MAGNETIC RESONANCE IMAGING (MRI) AND MAGNETIC RESONANCE SPECTROSCOPY (MRS) OF INTRACRANIAL LIPOMAS

R. Fründ, A. Geissler, M. Gliese, J. Seitz, S. Feuerbach

Rüdiger Fründ Institut für Röntgendiagnostik Klinikum der Universität Regensburg, D-93042 Regensburg Germany

Received 6/3/97 Accepted 9/16/97

4. RESULTS AND DISCUSSION

4.1. Imaging

figures 1 and 2 show a T1 weighted image of patient A with and without fat suppression. MR-imaging using fat suppression pre-pulses results in good signal extinction for the orbital region, subcutaneous fat and the lipoma in the area of the brain stem. Therefore lipomas can be easily detected. In addition fibrotic tissue and vascular structure of the lipoma can be seen in more detail with fat suppression for intracranial lipomas. Similar attempts with good results were made for tumors of the adrenal gland in order to distinguish benign fat tissue from malignant tissue (9). For the cases described here, the method can be used for a specific diagnosis of the tumor. This is most helpful in the differential diagnosis of fat, small amounts of blood and small lipomas.

figure 1: Axial T1 weighted brain images of patient A . Left without fat suppression, right with fat suppression, arrow indicates lipoma.

figure 2: Sagital T1 weighted brain images of patient B without fat suppression.

4.2. Spectroscopy

Voxels placed in the lipoma showed strong signals at 2.15, 1.3 and 0.9 ppm (figure 3). The signal at 1.3 ppm can be assigned to CH₂ groups in long alkyl chains. The resonance at 0.9 ppm stems from CH₃ groups of these chains. The broad signal around 2.15 ppm can be assigned to CH₂ groups in the neighbourhood of unsaturated CH-groups and to CH₂-groups next to the carboxygroup in carbonic acid. Due to the slightly better resolution of the ex vivo spectra, an additional peak at 1.6 ppm is visible. In the spectrum without water suppression the CH=CH groups of unsaturated carbonic acids was resonating at 5.3 ppm (figure 4). No signals of normal human brain metabolites (10) like N-acetylaspartat, choline, creatine or myo-inostitol were detected. The in vivo measurements of normal subcutaneous fat did not differ from intracranial lipomas. However, lipid signals often seen in high grade glioblastomas had lower CH₂ to CH₃ intensity ratios. The ex vivo measurements differs not significantly from the in vivo measurements. Only the resolution is slightly better for the ex vivo measurements. In table 1, the ratios for the signal intensities at 1.3 and 0.9 ppm of different tissues are shown.

In contrast to MR-imaging, MR-spectroscopy gave more detailed information about the chemical composition of the examined tissues. Lipids in human adipose tissue mainly consist of tri-acylglycerol. Themost frequent fatty acids are oleic acid (39 %), palmitic acid (26 %) and linolic acids (15%) (11). MR-spectroscopy can determine the amount of CH₃ and CH₂ groups in these fatty acids (12). Methyl groups at the end of the fatty acids chain induce a shift of 0.9 parts per million (ppm). Methylene groups in the middle of the chain introduce a shift at 1.3 ppm (12). The ratios of the peak areas indicate the length of saturated alkyl chain. Table 1 shows similar ratios for the two lipomas and for the subcutaneous fat. Therefore, the length of the chain of fatty acids in lipomas and subcutaneous fat must be in the same range and should be structuraly similar. The calculated theoretical ratio of intensities for the signals at 1.3 ppm and 0.9 ppm of olefinic acid is 6.7:1 because only 10 CH₂ groups have a resonance around 1.3 ppm. The other CH₂ groups are either in the vicinity of the carboxy group resulting in a resonance frequency of 2.36 ppm for the closest CH₂ group and 1.65 ppm for the following CH₂ group or 2.07 ppm in the neighbourhood of the ethylene group (12). This calculated ratio is very close to the measured value of lipoma from patient B and within the limits of measurements for lipoma from patient A and the subcutaneous fat.

In comparison, the ratios found for glioblastomas are much smaller, indicating shorter hydro-carbon chains for the lipid-like substances in tumor tissues. If all these substances consist of saturated fatty acids, the MR-data suggest a C8:0 type. The reduced linewidth found for the lipid resonances in glioblastomas support the idea of a shorter alkyl chain length in these tissues due to a higher mobility of molecules.

The unsaturated parts of the chains are visible in the in vivo spectra of lipomas and fat tissue acquired without water suppression. figure 4 shows an extra resonance at 5.3 ppm originating from unsaturated parts of the chains (12, 13).

Our ex vivo data confirm the in vivo results (Table 1). Taking into account the long repetition time and the missing echo time, the ex vivo values were not effected by T1 and T2, suggesting that the in vivo ratios were not influenced by this two relaxation times either.

MRS allows the in vivo evaluation of the structure of intracerebral lipoma tissue and may be helpful for the differentiation of benign and malignant tumors.