cni » projects » image analysis » guidelines to identify...
guidelines to identify gadolinium enhancing ms lesions
(Barkhof F, et al. Improving interobserver variation in reporting gadolinium-enhanced MRI lesions in multiple sclerosis. Neurology 1997;49(6):1682-1688.)
An enhancing lesion is a well-demarcated area of unequivocally increased signal intensity on T1-weighted spin-echo (SE) images compared with normal-appearing, surrounding brain tissue following intravenous injection of a gadolinium chelate. Although size definitions are difficult to give, tiny foci (1 pixel) should be excluded, and the signal increase should not be related to a normally enhancing structure (such as vessel or choroid plexus), and not identified as a flow-related artifact.
Although gadolinium enhancement can occur in isolation, preferably it should be accompanied by an abnormality on corresponding PD/T2-weighted images. T2 confirmation is mandatory for difficult areas, such as the posterior fossa (with many flow artifacts and confusing tentorial vessels) and the temporal lobes.
A new enhancing lesion is defined as an area of enhancement in an place that showed no enhancement on a previous scan. Most convincing is a new T2 abnormality on serial scans, but re-enhancement can of course occur in a preexisting T2 lesion, although it is considered less persuasive. A persistently enhancing lesion is defined as area of enhancement in a place that did show enhancement on a previous scan. Usually the accompanying T2 abnormality persists, but can change in size (e.g., enlarge).
Low levels of enhancement can be difficult to interpret. Care should be taken to avoid too small a window in preparing the hard copies. In the vicinity of lesions that are hypointense on T1-weighted SE images, enhancement might be falsely suggested.
With the knowledge that an enhancing lesion was seen in a particular area on a previous scan, it can sometimes be difficult to decide whether very slight persistent enhancement is still present in that location at follow-up. One should essentially use the same criteria as for new lesions: Would you identify it as a new enhancement? Because MS lesions frequently arise in close proximity, differentiating new from persistent enhancements can be difficult, especially when the repositioning of the imaging slices is not optimal. One can try to measure the distance from the midline or superimpose the images (when the images have the same zoom factor). As an alternative, one could try not to separate new and persistent lesions, but simply to register all areas of enhancement as activity.
Even using these definitions, there are many circumstances that will interfere with the interpretation of enhancing lesions:
- MS lesions are frequently seen around the ventricles, but care should be taken not to include subependymal veins. These structures differ from MS lesions in that they are longitudinal, thin, and frequently show signal void on PD/T2-weighted images. Likewise, in the basal ganglia, Virchow-Robin spaces with enhancing vessels should be ruled out.
- The thickness of the imaging slices is usually 3 mm, which allows partial volume averaging between brain tissue and other structures, such as the choroid plexus. Erroneous interpretation of the choroid plexus is most likely to occur at the level of the trigone of the lateral ventricle. To avoid confusion with the choroid plexus in adjacent slices, preferably one normal slice in between an area of enhancement and the choroid plexus should be present.
- Flow-related artifacts become more conspicuous after gadolinium injection and are particularly troublesome in the posterior fossa. In this area PD/T2-weighted images should always be abnormal. This is of course at the risk of either a false-positive decision because of a preexisting T2 abnormality or a false-negative in the case of no T2 abnormality, yet.
- Brainstem lesions can be small and difficult to recognize. The brainstem should be scrutinized separately, and even very small lesions should be considered when the PD/T2-weighted image is abnormal. The medulla oblongata is even more difficult. There should be no flow artifacts, and a definite T2 abnormality is mandatory. In addition, the area of enhancement should either be very clearly demarcated or surrounded by a clear rim of normal brainstem tissue.
- In the case of enhancement in the cortical/subcortical area, it might be difficult to confirm the presence of a lesion on PD/T2-weighted images with conventional SE imaging. Enhancement should be seen ideally only if visible on two consecutive slices. Specifically, if a sulcus is present on the slice immediately above or below, the abnormality should be disregarded if it is very small and only seen on one slice.
- Special care should be paid to the basal frontal area, where false-positive abnormalities can be induced by the sinuses due to either susceptibility effects at the interface with aerated sinuses or due to partial volume averaging in case of fluid-filled sinuses.
- Ring enhancement is frequently seen in MS lesions. Identification of enhancement in such lesions can be difficult when only parts of an incomplete ring are enhancing. If such areas of enhancement from parts of an incomplete ring, whereas the PD/T2-weighted image shows only one lesion, it can best be considered a single enhancing lesion. However, if the PD/T2-weighted image shows two contiguous lesions and two homogeneous areas of enhancement are seen, they should be considered as two separate enhancing lesions, even if they are contiguous.
Definition of T1 Hypointense Lesions
(van Walderveen MA, et al. Hypointense lesions on T1-weighted spin-echo magnetic resonance imaging: relation to clinical characteristics in subgroups of patients with multiple sclerosis. Arch Neurol 2001;58(1):76-81.)
T1 lesions were defined as regions with a signal intensity similar to or lower than the signal intensity of gray matter and corresponding to a hyperintense region on PD/T2-weighted MRI.
For more information consult the following manuscripts:
- Van Walderveen et al. Hypointense Lesions on T1-Weighted Spin-Echo Magnetic Resonance Imaging, Arch Neurol. 2001;58: pp.76-81 (download PDF)
- Barkhof F et al. Improving interobserver variation in reporting gadolinium-enhanced MRI lesions in multiple sclerosis. Neurology 1997 Volume 49(6).December 1997.pp.1682-1688 (download PDF)