GPI Seminar Series: Carlos Ciller
 |
Carlos Ciller, Medical Image Analysis Laboratory (MIAL) Center for Biomedical Imaging (CIBM) of the University of Lausanne (UNIL) |
Multi-modal patient-specific eye-modeling: Robust and accurate treatment planning systems for intraocular tumors
Abstract:
In the past few years, retinoblastoma and uveal melanoma, the most common cancers of the eye, have witnessed an important advance in the field of medical imaging and its applications in ophthalmology. Magnetic Resonance Imaging (MRI), Computed Tomography (CT), Fundus Photography and Ultrasound (US) are among the image modalities of reference for treatment planning and diagnosis confirmation of the disease. However, existing methods for modeling the eye are still imprecise and they do not benefit from state of the art techniques developed in other medical image processing applications. In addition to that, there is a disconnection between the different image modalities due to the lack of common anatomical landmarks and to the difficult method standardization in ophthalmic radiation oncology.
This situation mainly affects the treatment planning (radiation dose optimization) step involved in External Beam Radiotherapy (EBRT), Cryotherapy and Brachytherapy, thus precluding optimal preservation of healthy tissue in patients during treatment. In this context, having a robust and accurate segmentation tool for the eyes in the MRI, and a method to combine together Fundus Photography with MRI by connecting common anatomical landmarks would represent a breakthrough towards multi-modal patient specific eye modeling. In this presentation we introduce i) the first fully automatic segmentation of the eye in the MRI for the regions of the sclera, the vitreous humor and the lens in children, and ii) a new approach to combine Fundus photography with MRI by combining the position of the optic disc and the fovea with the position of the optic nerve.
Carlos' short bio:
Carlos holds a Telecommunications degree from the ETSETB/UPC. He carried out his Master Thesis project at EPFL, and later joined the CIBM/MIAL of the University of Lausanne, where he is currently pursuing a Phd, focusing on novel Image processing methods for A Multi-modal imaging framework for ocular tumor planning, under the supervision of Dr. Meritxell Bach and Prof. Jens Kowal (ARTORG, Bern), with the support of the Swiss Cancer League.
Carlos' research interests include signal and image processing, machine learning techniques, medical imaging, image registration and statistical shape modeling. His PhD focuses on computer assisted treatment planning systems for intraocular tumors.