Panagiotis Korfiatis PhD, Presenter: Nothing to Disclose

Leland S. Hu MD, Abstract Co-Author: Nothing to Disclose

Zachary Samuel Kelm BS, Abstract Co-Author: Nothing to Disclose

Bradley J. Erickson MD, PhD, Abstract Co-Author: Stockholder, Evidentia Health, Inc

Cerebral blood volume (CBV) is an emerging magnetic resonance imaging biomarker used to differentiate between progression and pseudo progression in patients with glioblastomas (GBM), a challenging task even for experts. Dynamic susceptibility-weighted contrast (DSC) imaging is the most commonly used clinical technique for measuring CBV in brain tumors. However, accurate CBV quantification is difficult especially in cases of GBM, where the blood-brain barrier has been interrupted and contrast leakage occurs. Several software tools exist, allowing for semi-automatic calculation of relative CBV. Some of these software tools incorporate mathematical techniques to deal with contrast agent leakage. Assessing the accuracy and reliability of leakage correction methods can be achieved only with the use of an appropriate phantom. However, the development of digital phantoms that simulate contrast extravasation is an open research issue. The purpose of the current study is the development of a DSC-MRI digital phantom containing regions with a known amount of simulated contrast leakage, allowing evaluation of software tools.

The created phantom was successfully able to simulate contrast leakage effects on the time-concentration curves.

An algorithm was developed, capable of producing a 4D digital phantom containing areas simulating white matter, gray matter, cerebrospinal fluid, and 4  regions with varying levels of contrast leakage. The time-concentration curves for each pixel were simulated based on the signal proposed by Liu et al1. The model utilized can describe the combined T1 and T2* effects from contrast agent leakage in the measured signal. The algorithm can produce datasets simulating various image acquisition parameters, such as TE, relaxation rate, and mean transit time, while also allowing addition of Gaussian noise.

The proposed digital phantom may be used in performance evaluation of clinically available software tools and may further enable the development and validation of algorithms aimed at CBV quantification leakage correction. References 1.Liu, H-L, et al. Med. Phys. 38: 2 (2011): 802–809.

Korfiatis, P, Hu, L, Kelm, Z, Erickson, B, A DSC Digital Brain Phantom for Assessment of Leakage Correction Methods.  Radiological Society of North America 2013 Scientific Assembly and Annual Meeting, December 1 - December 6, 2013 ,Chicago IL. http://archive.rsna.org/2013/13024766.html