SCIDOT-38. DEVELOPMENT OF AN IMAGE-INFORMED MATHEMATICAL MODEL OF CONVECTION-ENHANCED DELIVERY OF NANOLIPOSOMES FOR INDIVIDUAL PATIENTS. (11th November 2019)

• Record Type: Journal Article
• Authors: Woodall, Ryan; Hormuth, David; Abdelmalik, Michael; Wu, Chengyue; Feng, Xinzeng; Phillips, William; Bao, Ande; Hughes, Thomas; Brenner, Andrew; Yankeelov, Thomas
• Is Part Of: Neuro-oncology. Volume 21(2019)Supplement 6
• Page Start: vi279
• Page End: vi279
• Publication Date: 2019-11-11
• Subjects: Brain Neoplasms -- Periodicals; Brain -- Tumors -- Periodicals; Brain -- Cancer -- Periodicals; Nervous system -- Cancer -- Periodicals; 616.99481
• DOI: 10.1093/neuonc/noz175.1174
• Languages: English
• ISSNs: 1522-8517

Abstract

Abstract: 186-Rhenium nanoliposomes (RNL) are an experimental theranostic being investigated for the treatment of recurrent Glioblastoma. While traditional external beam therapy exposures healthy tissue to radiation, RNL has the potential to deliver extremely large doses (> 2000 Gy) of localized radiation, minimally exposing surrounding tissue. RNL is delivered directly to the malignancy by convection-enhanced delivery (CED) via intracranial catheter. For this reason, accurate and precise delivery of RNL to the target region is an imperative. While models of CED for molecular agents exist, we know of no such models for CED of liposomal nanoparticles. To that end, we are developing a patient-specific advection-diffusion model of RNL delivery and distribution, informed by pre-delivery quantitative magnetic resonance imaging (MRI) parameters, and validated by intra-delivery single-photon emission computed tomography (SPECT). Apparent liposome diffusivity and interstitial hydraulic conductivity are spatially informed by diffusion weighted MRI, while the clearance of interstitial fluid is spatially informed by the T1 enhancement ratio after contrast agent delivery. The model output is compared to SPECT images at two time points, acquired mid-way through and immediately following the RNL infusion. At the time of submission, the model has been calibrated by patient-specific data to match the spatiotemporal distribution of RNL in four patients. After calibration, the concordance correlation coefficient between the model and SPECT measurements was 0.80 +/- 0.23 mid-way through the infusion volume, and 0.86 +/- 0.14 immediately post-infusion. The DICE coefficient between the modeled delivery volume and measured delivery volume was 0.86 +/- 0.10 mid-way through the infusion volume, and 0.81 +/- 0.14 immediately post-infusion (reported as mean +/- 95% confidence intervals). These results provide preliminary evidence that the model can capture the spatiotemporal distribution of RNL during and after delivery, and may enable physicians to better plan CED procedures for liposomal nanotherapeutics in the future.