RTV in the post-resection MR image were significantly correlated (correlation coefficient = 0.960; P <0.001). The presence of patient-wise residual tumor mass was detected with a sensitivity of 81.8% and a specificity of 92.9%. The individual residual tumor mass was detected with a positive predictive value of 72%. Estimation of residual tumor with adequate log coverage appears to be a suitable method with a high reliability. This method can support rapid decision-making during resection.

show abstract

Prediction of lower-grade glioma molecular subtypes using deep learning

Cited by 74 publications

References 34 publications

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

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

Machine Learning-Based Radiomics Predicting Tumor Grades and Expression of Multiple Pathologic Biomarkers in Gliomas

Reliability of Residual Tumor Estimation Based on Navigation Log

Contact Info

Product

Resources

About