Can conventional magnetic resonance imaging predict survival in pediatric diffuse intrinsic pontine glioma? A single institution experience

MR imaging features of diffuse intrinsic pontine glioma and relationship to overall survival: report from the International DIPG Registry

BACKGROUND

This study describes imaging features of diffuse intrinsic pontine glioma (DIPG) and correlates with overall survival (OS) and histone mutation status in the International DIPG Registry (IDIPGR).

METHODS

Four hundred cases submitted to the IDIPGR with a local diagnosis of DIPG and baseline MRI were evaluated by consensus review of 2 neuroradiologists; 43 cases were excluded (inadequate imaging or alternative diagnoses). Agreement between reviewers, association with histone status, and univariable and multivariable analyses relative to OS were assessed.

RESULTS

On univariable analysis imaging features significantly associated with worse OS included: extrapontine extension, larger size, enhancement, necrosis, diffusion restriction, and distant disease. On central review, 9.5% of patients were considered not to have DIPG. There was moderate mean agreement of MRI features between reviewers. On multivariable analysis, chemotherapy, age, and distant disease were predictors of OS. There was no difference in OS between wild-type and H3 mutated cases. The only imaging feature associated with histone status was the presence of ill-defined signal infiltrating pontine fibers.

CONCLUSIONS

Baseline imaging features are assessed in the IDIPGR. There was a 9.5% discordance in DIPG diagnosis between local and central review, demonstrating need for central imaging confirmation for prospective trials. Although several imaging features were significantly associated with OS (univariable), only age and distant disease were significant on multivariable analyses. There was limited association of imaging features with histone mutation status, although numbers are small and evaluation exploratory.

A Cascaded Deep Convolutional Neural Network for Joint Segmentation and Genotype Prediction of Brainstem Gliomas

GOAL

Automatic segmentation of brainstem gliomas and prediction of genotype (H3 K27M) mutation status based on magnetic resonance (MR) images are crucial but challenging tasks for computer-aided diagnosis in neurosurgery. In this paper, we present a novel cascaded deep convolutional neural network (CNN) to address these two challenging tasks simultaneously.

METHODS

Our novel segmentation task contains two feature-fusion modules: the Gaussian-pyramid multiscale input features-fusion technique and the brainstem-region feature enhancement. The aim is to resolve very difficult problems in brainstem glioma segmentation. Our prediction model combines CNN features and support-vector-machine classifier to automatically predict genotypes without region-of-interest labeled-MR images and is learned jointly with the segmentation task. First, Gaussian-pyramid multiscale input feature fusion is added to our glioma-segmentation task to solve the problems of size variety and weak brainstem-gliomas boundaries. Second, the two feature-fusion modules provide both local and global contexts to retain higher frequency details for sharper tumor boundaries, handling the problem of the large variation of tumor shape, and volume resolution.

RESULTS AND CONCLUSION

Experiments demonstrate that our cascaded CNN method achieves not only a good tumor segmentation result with a high Dice similarity coefficient of 77.03%, but also a competitive genotype prediction result with an average accuracy of 94.85% upon fivefold cross-validation.