Multimodal Deep Learning-Based Prognostication in Glioma Patients: A Systematic Review

Simple Summary Primary malignant tumors of the brain are relatively rare, but their contribution to death due to cancer is disproportionately large. The use of multimodal data in machine learning techniques (such as deep learning) is still relatively new, but its implications for predicting brain tumor characteristics, treatment response, and patient survival are robust. In this study, we sought to review the current state of glioma prognostication using deep learning methods. A systematic review of the deep learning-based prognostication of gliomas was performed in accordance with PRISMA guidelines. All included studies focused on the prognostication of gliomas, and predicted overall survival, overall survival along with genotype characteristics, and response to immuno-therapy. Multimodal analyses were varied, with 6 studies combining MRI with clinical data; 6 studies integrating MRI with histologic, clinical, and biomarker data; 3 studies combining MRI with genomic data; and 1 study combining histologic imaging with clinical data. Overall, the use of multimodal data in deep learning assessments of gliomas leads to a more accurate prediction of overall patient survival as compared to unimodal models. As data collection and computational capacity expands, further improvements are likely from the continued integration of different data modalities into deep learning models. Malignant brain tumors pose a substantial burden on morbidity and mortality. As clinical data collection improves, along with the capacity to analyze it, novel predictive clinical tools may improve prognosis prediction. Deep learning (DL) holds promise for integrating clinical data of various modalities. A systematic review of the DL-based prognostication of gliomas was performed using the Embase (Elsevier), PubMed MEDLINE (National library of Medicine), and Scopus (Elsevier) databases, in accordance with PRISMA guidelines. All included studies focused on the prognostication of gliomas, and predicted overall survival (13 studies, 81%), overall survival as well as genotype (2 studies, 12.5%), and response to immunotherapy (1 study, 6.2%). Multimodal analyses were varied, with 6 studies (37.5%) combining MRI with clinical data; 6 studies (37.5%) integrating MRI with histologic, clinical, and biomarker data; 3 studies (18.8%) combining MRI with genomic data; and 1 study (6.2%) combining histologic imaging with clinical data. Studies that compared multimodal models to unimodal-only models demonstrated improved predictive performance. The risk of bias was mixed, most commonly due to inconsistent methodological reporting. Overall, the use of multimodal data in DL assessments of gliomas leads to a more accurate overall survival prediction. However, due to data limitations and a lack of transparency in model and code reporting, the full extent of multimodal DL as a resource for brain tumor patients has not yet been realized.