Lhermitte duclos disease: what to expect during surgery?

Lhermitte and Duclos first described the dysplastic gangliocytoma of the cerebellum in 1920. In the last 100 years, its clinical presentation, radiological features, pathological characteristics, and association with Cowden syndrome have been well described. However, documentation of surgical experiences is lagging. We here describe intraoperative experience during the removal of the tumor, which could help the operating surgeon plan and mental makeup.


To the Editor
The present literature on Lhermitte Duclos Disease (LDD) focuses on its presentation, radiologic and pathological features [1][2][3]. They lack detailed intraoperative pictures, findings, and surgical nuances. Its rarity may perplex a young neurosurgeon regarding intraoperative differentiation between dysplastic and normal folia, vascularity, consistency, ease of removal, and radicalism of resection. We have focused here on filling in such details.
A 19-years male had undergone a ventriculoperitoneal shunt two years back for headache, visual blurring, and drowsiness at a peripheral hospital. There were no previous imaging records. Following resolution of the symptoms, he did not seek further treatment until eight months when the headache reappeared, and gait disturbances worsened progressively. Examination revealed truncal and appendicular ataxia, left upper limb dysmetria without papilloedema. Radiological evaluation showing characteristic "tiger stripe" appearance has been described (Figs. 1, 2, 3).
Surgery was planned in view of brainstem compression in a prone position. The Head was slightly turned to the opposite side and flexed to make the tent perpendicular to the floor (Fig. 4a). A left lateral sub-occipital craniotomy was made across the midline, including the foramen magnum's rim, up to the transverse and sigmoid sinuses (Fig. 4b) as the dura was tense, hyperventilation and decongestants were used before dural opening in a Y-shaped manner up to the transverse sinus to expose the tentorial surface (Fig. 4b). Cisterna-magna was drained for further relaxation. No lesion was visible on the suboccipital surface (Fig. 4c). On gentle retraction over the tentorial surface, pale white, widened folia could be distinguished from normal folia (Fig. 4d). Normal draining veins were seen entering the tentorial sinuses from dysplastic tissue. A normal-sized superior petrosal vein was seen.
Dysplastic folia were firm, rubbery, and not easily suckable but amenable to CUSA. The dysplastic tissue had low vascularity, but the vessels traversing between the folia were the source of hemorrhage. Subpial resection helped preserve veins. After significant decompression, the trochlear nerve, pre-central cerebellar vein, and vermian veins draining into the vein of the Galen complex were identifiable. We then focussed the microscope at the under-surface of the cerebellum for further tumor resection. The planes between dysplastic and normal folia were well-defined superficially but turned progressively indistinguishable based upon color, width, and consistency in the deeper areas of the inferior and anterior cerebellar surfaces, which precluded further resection. At the end of the surgery, the posterior fossa was lax. Dura was closed primarily, and the bone flap was replaced.
Postoperatively, there were no new neurological deficits. CT showed adequate decompression; however, the residual lesion was seen on T2WI MRI in the inferior cerebellar regions (Fig. 5a, b). Histopathological examination and PTEN-positive mutation confirmed the diagnosis of LDD. At six months follow-up, here is a partial improvement in ataxia.
This pale white lesion [4] has a distinct plane superficially, but as one digs deeper surgical planes get obliterated, likely due to an "Arbor-Vitae" arrangement of cerebellar white matter. The dysplastic folia have a firm, rubbery, low-grade glioma-like consistency and may be held with forceps, but excessive pulling during dissection may damage the normal folia. Suctioning does not efficiently decompress the tumor, and CUSA is a boon. Cutting Bi-polar forceps may also serve the purpose.
Although the lesion itself has low vascularity, one should not presume it to be a bloodless procedure. The dysplastic folia do not enhance on contrast; streaks of enhancement inside the lesion denote traversing vessels between folia. Increased regional blood flow on perfusion-weighted MRI and hyper-metabolism in FDG-PET has been described [5]. Careful transgression across multiple sulcal spaces, harboring vessels, alleviates bleeding. Preservation of veins is of paramount importance as they drain normal cerebellar tissue along with the dysplastic ones. These veins may be preserved using subpial resection, especially towards the brainstem.
To the best of our knowledge, there is no published report for intraoperative ultrasonography use, aiding resections of LDD. Although we had not used it, we believe that the intraoperative ultrasonography could have helped detect the lesion and decide the margins. Similarly, no published report suggests the use of Fluorescence-guided resection in LDD. The intensity of Fig. 2 Preoperative MRI scan. a T1WI shows hypo-intense lesion involving the tentorial and petrosal areas of the cerebellum, upper 2/3rd of the vermis, with an indentation on the brain stem. b T2WI shows a hyperintense lesion with alternating hypo-intense bands giving the characteristic "tiger stripe" appearance. There is thinned-out cerebellar parenchyma overlying the lesion. There is no perilesional edema (Original) fluorescence correlates with tumor cell density, proliferation rate, vascularity, and blood-brain-barrier integrity. LDD is a grade I tumor with low proliferation rate, vascularity, and does not enhance on contrast, suggesting the intact blood-brain-barrier. We had not used fluorescent dye in this case and are unsure about its utility. Hydrocephalus in these patients occurs secondary to fourth ventricle compression. We LDD is a WHO grade-I tumor. Gross total resection is recommended but is often not possible [2]. The preoperative images of our patient showed thinned-out normal cerebellar parenchyma with intraoperative loss of plane of differentiation during resection. The decompression was therefore done conservatively to prevent damage of left-over functional cerebellar tissue. Recurrence after subtotal resection has been reported [1]. We plan to follow our patient with radiological evaluation and proceed accordingly.