Abstract
Article abstract-Lhermitte-Duclos disease (LDD) is a disorder sometimes referred to as a dysplastic gangliocytoma of the cerebellum. This is a focally indolent growth of the cerebellar cortex in which the folia enlarge due to a profusion of dysplastic cortical neurons and a thickening of the molecular layer. Loss of Purkinje cells and thinning of medullary white matter results. The enlarged folia lose their secondary foldings and asymmetrically expand the cerebellar hemisphere. These morphologic features produce a characteristic pattern on some CTs and all MRIs, affording an opportunity for a preoperative diagnosis. MRI of a pathologic specimen suggested that the abnormal T1 and T2 signals corresponded to the atrophic folial white matter, thickened granule cell layer, and outer molecular layer. Because of inherent Hounsfield artifact in posterior fossa with CT, MRI is the imaging modality of choice, with better visualized striated pattern. The uniqueness of these imaging features obviates the need for an obligatory biopsy for asymptomatic patients and either permits more definitive planning for surgical decompression or, in restricted lesions, guides more assuredly complete excision of the cerebellar mass.
NEUROLOGY 1997;48: 725-731
Lhermitte-Duclos disease (LDD), dysplastic gangliocytoma of the cerebellum, is a cerebellar mass that typically presents in young adults with symptoms of increased intracranial pressure. Until recently, diagnosis of the cerebellar lesion was made only postoperatively. The cerebellar folia are grossly enlarged and histologically are characterized by hypertrophic granular cell neurons, hypermyelination of the molecular layer, Purkinje cell loss, and atrophy of white matter within the affected folia. [1,2] Cerebellar folia enlarge up to a width of 7 mm. [3] For the past decade, a spate of case reports have described a consistent appearance on MRI. [4-8] We present six patients with LDD, summarize their clinical findings, and analyze the characteristic features found on CT and MRI. To improve consistency in diagnosis and draw further attention to this fascinating condition, we have reviewed the clinical features and biological behavior of LDD hoping to refine the approach to and management of this posterior fossa mass.
Methods.
We have reviewed medical records, six MRIs, and five CTs of six patients with histologically confirmed LDD.
Multiple MRIs were made of an available postmortem specimen from patient 6, with various T1 and T2-weighted spin-echo and gradient-echo sequences. Sections were 3-mm thick and were obtained with a standard knee coil as well as with a quadrature head coil. The specimen was submerged in 10% neutral buffered formalin in a small container with small plastic rods in between the enlarged folia to demarcate the cortical surface. The postmortem specimen and the postmortem scans were correlated.
Results.
The major clinical and neuroimaging findings of these patients are collated in Table 1. CTs and MRIs both demonstrated abnormal laminated patterns of cortical architecture within the mass lesions of LDD (Figure 1 and Figure 2). Although alternating isodense and hypodense layers are discernable by CT (Figure 2), the architecture of the lesion was more clearly evident in each of the MRIs. The hypodense layers in CT were hypointense on T1-weighted and hyperintense on T2-weighted MR images. The isodense layers in CT were isointense on both T1- and T2-weighted MRI series. Two cases demonstrated some variations from the typical MR images. Portions of the isointense bands demonstrated T1 shortening in patient 2 (Figure 3), probably from calcifications, and contrast enhancement involving the isointense layers was evident on MRI studies in patient 6 (Figure 4). There was also contrast enhancement in isodense regions of this patient's CT.
Table 1. Clinical and neuroimaging findings in LDD
Figure 1. (patient 5). T1-weighted sagittal (A), axial (B), and T2-weighted axial (C) images. These are representative of the typical pattern visualized with Lhermitte-Duclos disease. The T1-weighted images (A and B) demonstrate alternating layers of T1 isointensity and hypointensity creating mass effect and compression of the fourth ventricle and cerebral aqueduct. The areas of low signal intensity on T1-weighted images demonstrate increased intensity on T2-weighted images (C), whereas those that appear isointense with normal cerebellum on T1-weighted images again remain isointense on T2-weighted images.
Figure 2. (patient 3). CT (A) and T1-weighted MRI (B). The areas that appear hypointense on T1-weighted images and hyperintense on T2-weighted images appear hypodense on CT, whereas those that appear isointense to the normal cerebellum appear isodense to the normal cerebellum. Again the curvilinear laminated pattern and mass effect are evident.
Figure 3. (patient 2). Axial T1-weighted image. Linear regions of T1 shortening (arrow) are a rare finding involving the normally isointense zones in between the hypointense layers. This most likely represents early calcification.
Figure 4. (patient 6). Sagittal T1-weighted image (A). Multiple enlarged folia are evident, with the central portions appearing hypointense and the outer surface (arrowheads) appearing isointense. Coronal T1-weighted image with contrast (B) showed marked contrast enhancement within portions of the isointense regions.
