Ehsan Golkar

Radiation therapy is one of curative treatment ways of cancerous tumors in the thoracic-abdominal regions and respiratory motion is a major obstacle which is associated with the irradiation therapy. Holding breath during the treatment for example is not a good solution since many patient are not able to hold their respiration for long time. Gated treatment and tracked treatment are suggested to overcome the drawback of respiratory motion. Both methods are able to compensate the respiratory motion effectively; however, it is very difficult to monitor the tumour during the treatment. In addition, extra dose exposure of tracking tumour during treatment is the disadvantage of these methods.  One method which has been recently studied is to deduce internal motion using depth camera systems that images the surface of the body during external beam radiotherapy (EBRT), reducing the need for additional imaging and reducing radiation dose. 

 Methodology 

  • Internal motion

    Deformable image registration is a well-known tools to find the displacement field of internal organs between different time points. Then the displacement field is corresponding with organ motion between pair frames.  My current research focuses on an image registration framework (Creg) to estimate thoracic-abdominal respiratory motion using image registration of 4D (Ehsan Golkar, A.A.A. Rahni,MIC 2015). 
Displacement field extracted from Creg
  •  External motion

    Range imager which is known as time-of-flight (ToF) 3D camera or stereo camera is widely accepted in academic research recently. Range imager provides depth image which contains 3D geometric information from a real-time environment. A simulation of range imager is carried to acquire multi-dimensional surrogate signals of 4D MRI (Ehsan Golkar, A.A.A. Rahni,MIC 2016). 
3D mesh presentation of phantom example of surrogate signal
  • Correspondence modelling

    Correspondence model relates external surrogate signals and internal respiratory motion which compensate respiratory motion and applicable in RT practices.
Correspondence modelling