Eighteen patients with a colloid cyst of the 3rd ventricle presented to the Frenchay Hospital Neurosurgery Unit. The investigations included computed tomography (CT). Of these patients, 9 had radiographically isodense cysts that caused considerable diagnostic difficulties, being described only as an anterior 3rd ventricular space-occupying lesion and lateral ventricular dilatation in all but 2 cases. The clinical findings are described and the CT appearance is discussed. The new use of ventriculoscopy may resolve the problem rapidly; ventriculoscopy made the diagnosis in 5 cases and confirmed a CT diagnosis in 3 others. (It has also confirmed the diagnosis in 2 cases of hyperdense cysts.) Lately, it has been possible to aspirate the cyst contents endoscopically or, if the colloid is too viscous, to remove it with endoscopic rongeurs. After either of these procedures, the cyst wall can be coagulated with diathermy endoscopically. The method is described, and the advantages over other investigations and treatment are discussed.

To assess the value of computed tomography in investigating patients with dementia, 500 consecutive patients presenting with a provisional clinical diagnosis of dementia of recent onset were reviewed and the results analysed in a computer. Most patients had either cerebral atrophy or infarction, but 82 patients had a normal scan and 42 others had tumours. More than 10% of all patients, including 5% with no other symptoms or signs, had a treatable lesion. Various associated symptoms and signs were useful pointers to such a treatable lesion and clearly indicated computed tomography. Accurate diagnosis is the cornerstone of proper management, and if a few patients with treatable lesions can be identified then the benefits to all concerned may be incalculable.

No abstract

The pattern of cerebral hamartomas among a population of patients with tuberous sclerosis and normal intellect was determined. All patients with tuberous sclerosis over 5 years old with normal intellect who were resident in the Bath health district were offered cranial scanning by magnetic resonance imaging. Cerebral axial and coronal images were obtained in 10 mm contiguous sections with a Picker 0-5 tesla magnetic resonance imaging unit. The number, size, and distribution of lesions found was recorded. Eleven of 13 eligible patients underwent scanning. Two patients had normal scans. Seven patients had between one and five subependymal nodules. Nine patients had between two and nine cerebral tubers best seen on T2 weighted images. Our findings suggest that the wrong conclusions may be drawn if the number of lesions alone is used to predict neurological outcome in tuberous sclerosis.

No abstract