Characterization of a First-Pass Gradient-Echo Spin-Echo Method to Predict Brain Tumor Grade and Angiogenesis
Kathleen M. Schmaindaa,b,
Scott D. Randa,
Allen M. Josephb,d,
Rebecca Lundb,d,
B. Doug Wardb,
Arvind P. Pathakb,d,
John L. Ulmera,
Michael A. Baddrudojac and
Hendrikus G. J. Krouwerc
a Department of Radiology, Medical College of Wisconsin, Milwaukee, WI
b Department of Biophysics, Medical College of Wisconsin, Milwaukee, WI
c Departments of Neurology and Neurosurgery, Medical College of Wisconsin, Milwaukee, WI
d Department of Biomedical Engineering, Marquette University, Milwaukee, WI
View larger version (58K):
[in a new window]
|
FIG 1. Sample SE rCBV map shows hot-spot ROIs (black outlines and arrows), which represent the highest approximate 10–20% of microvascular rCBVs. These were used to mask the image maps and perform hot-spot analysis.
| |
View larger version (58K):
[in a new window]
|
FIG 2. T1-weighted contrast-enhanced images (a and e) GE rCBV maps (b and f), SE rCBV maps (c and g), and ratio maps (d and h) obtained in a 53-year-old man with an anaplastic oligodendroglioma (a–d) and a 55-year-old man with a glioblastoma multiforme (e–h).
| |
View larger version (17K):
[in a new window]
|
FIG 3. Whole-tumor ROI analysis for gliomas.
A, GE rCBV, which is sensitive to total blood volume, is significantly correlated with tumor grade (n = 72).
B, Conversely, SE rCBV, which is sensitive to microvascular blood volume, is not significantly correlated with tumor grade (n = 67).
C,  R2*/R2 ratio, an index of mVD, is significantly correlated with tumor grade (n = 67).
| |
View larger version (17K):
[in a new window]
|
FIG 4. Microvascular hot-spot ROI analysis for gliomas.
A, GE rCBV results maintain a significant correlation (n = 67).
B, Contrary to whole-tumor results, SE rCBV shows a significant correlation with tumor grade (n = 67).
C, While a significant correlation with tumor grade is retained, the correlation is weaker than that of whole-tumor ROIs (n = 67).
| |
View larger version (26K):
[in a new window]
|
FIG 5. SE rCBV data as a function of GE rCBV data. The K nearest-neighbor analysis (K = 5) resulted in correct classification of five (69%) of 16 grade I–II tumors (circles) and 49 (95%) of 51 grade III–IV tumors (squares). Asterisk indicates misclassified data points; the arrows indicate two low-grade tumors where diagnosis was based on biopsy.
| |
View larger version (86K):
[in a new window]
|
FIG 6. MRI-derived (a) post-contrast (b) GE rCBV, (c) SE rCBV and (d) ratio maps obtained in a 66-year-old female patient prior to CT-guided biopsy. The biopsy pathologic results suggest low-grade astrocytoma. This diagnosis is not consistent with either the rCBV results or the clinical course of this patient.
| |
