Introduction: Several studies have examined correlates between imaging features of neoplasm and patient survival or tumor genetic composition; however, few have generated predictive models robust enough to enter clinical practice. In this study, we use advanced pattern analysis and machine learning to identify a combination of imaging features on initial magnetic resonance (MR) images to predict overall survival and molecular subtype in patients with glioblastoma (GB).
Methods: We performed a retrospective followed by a prospective cohort study of GB patients. Imaging features were extracted from structural, diffusion, and perfusion MR images at time of diagnosis. A machine learning algorithm was used to examine multiple features simultaneously to determine which set of features was most predictive of survival. The model was tested prospectively in a separate cohort of patients. In a subset of patients for which genetic data was obtained, machine learning was used to classify the likelihood of molecular subtype affiliation based on imaging. Ten-fold cross validation was performed.
Results: The accuracy of the model in predicting survival was 77% in the retrospectively study (n=105) and 79% in the prospective study (n=29). Constellations of imaging markers related to infiltration and diffusion of tumor cells into edema, microvascularity, and blood-brain barrier compromise were predictive of shortened survival. A separate model was generated to predict molecular subtype. The accuracy of individual subtype predictions was 85% for classical (n=20) 84% for mesenchymal (n=28), 88% for neural (n=29), and 86% for proneural (n=22).
Conclusions: Unlike prior studies, we analyzed the entirety of imaging data in an integrative fashion, leveraging the power of pattern analysis and machine learning to predict survival and molecular subtype with high accuracy and reproducibility in GB. Our non-invasive model utilizes multi-parametric imaging obtained routinely for GB patients, making it readily translatable to the clinic.
Patient Care: Our findings provide a robust mechanism for predicting survival among GB patients and for estimating molecular subtype non-invasively. Furthermore, GB tumors exhibit regional differences in genomic alterations, enhancement, cell density, and necrosis, and imaging provides a means of capturing such spatial heterogeneity of tumors, especially when pathology sampling is suboptimal or not possible due to tumor location. By utilizing standard imaging sequences that are commonly employed in clinical practice, our findings are readily translatable to the clinic and follow the paradigm of personalized medicine.
Learning Objectives: By the conclusion of this session, participants should be able to: 1) Describe the strengths of image analysis and machine learning as tools to analyze the entirety of imaging data and to uncover clinically relevant patterns not detectable by visual assessment alone, 2) Discuss, in small groups, ways in which non-invasive predictions of survival and molecular subtype would help to guide patient management in the field of Neuro-Oncology, 3) Identify ways in which informatics-derived imaging biomarkers can be incorporated into clinical practice and applied in future studies to evaluate treatment response over time and response to targeted agents.