Abstract
BACKGROUND
The importance of detecting the genomic status of gliomas is increasingly recognized and IDH (isocitrate dehydrogenase) mutation and TERT (telomerase reverse transcriptase) promoter mutation have a significant impact on treatment decisions. Noninvasive prediction of these genomic statuses in gliomas is a challenging problem; however, a deep learning model using magnetic resonance imaging (MRI) can be a solution. The image differences among facilities causing performance degradation, called domain shift, have also been reported in other tasks such as brain tumor segmentation. We investigated whether a deep learning model could predict the gene status, and if so, to what extent it would be affected by domain shift.
METHOD
We used the Multimodal Brain Tumor Segmentation Challenge (BraTS) data and the Japanese cohort (JC) dataset consisted of brain tumor images collected from 544 patients in 10 facilities in Japan. We focused on IDH mutation and TERT promoter mutation. The deep learning models to predict the statuses of these genes were trained by the BraTS dataset or the training portion of the JC dataset, and the test portion of the JC dataset evaluated the accuracy of the models.
RESULTS
The IDH mutation predicting model trained by the BraTS dataset showed 80.0% accuracy for the validation portion of the BraTS dataset; however, only 67.3% for the test portion of the JC dataset. The TERT promoter mutation predicting model trained by the training portion of the JC dataset showed only 49% accuracy for the test portion of the JC dataset.
CONCLUSION
IDH mutation can be predicted by deep learning models using MRI, but the performance degeneration by domain shift was significant. On the other hand, TERT promoter mutation could not be predicted accurately enough by current deep learning techniques. In both mutations, further studies are needed.
A Deep Learning Model Based on Concatenation Approach for the Diagnosis of Brain Tumor
