Interpreting limits to nanoparticle delivery in high-stroma low-perfusion tumors

点击翻译按钮获取中文摘要

4.4.7 Project Summary/Abstract The central theme in this work is that critical spatial patterns exist in highly resistant cancer stroma and vascular density that inherently inhibit larger nanoparticle penetration into cancer, and that these phenotypes can be imaged in vivo. We will use in vivo diagnostic imaging, combined with ex vivo analysis to test this in pancreatic cancer, which has as well known drug penetration limitation. Specifically, we will quantify nanoparticle penetration in pancreas cancer, which has high stroma content and low vascular density. The analysis and prediction of efficacy will be quantitatively developed by methodological correlation of in-vivo and ex vivo images using Fourier spatial frequency analysis. We will determine the characteristic spatial patterns of these tumor microstructures that present as barriers to nanoparticle transport, as assayed through in vivo/ex vivo studies. We have seen that these characteristic spectral features appear in high-field magnetic resonance imaging (HF-MRI) scans and micro-Computed Tomography (uCT) scans of tumors imaged within the ongoing nanoparticle project at DHMC. The scope of this project is to conduct a secondary analysis on the images that are being produced within these projects, with two specific aims. 1) We will directly correlate nanoparticle penetration and distribution to the Fourier spatial frequencies found in in vivo images by Fourier spatial frequency analysis in which we have demonstrated expertise. The in vivo images will be analyzed by correlating them with histological sections of nanoparticle distribution post-treatment. Tumors will be classified on two levels as either a high or low permeability to a specific nanoparticle formulation (to quantify the amount of agent delivered), and as having high or low isotropy (to quantify the dispersion of the agent). 2) We will then apply this characteristic morphology analysis to pre-treatment, pre-operative HF-MRI, uCT images, and analyze their value as a potential diagnostic classifier. We will use a Support Vector Machine Analysis to predict the permeability and isotropy of unknown tumors, and validate our results against experimental outcomes. An iterative strategy will optimize the predictive power of the method, and be used to distinguish between characteristic spectra that are good and bad classifiers. The research will be produced using the unique software systems that we have designed during preliminary studies, and will be deployed on an analysis platform that can be integrated with the hospital- based DICOM and virtual pathology environment to allow clinical investigators to plan adjuvant therapies to promote nanoparticle efficacy. Several hundred high-quality scans are now available for analysis, which will be processed and reported on within the first year of funding. By year two, the established system is projected to be able to analyze images within a few minutes post-scan. These analysis methods will give us the key background needed to advance our fundamental understanding of nanoparticle in-vivo delivery, and test ways to interrupt transport barriers in interventional future work.