Improved tumor diagnosis with BOLD and 1H-MRS imaging

Improved tumor diagnosis by measuring tumor blood content and metabolism combining MR imaging and spectroscopy methods at 3 Tesla Primary brain tumors are recognized and characterized by the presence of pathological vascularity and metabolism, however, in a very heterogenous fashion. This limits both established diagnostic approaches, namely conventional MRI with contrast agents, and brain biopsy. In addition, both method are invasive and, thus, bear some discomfort and risk for the patient. Therefore, the ability to non-invasively image both the vasculature and metabolism with high spatial resolution in tumors is a key component in diagnosing their presence, differentiating tumors from otherwise similar appearing pathological processes (e.g. brain abscess), characterizing their stage of development and assessing their activity levels. In this exploratory proposal, we present (a) the application of venographic imaging in detecting disease to evaluate the tumor blood pool and (b) the application of ultrahigh-resolution, high field 1H-spectroscopic imaging to evaluate the distinct metabolic patterns of neoplastic (i.e., active tumor of different stages, necrosic, edema) vs. non-neoplastic (i.e., abscess, dernyelination etc.) processes. Specifically, venographic imaging is based on the blood oxygen level dependent effect (BOLD), commonly used in functional MRI, which is related to the local oxygen utilization to detect venous microvasculature within tumors. This method has already been shown to be an excellent way to detect small veins and abnormal venous microvasculatur in other diseases. Complementary, 1H-MR spectroscopy and spectroscopic imaging have been applied to differentiate brain vs. tumor metabolism in research and clinical work for more than a decade, however, tissue and tumor heterogeneity prevented a widespread clinical use as clinical scanners (max. 1.5 T) do not allow sufficiently high spatial resolution for spectroscopy (commonly 3-8 ccm in single voxel and 0.75-1 ccm in multivoxel techniques). Recent developments in MR hard- and software enabled us to achieve imaging resolution (min. 3x3x5 mm) together with excellent spectral resolution (3-5 Hz in vivo) also in ultra-high resolution proton spectroscopic imaging (UHRPSI) at 3 T. Due to the strongly reduced magnetic and tissue heterogeneity, both, sensitivity (SNR(NAA) > 6) and specificiy (high resolution metabolic patterns from 50 -220 microl (0.05 - 0.22 ccm) voxels) are outstanding for the unsurpassed spatial resolution (i.e., 5-150 times the currenty available resolution). We believe that these techniques will be valuable in the early detection of (small) brain tumors and may help differentiate malignant from less agressive tumors or brain abscesses. The results from this study will further research into studies of angiogenesis and the treatment of tumors. Further, as 1H-MRS is already approved as a clinical tool by the FDA and 3 T MR-systems are on the verge to become clinical soon, our research might become integrated into clinical protocols for tumor diagnosis and theraphy control in the near future.