| Radiologic findings of the head and spine in neurofibromatosis 1 (NF1) in Northern Finland | ||
|---|---|---|
| Prev | Chapter 2. Review of the literature | Next |
Since the introduction of CT, and especially MRI, CNS abnormalities in NF patients have been shown to be much more common than it had been thought earlier. Gardeur et al. (1983) found normal CT scans in only 17 out of 77 NF patients, including both NF1 and NF2. Jacoby et al. (1980) found normal CT in 8 out of 29 NF patients, while 21 (72%) had abnormalities of the orbits or brain. Gayet-Delacroix et al. (1995) presented the MRI findings of 40 asymptomatic NF1 patients (mean age 22), pointing out that MRI demonstrated various abnormalities in 24 patients (60%). Shu et al. (1993) found abnormalities in 66 (94%) of their 70 NF patiens on cranial MRI examinations.
MR has been recommended as the method of choice for examining the head and spine in NF patients (Menor et al. 1991). It has been found better than CT in detecting and defining many brain and spine lesions, especially ones in the posterior fossa and spinal cord. MRI has been pivotal in the detection of T2 hyperintensities in the brain, and intramedullary tumours in NF have become more easily visualized. The use of contrast agents in MRI has further improved the detection of lesions in NF1. Gadolinium-DTPA (Gd-DTPA) is over 25 times safer than iodinated contrast agents. The NIH conference report of 1990 (Mulvihill 1990) recommends the use of intravenous contrast enhancement in MRI examinations of patients with neurofibromatosis.
There are only a few real comparisons concerning the sensitivity of CT and MRI.
Menor et al. (1991) compared CT and MRI in an evaluation of the incidence of CNS lesions in 41 children with NF1. CT detected 32% globus pallidus lesions which showed hyperintensity on T2-weighted MR images. On T2-weighted images, 51% of the patients showed cerebellar and 42.5% brainstem hyperintensities, which were not visible on CT. One of the four brain tumours did not show on CT, and two cases of Chiari I malformations remained undetected on CT. Involvement of the chiasm or optic radiation were better seen on MRI.
The MRI imaging protocols have varied in different studies. Menor et al. (1991), for example, recommended for brain imaging a MRI protocol consisting of an initial sagittal T1-weighted SE and axial T2-weighted SE sequences as well as a T1-weighted SE study parallel to the hard palate, which is optimal for visualizing the optic pathways. According to them, gadolinium should be reserved for the evaluation of aqueductal stenosis, the posterior extent of the optic pathway and parenchymal tumours and lesions with a mass effect or perilesional edema and hypointensity with a tumour suspicion on T1 sequences. In the case of a typical NF lesion, yearly MRI control was recommended, and controls at shorter intervals were to be made only if clinically needed. Lesions with gadolinium uptake should be biopsied or controlled with MRI within less than six months.
There are several advantages in the use of MRI in the imaging of the optic pathways: the intracranial extent of the optic tumour is easier to detect, the anatomic delineation of the optic glioma is visualized better, other than optic pathway lesions are easily detectable on the same examination, no ionizing radiation (or i.v. contrast) is needed, bone and nonferromagnetic metallic implants produce no artifacts, as they do on CT, and direct coronal, sagittal and oblique imaging is possible (Pomeranz et al. 1987). Parazzini et al. (1995) recommended MRI screening for patients with NF1 to detect clinically silent optic pathway lesions and serial MRI studies to follow up the evolution of these lesions. No aggressive treatment should be undertaken without follow-up, and surgical treatment should be limited to cases that are clinically and radiologically in progression. Listernick et al. (1999), on the other hand, emphasized the importance of yearly ophthalmological examinations of a child with NF1 and regarded radiological screening for optic pathway gliomas as unnecessary.
Balestri et al. (1993) were able to show that the likelihood of detecting imaging abnormalities in patients with NF1 is increased by systematic follow-up. They found three cases of optic glioma, six cases of T2 hyperintensities and a suspected parenchymal tumour in one case at the initial exmination. After follow-up for two to four years, they found three more cases of optic glioma and T2 lesions in nine more patients. The authors concluded that systematic follow-up MRI examinations are indicated in NF1 patients, but invasive treatment should be avoided and used only in rapidly progressive and symptomatic lesions.
The NIH Consensus Group on Neurofibromatosis (1988) regarded radiological examinations of asymptomatic NF patients as unnecessary. Nor did Guttmann et al. (1997) recommend routine CT or MRI screening of the CNS of individuals with NF1, whereas patients at risk for NF2 should undergo gadolinium-enhanced MRI of the head and the whole spine, to rule out vestibular schwannomas and spinal tumours.
Broniscer et al. (1997) analyzed the proton magnetic resonance spectroscopic (MRS) findings of patients with diffuse pontine glioma (PG) with and without NF1 and healthy children. The neuronal markers N-acetyl aspartate (NA) and the vector sum of the metabolites choline and creatine/phosphocreatine (M-CHO-CR) were compared in different groups. The mean NA levels and the mean M-CHO-CR levels did not differ significantly between NF1 patients and controls. Patients in the PG group had significantly lower NA and M-CHO-CR levels than did patients in the control and NF1 groups. Patients with NF1 and symptomatic brainstem enlargement had significantly higher (MRS) values for the variables studied compared to patients with PG. The authors presume the higher NA values to represent preservation of brainstem neuronal elements in NF1 patients, consistent with fewer symptoms upon brainstem involvement, and the results may be helpful in planning appropriate therapy.
Gonen et al. (1999) studied the brains of NF1 children and healthy control subjects with three-dimensional multivoxel proton MR spectroscopy and found distinct metabolic features in normal brain and T2 hyperintense lesions and tumours in children. The T2 hyperintense lesions were characterized by significantly elevated choline (Cho), reduced creatine (Cr), 2 > Cho:Cr > 1.3 and nearly normal N-acetyl aspartate (NAA) levels. The tumours had Cho:Cr > 2 and no NAA. The T2 hyperintense lesions had no intrinsic lipid or lactate signals, and they were more extensive than shown by MRI.
Wilkinson et al. (2001) studied NF1 children with typical T2 hyperintense brain lesions and lesions with atypical or tumorous features and found that MR proton spectroscopy can help to differentiate between T2 hyperintense brain lesions and brain (non-optic/hypothalamic) glioma. They found a significant increase in choline and myo-inisitol in tumours compared to typical T2 hyperintense brain lesions or controls, a reduction in the N-acetyl levels in T2 hyperintense lesions compared to controls, a reduction in N-acetyl in tumours compared to controls and a reduction in glutamate/glutamine in tumours compared to controls. The MR proton spectroscopy findings of Jones et al. (2001) suggested that metabolic changes may be present without visible changes on MRI.
Sener et al. (2000a) and Eastwood et al. (2001) evaluated changes in brain water diffusibility by MRI in different areas of the brain in children with and without NF1. Eastwood et al. measured significant ADC (apparent diffusion coefficient) increases both in hyperintense lesions and in normal-appearing areas of the brain in children with NF 1. Sener et al. obtained higher ADC values in optic radiations affected by glioma than values of normal white matter, suggesting relatively high molecular motion in the affected areas, probably due to myelin vacualization. ADC values of thalamic hamartomas were lower, and it might be possible to differentiate optic gliomas from hamartomas.
Balestri et al. (1994) presented four patients with NF1 who underwent cerebral PET with (18F)-2-fluoro-2-deoxy-D-glucose. Widespread hypometabolism was found in three of them. The lesions on MRI, which were localized in the subcortical white matter and grey matter structures had normal rates of glucose metabolism. This suggested that the lesions seen on MRI were not due to defective blood supply