Abstract 5200: Multiparametric MRI along with machine learning informs on molecular underpinnings, prognosis, and treatment response in pediatric low grade glioma

Abstract

Abstract We report a comprehensive radiogenomic analysis, including multiparametric MRI and RNA sequencing of 502 tumors across 13 histological types of pediatric low-grade glioma (PLGG). Through an unsupervised analysis of treatment-naïve imaging, we were able to identify 3 distinct imaging clusters that were associated with different molecular pathways according to transcriptomic pathway analysis. Using xCell enrichment scores as input for clustering, we discovered three distinct immunological groups based on immune cell infiltration. The results suggested the presence of an ‘immune-hot” cluster within the tumors with a poorer prognosis, which may potentially benefit from emerging immunotherapies. We subsequently generated a radiomic signature that was able to predict immunological profiles. We also developed a clinicoradiomic predictive model that can assess patients’ risk of progression and predict treatment response. Finally, we evaluated the transcriptomic pathways that were most predictive of clinicoradiomic risk scores and identified the biological pathways most predictive of risk of progression in PLGG, when adjusting for molecular subtype, race, reported gender, age at diagnosis, and the anatomic location of the tumor. This is the first large-scale radiogenomic analysis in PLGG that can aid in prognostication, assessing patients’ risk of progression, predicting treatment response in conventional therapies, and early stratification of patients to identify potential candidates for targeted therapies. Citation Format: Ali Nabavizadeh, Adam Kraya, Komal Rathi, Meen Chul Kim, Nastaran Khalili, Ariana Familiar, Varun Kesherwani, Hannah Anderson, Deep Gandhi, Aria Mahtabfar, Sina Bagheri, Xiaoyan Huang, Yuankun Zhu, Alexander Sickler, Saksham Phul, Sabine Mueller, Mateusz P Koptyra, Philip B Storm, Jeff Ware, Yuanquan Song, Michael J Fisher, Arastoo Vossough, Christos Davatzikos, Adam C Resnick, Anahita Fathi Kazerooni. Multiparametric MRI along with machine learning informs on molecular underpinnings, prognosis, and treatment response in pediatric low grade glioma [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2024; Part 1 (Regular Abstracts); 2024 Apr 5-10; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2024;84(6_Suppl):Abstract nr 5200.