| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Letters to the Editor |
Institute of Radiology and Interventional Diagnostics, Tbilisi, Georgia
Bethel Epilepsy Center, Mara Hospital, Maraweg 21, D-33617 Bielefeld, Germany, e-mail: fgw@mara.de
Editor:
We read with interest the article by Dr Hirai and colleagues in the May 2000 issue of Radiology (1) on the evaluation of the human limbic lobe with the turbo fluid-attenuated inversion-recovery (FLAIR) sequence. In the past, FLAIR imaging proved to increase sensitivity for the detection of hippocampal sclerosis, cortical lesions bordering cerebrospinal fluid spaces, and white matter lesions. It is widely used in magnetic resonance imaging protocols for epilepsy (2). At the same time, FLAIR images are not free of artifacts and may show inhomogeneity of signal intensity in neurologically normal control subjects (1). The visual detection of the difference in signal intensity between cortical structures might lead to false-positive or false-negative results (eg, in patients with bilateral symmetric lesions). It is of great clinical importance to distinguish artifacts from the pathologic changes in signal intensity.
The visual assessment and the evaluation of contrast-to-noise ratio used by Dr Hirai and colleagues (1) revealed higher signal intensity in the gray matter of phylogenetically older structures (the term limbic lobe seems not precise enough, because the insula is not anatomically part of this lobe).
We tried to replicate these findings by measuring transverse relaxation time in different cortical structures and the amygdala in healthy subjects. Ten healthy control subjects (six women, four men; age range, 27–48 years; mean age, 39.6 years) were examined. T2 maps were calculated from coronal images (3,000/22.5–360 [repetition time msec/echo time msec], 5-mm section thickness, eight sections acquired) obtained with a 16-echo Carr-Purcell-Meiboom-Gill sequence by using a 1.5-T system (Magnetom Symphony; Siemens, Erlangen, Germany). Regions of interest were manually traced to avoid boundaries with white matter and cerebrospinal fluid and placed bilaterally in six structures: hippocampus, parahippocampal gyrus, cingulate gyrus, insula, amygdala, and frontal cortex. Statistical differences were tested by using analysis of variance with post hoc tests. Mean T2 data are presented in the Table. There was no statistically significant difference between the T2 time of the frontal cortex and that of the phylogenetically older cortical structures and amygdala. Our quantitative results failed to help confirm visually detected and thus relatively gross differences.
Dr Hirai and colleagues discussed results of a T2-relaxometry study by Whittall et al (3) that supported their findings. However, data by Whittall et al seem to have partial volume effects, as they readily concede: "...low cortical grey values were probably caused by the thick slice (10 mm) averaging adjacent white matter. ..." (3).
Is it possible that the higher signal intensities Dr Hirai and colleagues describe for structures like the hippocampus are the result of coronal FLAIR images obtained perpendicular to the long axis of hippocampus and that they solely represent higher inclusion of gray matter per voxel? Did the authors evaluate the phenomenon published here in the transverse and sagittal planes?
|
Department of Radiology, Amakusa Medical Center, 854-1 Kameba, Hondo, Kumamoto 863-0046 Japan* e-mail: toshinor@beige.ocn.ne.jp Department of Radiology, Kumamoto University School of Medicine, Kumamoto, Japan
We thank Drs Okujava and Woermann for their interest and questions about our article (1). We would like to respond to their comments as follows.
It is true that FLAIR images are not free of artifacts due to cerebrospinal fluid inflow effects and so on, and sometimes they fail to demonstrate brain lesions located in the basal ganglia and brainstem (2). However, in the cerebral cortices evaluated in our article, there seem to be fewer artifacts than in the locations just mentioned.
Drs Okujava and Woermann have measured the transverse relaxation time in cortical structures, but they have not found a statistically significant difference between the frontal cortex and the phylogenetically older cortical structures. According to their data, we suppose that the T2 relaxation time may not affect the difference in the signal intensity of the cerebral cortices on FLAIR images. The inversion pulse in FLAIR sequences introduces considerable T1 weighting, which acts antagonistically to the T2 contrast (3). Thus, FLAIR images do not merely reflect T2 relaxation time but are also influenced by T1 relaxation time. The T1 contrast of the cerebral cortices might have a greater effect on the difference of the cortical signal intensity seen on FLAIR images.
The other explanations regarding signal intensity of the limbic cortical areas are cellular composition and vascularization of the cortex, which were described in our article (1). Among the cortices of the limbic lobe seen on FLAIR images, the hippocampus had the highest signal intensity. The hippocampus has the simplest and most primitive parts of the cortex and different vascular networks, as compared with the more complex and developed isocortex. We think that these factors are probably associated with cortical signal intensity seen on FLAIR images.
Drs Okujava and Woermann also mentioned imaging planes. Although we did not evaluate the FLAIR images at other planes in the same patient, we believe that the coronal plane is the most suitable for evaluating the structures of the limbic lobe, because the coronal plane is generally perpendicular to the long axis of these structures and the partial volume effect should be smallest in this plane.
REFERENCES
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| RADIOLOGY | RADIOGRAPHICS | RSNA JOURNALS ONLINE |
