Visualization of Liliequist's membrane prior to endoscopic third ventriculostomy

Endoscopic third ventriculostomy (ETV) is an effective treatment in patients with obstructive hydrocephalus caused by aqueductal stenosis. Operative failure may occur if an unnoticed membrane below the floor of the third ventricle is present, such as Liliequist's membrane (LM). To analyze how often LM can be visualized by high-resolution heavily T2-weighted 3D-MRI prior to ETV, and to find out potential reasons for diagnostic failure.

Preoperative 3D-MR images of 37 consecutive patients (19 female, median 42 years) were retrospectively analyzed. Visualization of three LM segments (sellar, diencephalic, mesencephalic), dimensions of the space below the third ventricle, and extent of hydrocephalus were measured. Image quality was scored (score 1[poor] to 3[excellent]). Preoperative imaging findings were compared with intraoperative findings.

Patients were subdivided into group 1 (no segment of LM identified, n = 18), and group 2 (at least one segment of LM was identified, n = 19). The sellar segment of LM was most often positively identified (10 out of 19 cases). The mean distance between the pons and the sella/clivus was significantly shorter in group 1 than in group 2 (3.7 vs. 6.2 mm; p < 0.01). Other variables, such as the distance between tip of the pons and the mamillary bodies as well as the image quality, were not significantly different between both groups. Intraoperatively, LM was present in 78 % of group 2 patients, and in 28 % of group 1 patients, respectively (p < 0.03).

LM can be detected in about half of patients prior to ETV. Reduced PSD influences visualization of LM.