DCE-MRI Tumor Vascular Parameters in Two Preclinical Patient-Derived Orthotopic Xenograft Models of Glioblastoma
Recommended Citation
Acharya PC, Nagaraja TN, Brown SL, deCarvalho AC, Tabbarah AZ, Cabral G, Knight RA, Lee I, Divine GW, and Ewing JR. DCE-MRI Tumor Vascular Parameters in Two Preclinical Patient-Derived Orthotopic Xenograft Models of Glioblastoma. NMR Biomed 2025;38(8):e70089.
Document Type
Article
Publication Date
8-1-2025
Publication Title
NMR in biomedicine
Abstract
Two preclinical patient-derived orthotopic xenograft (PDOX) models of glioblastoma (GBM) were characterized using measures of tumor physiology. Plasma volume fraction (v(p)), blood-to-tissue forward volumetric transfer constant (K(trans)), and interstitial volume fraction (v(e)) were estimated via dynamic contrast-enhanced (DCE) MRI. Tumor blood flow (TBF) was estimated via continuous arterial spin-labeling and apparent diffusion coefficient of water (ADC) via spin-echo diffusion-weighted imaging. Tumor distribution volume at the tumor rim (V(D)) and peritumoral flux (Flux) were also estimated. Two neurosphere cell lines, taken from a primary human GBM (HF3016) and its recurrence (HF3177), were used in 15 immune-compromised athymic rats (n = 7 for HF3016; n = 8 for HF3177). When the tumors grew to about 3-4 mm in diameter, DCE-MRI data were acquired in a 7T magnet using a low molecular weight gadolinium-chelate contrast agent. DCE data were analyzed voxel-by-voxel using Patlak, extended Patlak, and Logan graphical methods. A data-driven model selection approach was applied to segment the tumor region, and regions of interest (ROIs) based on that segmentation were selected in the imaging slice having the largest tumor cross section. Summary ROI statistics of vascular measures were produced. The parameter estimates K(trans), v(e), v(p), V(D), ADC, TBF, and growth rates between the two models varied slightly, but the differences were not statistically significant (p > 0.05; t-tests). Flux estimates were found to be strongly correlated with V(D) values at the tumor rim in both tumor models (R(2) = 0.84 and 0.91 for HF3016 and HF3177, respectively). These data report physiological properties of untreated GBM models that are representative of human disease both geno- and pheno-typically. Imaging biomarkers of vascular function in GBMs may aid in testing novel antiglioma therapies using these and other similar PDOX models for longitudinal, minimally invasive evaluations of treatment effects.
Medical Subject Headings
Glioblastoma; Animals; Humans; Rats, Nude; Contrast Media; Magnetic Resonance Imaging; Brain Neoplasms; Rats; Cell Line, Tumor; Male; Neovascularization, Pathologic; Female; Dynamic Contrast Enhanced Magnetic Resonance Imaging
PubMed ID
40635263
Volume
38
Issue
8
First Page
70089
Last Page
70089
