Proximity Map Projection: Interactive Visualisation for Image-Guided Surgery
Abstract
This thesis describes a new interface technique for neurosurgeons
and interventional radiologists performing image-guided therapies
such as the ablation of brain tumours. This new technique is
called Proximity Map Projection (PMP).
Based on an analysis of related work, including the documented
recent progress in enabling technologies, a case is made that
present-day interactive visualisations supporting image-guided
treatment of tumours will need to be dramatically improved to
take advantage of the increased image refresh rates available as
soon as 2020. This probable requirement for improved
visualisation technology in the very near future motivated the
invention and investigation of the PMP technique described in
this thesis. The PMP technique is an interactive 2-D visual
projection of the proximity of two 3-D surfaces – in
particular, the surface representing the boundary of a thermal
treatment region, and the surface of a tumour that is the target
of this treatment. By clicking on interesting points in the PMP,
surgeons are able to quickly select the 2-D MRI slices
corresponding to those interesting points. The PMP provides a
quick way of selecting a desired image from a large stack of 2-D
MRI data, thus freeing up surgeons to spend a greater proportion
of their time applying their expertise to decision making, rather
than to navigating through image data.
In this thesis, the PMP technique is presented and then refined
as user studies are undertaken. In a series of investigations
exploring its effectiveness, it is shown that the PMP technique
enables non-expert users to quickly and accurately navigate to,
and observe, desired individual medical images within large
stacks of such images. A further experiment finds no significant
differences in the way that medically experienced and
inexperienced users use PMP to complete tasks. That study also
verifies that users pay visual attention to PMP, regardless of
whether or not they have interacted with it via the mouse.
Observation of the visual attention of users during simulated
tasks is used to provide further explanation of why PMP is
effective. PMP’s potential to be used by medical professionals
is then assessed via a series of semi-structured interviews with
surgeons and interventional radiologists. Such experts are found
to be optimistic about the potential for PMP to be incorporated
into their workflows. This last phase of the research then
culminates with observations of a number of medical procedures on
human patients that are similar to the kinds of procedures to
which PMP might be applied. From these observations it appears
that, while they do enable life-saving therapies, present
software interfaces are not entirely satisfying for the surgeons
who use them. Opportunities for significant future research
collaborations were identified during these interviews and
observations.
This thesis concludes by describing a practical path towards
achieving its ultimate goal: the use of PMP in real-time
image-guided medical procedures on human patients. Key activities
on this path include: integration of PMP into the training and
simulation version of a collaborator’s therapy system;
conducting a case study to allow further refinement of the PMP
technique; and inclusion of PMP in a clinical trial with
surgeons.
Description
Citation
Collections
Source
Type
Book Title
Entity type
Access Statement
License Rights
Restricted until
Downloads
File
Description