# OMT and tensor SVD–based deep learning model for segmentation and predicting genetic markers of glioma: A multicenter study

## TL;DR

This study introduces a deep learning model using OMT and tensor SVD to accurately segment and predict genetic markers of glioma from MRI scans, outperforming radiologists.

## Contribution

A novel deep learning model combining optimal mass transport and multimode tensor SVD for glioma segmentation and genetic marker prediction.

## Key findings

- The OMT segmentation model achieved a mean Dice score of 0.880 for tumor regions.
- The OMT-APC model outperformed four radiologists in predicting WHO grade, IDH mutation, and 1p/19q codeletion with high accuracy and AUC scores.
- The model demonstrated strong generalizability across 16 multicenter datasets from Asia, Europe, and America.

## Abstract

Accurate characterization of glioma is essential for effective clinical decision-making. Most current studies involve a limited number of patients and focus solely on single-gene tasks. This research introduces a novel deep learning model based on OMT and multimode tensor SVD to predict molecular markers using international multicenter datasets. Our approach efficiently compresses irrelevant information while enhancing tumor-region features through OMT. Additionally, we innovatively integrate an algebraic preclassification model, derived from multimode tensor SVD, with deep learning networks. This combination significantly improves the model’s ability to recognize tumor and classify genetic subtypes. Experimental validation on multicenter datasets demonstrates that our method is highly reproducible and generalizable, offering promising potential for glioma analysis and clinical applications.

Glioma is the most common primary malignant brain tumor and preoperative genetic profiling is essential for the management of glioma patients. Our study focused on tumor regions segmentation and predicting the World Health Organization (WHO) grade, isocitrate dehydrogenase (IDH) mutation, and 1p/19q codeletion status using deep learning models on preoperative MRI. To achieve accurate tumor segmentation, we developed an optimal mass transport (OMT) approach to transform irregular MRI brain images into tensors. In addition, we proposed an algebraic preclassification (APC) model utilizing multimode OMT tensor singular value decomposition (SVD) to estimate preclassification probabilities. The fully automated deep learning model named OMT-APC was used for multitask classification. Our study incorporated preoperative brain MRI data from 3,565 glioma patients across 16 datasets spanning Asia, Europe, and America. Among these, 2,551 patients from 5 datasets were used for training and internal validation. In comparison, 1,014 patients from 11 datasets, including 242 patients from The Cancer Genome Atlas (TCGA), were used as independent external test. The OMT segmentation model achieved mean lesion-wise Dice scores of 0.880. The OMT-APC model was evaluated on the TCGA dataset, achieving accuracies of 0.855, 0.917, and 0.809, with AUC scores of 0.845, 0.908, and 0.769 for WHO grade, IDH mutation, and 1p/19q codeletion, respectively, which outperformed the four radiologists in all tasks. These results highlighted the effectiveness of our OMT and tensor SVD–based methods in brain tumor genetic profiling, suggesting promising applications for algebraic and geometric methods in medical image analysis.