A theory on the natural history of colloid cysts of the third ventricle.

OBJECTIVE

Patients with third ventricular colloid cysts typically are diagnosed when they develop symptoms related to cerebrospinal fluid (CSF) obstruction at the foramen of Monro. However, the clinical and neuroimaging characteristics related to symptom development are poorly understood.

METHODS

From January 1974 to June 1998, 155 patients with newly diagnosed colloid cysts were managed at our center. Eighty-seven patients (56%) were thought to have tumor-related symptoms, and they underwent surgery (resection, n = 74; ventriculoperitoneal shunting, n = 11; stereotactic aspiration, n = 2). Sixty-eight patients (44%) had colloid cysts thought to be asymptomatic, and observation with serial neuroimaging was recommended. Recursive partitioning was used to examine the association of patient and tumor characteristics with the development of cyst-related symptoms.

RESULTS

Univariate analysis comparing symptomatic and asymptomatic patients revealed four factors associated with cyst-related symptoms: 1) younger patient age (44 yr versus 57 yr; P < 0.001); 2) cyst size (13 mm versus 8 mm; P < 0.001); 3) ventricular dilation (83% versus 31%; P < 0.001); and 4) increased signal on T2-weighted magnetic resonance images (44% versus 8%; P = 0.001). All four variables remained significant in a multivariate logistic regression model: patient age (P = 0.04; odds ratio, 1.0); cyst size (P = 0.04; odds ratio, 1.2); ventricular dilation (P = 0.02, odds ratio, 7.2); and increased signal on T2-weighted magnetic resonance images (P = 0.04; odds ratio, 2.7). The most significant variable was ventriculomegaly (yes versus no). Patients with normal-sized ventricles (n = 62) were further categorized by cyst size (< or = 10 mm versus > 10 mm). For patients with enlarged ventricles (n = 93), patient age (< or = 50 yr versus > 50 yr) was the most important variable. Patients older than 50 years also were split with respect to cyst size; patients aged 50 years or younger with enlarged ventricles were not affected by cyst size. The percentage of patients with cyst-related symptoms was 12, 50, and 85% in the three final patient classes, respectively. Multivariate analysis including the patient classes resulted in removal of the other significant variables from the model, whereas the patient classes remained significant (P < 0.0001; odds ratio, 6.3) for predicting patients with cyst-related symptoms.

CONCLUSION

The patient and neuroimaging characteristics of the different patient classes support a theory on the natural history of colloid cysts. Patients with third ventricular colloid cysts become symptomatic when the tumor enlarges rapidly, causing CSF obstruction, ventriculomegaly, and increased intracranial pressure. Some cysts enlarge more gradually, however, allowing the patient to accommodate the enlarging mass without disruption of CSF flow, and the patient remains asymptomatic. In these cases, if the cyst stops growing, the patient can maintain a steady state between CSF production and absorption and may not require neurosurgical intervention.