Statistical modeling

in medical imaging

 

  People

  Software


  MSMI main

      IBIA





Using InShape models for image

analysis and parameter prediction

One major applicaiton of InShape models is the automatic segmentation of images

(more>>).

However, InShape models have also been sucessfully used for the assessment of

pathogenetic factors, which manifest in variations of the appearance of an organ

and also for the automatic prediction of biomechanical parameters.

For this purpose, InShape models offer the possibility of using information

about the geometrical variations as well as variations concerning the inner

structure/appearance of an organ.

For more details on the developed methods and algorithms that have been

developed at the IBIA and MSMI, please also refer to the corresponding

publications.

 

 

Analysis of the proximal femur in

(clinical) CT data

 

Assessment of individual fracture risk and pathogenetic factors

in CT images of femur specimen with clinical resolution

Assessing the individual risk for femoral neck fractures of a patient in CT images

provides information that is necessary to start a therapeutic intervention on time.

Hence, the number of femoral neck fractures and especially the complications after

necessary surgical interventions could possibly be decreased. Some of the underlying

methods and algorithms were developed in the course of the project

"Local Bone Analysis, X-Ray and CT Analysis".

 

 

Regions of interest that were placed automatically for the assessment of the

individual fracture risk using CT images

 

 

Predicted versus real faliure load (~fracture risk) of 86 specimen

 

 

 

Principal components that describe variations, which are relevant
for the individual fracture risk of a patient using InShape models.

Relevant geometrical variations as well as variations concerning the inner
structure/appearance could be identified.

 

 

Assesment of local bone quality of the proximal femur

Relevant structural variations that influence local bone quality could be identified

and quantified. Having quantitative information about the local bone quality

prior to a surgical intervention helps the surgeon to plan the intervention and might

therefore reduce the risk of complications after the intervention (e.g. implant cut out)

For this purpose CT images with clinical resolution as well as radiographs

have been used.

 

CT images

Strcutural variations, which are relevant for the assessment of the local bone quality

inside a spherical region in the femoral head (location of the tip of a dynamic

hip screw (DHS)) as described by the InShape models

(also compare to the images below)

 

 

Actual variations that can be observed in CT data of specimen with low local bone

quality and high local bone quality

 

 

Radiographs

Strcutural variations, which are relevant for the assessment of the local bone quality

inside a spherical region in the femoral head (location of the tip of a dynamic

hip screw (DHS)) as described by the InShape models

(also compare to the images below)

 

 

 

Actual variations that can be observed in CT data of specimen with low local bone

quality and high local bone quality

 

 

 

Analysis of the ventricular movement of healthy

and diseased patients (pilot study)

Endocardial movement (between 0 ms and 100 ms of the cardia cycle) of a healthy heart

(left) and the heart of a patient suffering from cardiomypathy (right) in 2 different views.

 

 

Local cardiac movement (part of left ventricle) of 2 healthy hearts (left, middle)and

one heart after "simulated" myocardial infarction (right) mapped onto one

reference subject for better comparison

 

 

Contact:
Karl D. Fritscher, MD, PhD
Head of the Research Group
Room: G3 41
Tel: +43-50-8648-3866
Fax: +43-50-8648-673866
karl.fritscher@umit.at