Trajectories for frontal external ventricular drain placement: virtual cannulation of adults with acute hydrocephalus

Abstract

Objectives. External ventricular drains (EVDs) are commonly placed freehand using targeting landmarks unchanged since the pre-CT era; it is known to be an inaccurate procedure. To our knowledge, this is the first study to assess the geometric reliability of specific trajectories in a three-dimensional model. Design. Three-dimensional volume reconstruction of EVD trajectories in a Stealth Station S7. Subjects. Adults with a primary EVD sited for acute hydrocephalus secondary to spontaneous subarachnoid haemorrhage with CT angiography less than 24 hours previously. Methods. CT angiograms from 10 consecutive patients meeting the inclusion criteria were reconstructed. The surgical planning tool was used to construct three trajectories from Kocher's point: i) perpendicular to the skull (PTS) ii) towards the ipsilateral medial canthus coronally and the external auditory meatus sagitally (IMC) iii) towards the contralateral medial canthus coronally and the external auditory meatus sagitally (CMC). Their engagement with the frontal horn of the ipsilateral lateral ventricle (FILV) and distance from the ventricular wall and foramen of Monro were measured. Results. Mean supratentorial ventricular volume was 55.8 cc (range 35.2–83.4 cc). The IMC met the FILV in only one patient, on average missing the ventricular wall by 5.5±2.3 degrees (95% confidence interval). CMC and PTS met the FILV in 9 and 10 cases, respectively. Mean engagement was 16.3±5.1 mm (95% confidence interval) for PTS and 20.0±7.1 mm (95% confidence interval) for CMC. CMC and PTS gave significantly better engagement and aiming error margins than the IMC trajectory. Conclusions. Despite its widespread use, the IMC trajectory performed poorly; PTS and CMC trajectories are more reliable ways of targeting the FILV when placing an EVD.

Keywords::

• ventriculostomy
• hydrocephalus
• external ventricular drain
• subarachnoid haemorrhage
• trajectory

Acknowledgement

The authors are grateful to Mr Barrie White and Mr Stuart Smith for their suggestions regarding this paper.

Declaration of conflicts: The authors report no conflicts of interest. The authors alone are responsible for the content and writing of the article.

Appendix A – Locating the trajectory origin at Kocher's point vs. the cortical surface immediately deep to Kocher's point (CSDK point)

Modelling trajectories originating from the CSDK point rather than Kocher's point would produce the greatest difference in results to ours for trajectories which differ most from the perpendicular to skull trajectory. In the context of our results, the difference between these two estimation is unlikely to be significant for two reasons: First, the skull thickness is a relatively small part of the total catheter trajectory – our mean perpendicular skull thickness was 9.5±1.2 mm (p <0.05), while the mid-engagement point of our foramen of Monro trajectory (an estimate for the depth of the centre of the frontal horn of the ipsilateral lateral ventricle) lay at a mean of 53.9±1.6 mm (p <0.05) – and secondly a translation in the origin of either an ipsilateral medial canthus or contralateral medial canthus constructed line form Kocher's point to the CSDK point will necessarily be partially compensated by an adjustment in angulation in the opposite direction. Even for our most deviant trajectory, ipsilateral medial canthus, this gives a mean difference of 3.0 mm by the depth of CSDK. Making the assumption that the midpoint of the frontal horn of the ipsilateral lateral ventricle lies roughly halfway between the entry point and the final targeting landmarks used in the ipsilateral medial canthus and contralateral medial canthus trajectories, then partial compensation will have reduced this difference to 1.5 mm by the midpoint of the frontal horn of the ipsilateral lateral ventricle. This equates to a difference in the absolute angle between perpendicular to skull and ipsilateral medial canthus trajectories of 1.6 degrees. This is smaller than our 95% error bounds for our estimates of all relative coronal angulation and relative sagittal angulation values. The coronal and sagittal components of this absolute angle will necessarily be even smaller. It is also an order of magnitude smaller than the difference in relative sagittal angulation and relative coronal angulation seen between perpendicular to skull and ipsilateral medial canthus trajectories. Whether we model the origin of the trajectory at Kocher's point, or the cortical point immediately deep to it seems unlikely to have a significant effect on the overall result.