Novel application of deformable image registration using Computed Tomography for early diagnosis of lung parenchymal disease: focus on lung fibrosis

We aim to optimize advanced Computed Tomography (CT) imaging to provide non-invasive and accurate assessment of early lung fibrosis. Lung fibrosis causes inflammation and scarring of lung tissue and can be progressive, even fatal. Currently, detection of lung fibrosis relies on relatively insensitive tests, such as pulmonary function tests and static lung CT. Alternately, microscopic evaluation of lung cells (histology) can be used but this technique is invasive and provides inconsistent results in disease states that exhibit non-uniform lung involvement.

We propose to develop an ultralow dose dynamic CT protocol and use it in conjunction with lung deformable image registration to detect early changes in lung elasticity. Histology and other, well established clinical disease indicators will be used to confirm the correlation between the severity of lung fibrosis and the degree of change in lung deformation.

Faculty Supervisor:

Dr. Narinder Paul

Student:

Miho Horie

Partner:

Toshiba Medical Systems

Discipline:

Physics / Astronomy

Sector:

Life sciences

University:

University of Toronto