Dr. Meng Law of the departments of radiology and neurosurgery at Mount Sinai Medical Center in New York and his colleagues have been investigating the use of dynamic susceptibility-weighted contrast-enhanced perfusion MRI to gather physiologic information about vascular endothelial proliferation, vascular density, and angiogenesis. In particular, they hypothesized that this technique can provide a means of characterizing tumor biology and predicting tumor progression. They retrospectively evaluated whether relative cerebral blood volume (CBV) measurements could be used to predict clinical outcomes in patients with malignant high-grade gliomas and low-grade gliomas. Specifically, they looked at whether patients who have gliomas with high initial relative CBV have more rapid progression than do those who have gliomas with low relative CBV.

Dynamic susceptibility-weighted contrast-enhanced perfusion MRI takes advantage of signal changes that take place with the passage of paramagnetic contrast agents—such as gadopentetate dimeglumine—through the cerebrovasculature. Dynamic susceptibility-weighted contrast-enhanced perfusion MRI results in a drop in signal intensity due to the susceptibility of the gadolinium that is proportional to the blood volume.

The technique “can be utilized and translated to the clinic pretty readily and that really provides us with a new way to predict tumor biology—one that is much needed, given the limitations that we have” with current classification systems, said Dr. Law.

The study included 189 patients (65% male, mean age 43) with pathologically proved gliomas, using the World Health Organization four-tier classification of gliomas (Radiology 2008;247:490–8). Patients were referred for preoperative assessment of intracranial tumors. They could not have any evidence of systemic malignancy or immune status compromise. In all, 28 patients had low-grade fibrillary astrocytomas (WHO II), 11 had low-grade oligoastrocytomas (WHO II), 14 had low-grade oligodendrogliomas (WHO II), 72 had anaplastic astrocytomas (WHO III), 12 had anaplastic oligoastrocytomas (WHO III), and 52 had glioblastoma multiforme (WHO IV).

Patients were followed up a median of 334 days and assessed clinically and with MRI—1.5 T conventional, single-dimension measurements of contrast-enhanced T1 enhancement and T2 signal hyperintensity (for tumor size), and serial relative CBV measurements.

Each patient was assigned to one of four response categories, based on clinical chart review and MR findings: complete response (4 patients), stable disease (41 patients), progressive disease (130 patients), and death (14 patients). Complete response was defined as no visible tumor on MRI and no new neurologic deficit. Stable disease was defined as no change in the patient's neurologic examination results and Karnofsky score, and less than 25% change in tumor size on MRI. Progressive disease was defined as a decline in neurologic status and Karnofsky score, or an increase in tumor size of more than 25% on MRI. Patients were assessed at 3-month intervals by their neuro-oncologist. MRIs were performed at the same time.

Dynamic susceptibility-weighted contrast-enhanced perfusion MRI does not give an absolute measure of CBV. Instead, CBV in the area of interest is expressed as a ratio relative to the CBV measured in standard tissue—typically normal contralateral white or gray matter. The researchers developed color overlay maps of relative CBV. Regions of interest were placed in regions of greatest perfusion on the color overlay maps for each patient. A constant radius of 3.6 mm was used for all regions of interest. Four separate CBV measurements were made in these regions of interest and the maximal value was recorded.

The researchers calculated means, standard deviations, and medians of relative CBV measurements in the regions of interest for all patients in a clinical response category. Mean relative CBV values were 1.41, 2.36, 4.84, and 3.82 for the complete response, stable disease, progressive disease, and death groups, respectively.

Patients were also classified in groups with low or high relative CBV, using a threshold of 1.75. Dr. Law and his colleagues previously identified this threshold value to provide optimal sensitivity and specificity for differentiating low-grade gliomas from high-grade gliomas in a study of 160 patients (Am. J. Neuroradiol. 2003;24:1989–98).

Median time to progression for patients with relative CBV values less than 1.75 was 3,585 days. In comparison, median time to progression for patients with relative CBV values greater than 1.75 was 265 days, regardless of histopathologic tumor type. Use of the 1.75 threshold was significantly associated with time to progression among all patients, with or without adjustment for pathologic status. Age and relative CBV—but not gender—were significant predictors of disease progression and death, based on binary logistic regression. However, using the 1.75 CBV threshold was not significantly associated with survival.

The relative CBV measurement might provide an important imaging biomarker of glioma malignancy that could potentially affect therapeutic choices.

70187_fx1.sml

Glioma seen on conventional MRI (arrow) did not progress with time, indicated by susceptibility-weighted contrast-enhanced perfusion MR images (color slices). Courtesy Dr. Meng Law

Image of the Month