Related Experiment Video
Updated: Jul 23, 2025
Whole-brain Segmentation and Change-point Analysis of Anatomical Brain MRI—Application in Premanifest Huntington's Disease
Published on: June 9, 2018
Deep learning-based prediction of H3K27M alteration in diffuse midline gliomas based on whole-brain MRI
Bowen Huang1, Yuekang Zhang1, Qing Mao1
1Department of Neurosurgery, West China Hospital of Sichuan University, Chengdu, China.
Background:
H3K27M mutation status significantly affects the prognosis of patients with diffuse midline gliomas (DMGs), but this tumor presents a high risk of pathological acquisition. We aimed to construct a fully automated model for predicting the H3K27M alteration status of DMGs based on deep learning using whole-brain MRI.
Methods:
DMG patients from West China Hospital of Sichuan University (WCHSU; n = 200) and Chengdu Shangjin Nanfu Hospital (CSNH; n = 35) who met the inclusion and exclusion criteria from February 2016 to April 2022 were enrolled as the training and external test sets, respectively. To adapt the model to the human head MRI scene, we use normal human head MR images to pretrain the model. The classification and tumor segmentation tasks are naturally related, so we conducted cotraining for the two tasks to enable information interaction between them and improve the accuracy of the classification task.
Results:
The average classification accuracies of our model on the training and external test sets was 90.5% and 85.1%, respectively. Ablation experiments showed that pretraining and cotraining could improve the prediction accuracy and generalization performance of the model. In the training and external test sets, the average areas under the receiver operating characteristic curve (AUROCs) were 94.18% and 87.64%, and the average areas under the precision-recall curve (AUPRC) were 93.26% and 85.4%.
Conclusions:
The developed model achieved excellent performance in predicting the H3K27M alteration status in DMGs, and its good reproducibility and generalization were verified in the external dataset.
Insights
A deep learning model accurately predicts H3K27M mutation status in diffuse midline gliomas (DMGs) using whole-brain MRI. This automated approach aids in prognosis for DMG patients, overcoming pathological acquisition challenges.
Area of Science:
- Neuro-oncology
- Medical Imaging
- Artificial Intelligence
Background:
- Diffuse midline gliomas (DMGs) prognosis is linked to H3K27M mutation status.
- Pathological acquisition poses challenges for accurate H3K27M status determination in DMGs.
Purpose of the Study:
- To develop a fully automated deep learning model for predicting H3K27M alteration status in DMGs.
- To leverage whole-brain MRI for non-invasive H3K27M status prediction.
Main Methods:
- Utilized deep learning on whole-brain MRI data from 235 DMG patients (WCHSU and CSNH).
- Employed model pretraining on normal human head MRI and cotraining for classification and tumor segmentation tasks.
- Integrated information interaction between classification and segmentation to enhance predictive accuracy.
Main Results:
- Achieved average classification accuracies of 90.5% (training) and 85.1% (external test).
- Demonstrated strong performance with average AUROCs of 94.18% (training) and 87.64% (external test).
- Reported average AUPRCs of 93.26% (training) and 85.4% (external test), with pretraining and cotraining improving accuracy.
Conclusions:
- The developed deep learning model shows excellent performance in predicting H3K27M alteration status in DMGs.
- The model's reproducibility and generalization capabilities were validated on an external dataset.
- This automated approach offers a promising tool for DMG patient management and prognosis.
