Carotid angiography in the localisation of supratentorial neoplasms and hamartomata.

Abstract

The high accuracy of carotid angiography in demonstrating such vascular lesions as intracranial aneurysms, arteriovenous malformations, subdural haematomata, and the like, has been well established. The large majority now agree that cerebral angiography is the method of choice in the diagnosis of these vascular lesions. The place of carotid angiography in the diagnosis and localisation of supratentorial neoplasms is not so well established, and there is much difference of opinion concerning the relative merits of ventriculography, pneumoencephalography, and carotid angiography. The following figures are taken from an investigation by Hodges and Holt (1) into the relative reliability of different radiological methods in localisation of brain tumours. As is apparent, ventriculography was used in a high proportion of cases. Cerebral angiography is stated to have an accuracy of 72 per cent, against an accuracy of 60.5 per cent for encephalography, the definition of accuracy as used in this review being a localisation sufficient to indicate correctly the site and size of bone flap adequate for approach to the tumour. It is very doubtful, however, if radiologists and neurological surgeons should be satisfied with this criterion of accuracy, as it ignores the question of correct assessment of the extension in depth of a lesion. A second tabulation illustrates the findings of an investigation into the relative accuracy of pneumography and angiography in the diagnosis of space-occupying lesions, conducted by Martin and Webster (2). Again, a high percentage of cases was submitted to ventriculography. Encephalography was found to have an accuracy of 36 per cent, but there were 20 diagnostic failures in 90 encephalograms. Since encephalography was performed by massive replacement of cerebrospinal fluid by air (100 to 150 c.c. exchange), this high percentage of diagnostic failures is perhaps not surprising. This also affects the estimate of accuracy of encephalography. It is noted, too, that 17 of the 82 cases examined by encephalography in this series were cases of subdural haematoma, for the investigation of which encephalography is hardly the method of choice. Only half of the subdural haematomata were localised by angiography, a remarkably low figure. It is apparent from the above surveys that the methods of selection of cases for the three procedures, encephalography, ventriculography, and carotid angiography, left something to be desired and also that the technique of encephalography, in one series at least, was of debatable value. We have found, by careful selection of cases for each of these three procedures and by careful elaboration of technique, including particularly a controlled fractional air injection of the Lindgren type for encephalography, that ventriculography is required in very few cases of intracranial space-occupying lesion, whether supratentorial or infratentorial.