NIMG-49. RADIO-PATHOMIC MAPS OF DE-NOVO GLIOBLASTOMA IDENTIFY PHENOTYPES OF TUMOR INVASION ASSOCIATED WITH PROGNOSIS

Abstract PURPOSE Although tumor is known to exist beyond the contrast-enhancing margin, a lack of access pathological information enhancement has hindered study non-angiogenic characteristics and their effects on patient prognosis. This tested hypothesis that radio-pathomic maps developed using autopsy tissue samples can identify phenotypes relevant survival treatment outcomes. METHODS A dataset aligned MRI as ground truth from 65 glioma patients (training set n=43, test n=22) was used train machine learning models predict map cellularity (Cell), extracellular fluid (ECF), cytoplasm (Cyt) density following our previously published methodology (Bobholz et al. 2022). Cell, ECF, Cyt were then generated pre-surgical scans an independent 80 glioblastoma patients. Each separated into groups based appearance margin. Group included Well-Circumscribed (WC) tumors with no activity contrast enhancement, Necrotic Front (NF) areas necrosis extending Hypercellular (HF) increased surrounding tumor, Hybrid (HYF) both HF NF tumors. Cox regression assess differences between phenotypes, controlling for history. RESULTS Across patients, 22 classified WC, 14 HF, 24 NF, 20 HYF. HYF each showed significant/trending reductions in when compared WC (HR=2.02, p=0.03; HR=2.0, p=0.06; HR=1.75, p=0.09, respectively). CONCLUSION Radio-pathomic margin affect overall outcome glioblastoma.