The gross appearance of the tumor was of expanded folia that effaced sulci and clefts in the atrophic white matter that mimicked the sulci and leptomeningeal surface (Figure 5). Enlargement of the folia was caused by a thickened granular cell layer and hypermyelination of the molecular layer. Microcalcifications and severe loss of white matter within the cerebellar folia were also found. Variable degrees of microvascular proliferation were present, most notably in the leptomeninges and outer cortex (Figure 6).
Figure 5. (A) Residual tumor was present at autopsy in patient 6. Note the expansion of the cerebellar folia in the mass compared with the normal folia on the left (open circle) nearby. Within the lesion, the pale hypermyelinated molecular layer (fill square), darker abnormal granular cell layer (fill circle), and atrophic white matter ([open bullet]) that resembles a cleft are clearly seen (magnification x3.7, before 52% reduction). (B) T2-weighted (FSE 4,000/120Ef/1) MR image of the abnormal lobule demarcated by arrowheads in (A). The central portion corresponding to the inner molecular layer, the granular cell layer, and the slit-like cavity replacing the white matter appear hyperintense (*), while the outer molecular layer appears relatively isointense (arrowheads) to the normal brain. Plastic rods (arrows) are in the fluidfilled sulci.
Figure 6. Classic histologic features of LDD are seen in autopsy material from patient 6. (A) Note the prominent vascularity in the expanded molecular layer (arrows) (hematoxylineosin, scale: bar equals 0.40 mm). (B) Calcification is associated with blood vessels of the molecular layer (arrow) and overlying leptomeninges (hematoxylin-eosin, scale: bar equals 0.09 mm). (C) Large dysplastic neurons populate the granular cell layer. Purkinje cells are absent (hematoxylin-eosin, scale: bar equals 0.08 mm).
MR and pathologic evaluation of a postmortem specimen containing residual postoperative tumor from patient 6 demonstrated that the layers of T1 and T sub 2 prolongation corresponded to a zone consisting of the atrophic white matter, abnormal granular cell layer, and the inner portion of the expanded molecular layer. The intervening layers of MR isointensity were due to summation of the outer portions of the molecular layers of adjacent folia and leptomeninges sandwiched between them (Figure 7).
Figure 7. The repeating bilaminar cortical pattern evident on CT and MRI studies is compared with the histologic alterations that characterize Lhermitte-Duclos disease. The inner portion of the folia consisting of the white matter, the abnormal granular cell layer, and deep molecular layer, was hypodense on CT and hypointense on T1 and hyperintense on T2 MR images. The outer portion of the folia consisting of the outer molecular layer and leptomeninges within effaced sulci was isodense/isointense on CT and MR studies. Vascular proliferation in the pia and adjacent outer cortex may be associated with calcifications or rarely contrast enhancement within the isodense/isointense bands.
Discussion.
Ambler et al. [9] published an extensive review of 34 patients with LDD in 1969, and we have found 64 patients described in the literature since then. [2-8,10-44] Interestingly, because of the wide availability of MRI, there have been 45 LDD patients reported since 1989. [4-8,14-33] Patients typically present as young adults with a mean age of 34 years, but reported ages range from a newborn to 74 years. Symptoms are chiefly those of headache and papilledema from increased intracranial pressure due to obstructive hydrocephalus. Cerebellar signs are much less prominent. Familial occurrence is reported. [9,15] Numerous developmental abnormalities occur in LDD and include megalencephaly, heterotopia, microgyria, hydromyelia, polydactyly, peritheliomas, partial gigantism, macroglossia, and leontiasis ossea. [9,13] Coexisting conditions in LDD patients include neurofibromatosis [9] and severe postural hypotension. [3] Cowden's disease, or multiple hamartoma syndrome, an autosomal dominant disorder of the skin and mucous membranes, has coexisted with LDD in eight reports, [15-18,23-25,31] suggesting that this constellation of disease represents a phakomatosis. Characteristic skin lesions include multiple facial papules and cobblestone-like trichilemmomas. In this syndrome, thyroid disorders are also common, and malignancies of the breast, colon, and adnexa may occur.
The natural history of LDD is not well established. Protracted asymptomatic (4 years) and symptomatic (3 to 29 years) periods underscore the slow growth of the dysplastic tissue. [12] Initially, the outcome in non-operated patients was poor, but this result was likely due to the lack of control of increased intracranial pressure. [16] Roessmann and Wongmongkolrit [44] have incidentally found LDD in a newborn, which suggests the very slow evolution of this dysplastic process in subsequently symptomatic patients. A few cases had regrowth of tumor after surgical resection, [12,16] including an exceptional case in a 13-year-old girl who required three decompressive surgeries over 2 years. [13] Recurrences may occasionally follow prolonged disease-free intervals. Williams et al. [16] reported recurrence after gross removal in two patients postoperatively at 6 and 12 years. The possibility of recurrent disease many years after gross surgical excision necessitates long-term follow-up.
In 1988, Milbouw et al. [10] reported that MRI depicted the extent of LDD in a clearly superior manner compared with CT; however, no one has described diagnostic criteria of MRI findings in LDD to date. Subsequently, similar reports appeared, [4-8,11-14] some of which noted the laminated appearance of the cerebellar mass. Carter et al. [4] first reported a preoperative diagnosis of LDD based on MRI, and others have since accomplished this as well. [5-7,15-22] CT and MRI descriptions in LDD have directed attention toward the striated appearance, but there is no previously reported detailed MRI and CT evaluation in conjunction with historadiographic correlation.
Only recently, high-resolution CT has revealed the alternating serpiginous layers of low density and isodensity. In our patients, the low-density regions specifically correlated to regions of T1 and T2 prolongation on MRI, and the isodense regions were relatively isointense with the normal cerebellum.
Within the normal cerebellum, the hypointense/hyperintense signal on the respective T1/T2-weighted sequences is from the CSF in the sulci, the relatively hyperintense/hypointense signal on T1/T2 images is from the white matter, and the intermediate signal is derived from the cerebellar cortex. MRI of an available specimen showed a region of T1 and T2 prolongation that corresponded to the central portion of the enlarged folia while the cortical surface consisting of the outer molecular layer remained relatively isointense with the normal cerebellum (see Figure 5 and Figure 7). This impression was reinforced by the MRI in Figure 4A. Further support to this model is given by case reports in which calcifications appear to correlate with the regions of isodensity [10,32,37] as well as our one example of contrast enhancement that also involves this isointense layer. This correlates with frequent microscopic descriptions of a proliferation of blood vessels in the pia and adjacent molecular layer with concomitant calcareous deposits along the walls of the smaller blood vessels and in the perivascular tissue. [29,32] Increased vascularity was additionally noted in the molecular layer and adjacent pia of the contrast-enhancing specimen. Early calcium deposition would explain the regions of T1 shortening noted in the isointense layers in Figure 3.
Enlargement of the folia produces a mass effect that effaces the sulci so that they are not radiographically distinguishable. Within the folia, the loss of secondary arborization together with alternating layers of (1) opposing outer cortical and leptomeningeal surfaces (isodense on CT and isointense on MRI) and (2) more central cortex and white matter (hypodense on CT with T1 and T2 prolongation in MRI) creates a somewhat contorted alternating laminar pattern (see Figure 7). The laminar pattern on CT is much less evident than on MR due in part to the tremendous amount of Hounsfield artifact inherent in CT examination of the posterior fossa. Hence, MRI is the procedure of choice to diagnose LDD. We caution, however, that in the immature cerebellum of the infant and neonate, the signal characteristics may not be well defined. In fact, in one report of a 1-year-old boy, MRI did not show the laminated appearance observed in adults. [28]
The characteristic striated appearance of the cerebellum in a number of patients is sufficient enough to now regularly suggest the diagnosis of LDD preoperatively. [45] This capability allows for management options independent of a biopsy. Accurate MR diagnosis of incidental LDD in patients evaluated for unrelated reasons obviates the need for craniotomy for tissue diagnosis. These patients have the option of conservative management and follow-up with repeated scanning, probably annually, to exclude growth and complications such as hydrocephalus or tonsillar compression before resection. Screening MRI studies of patients with Cowden's disease and family members of patients with LDD who have megalencephaly is also recommended. For symptomatic patients, treatment should first be directed toward relief of obstructive hydrocephalus either by decompression of the posterior fossa or placement of a ventriculoperitoneal shunt. Partial or complete surgical resection may then be attempted. MRI will greatly assist the surgical approach by defining the extent of the disease that may not be apparent to the surgeon intra-operatively. Marano et al. [13] has questioned the value of irradiation therapy, and the indolence and apparently dysplastic nature of LDD tends to mitigate against its effectiveness. In all patients, long-term follow-up is advisable because of the occasional symptomatic recurrence.
Acknowledgments
We are grateful to Ms. Mary Althage for her secretarial assistance.
- Copyright 1997 by Advanstar Communications Inc.
REFERENCES
Disputes & Debates: Rapid online correspondence
REQUIREMENTS
If you are uploading a letter concerning an article:
You must have updated your disclosures within six months: http://submit.neurology.org
Your co-authors must send a completed Publishing Agreement Form to Neurology Staff (not necessary for the lead/corresponding author as the form below will suffice) before you upload your comment.
If you are responding to a comment that was written about an article you originally authored:
You (and co-authors) do not need to fill out forms or check disclosures as author forms are still valid
and apply to letter.
Submission specifications:
- Submissions must be < 200 words with < 5 references. Reference 1 must be the article on which you are commenting.
- Submissions should not have more than 5 authors. (Exception: original author replies can include all original authors of the article)
- Submit only on articles published within 6 months of issue date.
- Do not be redundant. Read any comments already posted on the article prior to submission.
- Submitted comments are subject to editing and editor review prior to posting.