Publication Date




Embargo Period


Degree Type


Degree Name

Doctor of Philosophy (PHD)


Biomedical Engineering (Engineering)

Date of Defense


First Committee Member

Weizhao Zhao

Second Committee Member

Xiaodong Wu

Third Committee Member

Fabrice Manns

Fourth Committee Member

Jorge Bhorquez

Fifth Committee Member

Edward A. Dauer

Sixth Committee Member

Yidong Yang


Uveal melanoma is the most common primary intraocular tumor in adults. One of the effective managements of ocular melanoma is the localized radiation treatment. In addition to the commonly performed invasive radioactive plaque brachytherapy, using focused external photon radiations, x-rays or gamma rays, have emerged as treatment options. However, these radiation delivery methods are either involved with the invasive surgical operation or limited by the gantry movement to account for the involuntary movement of the eyeball. The goal of this doctoral research is to design and validate a method that can offer a non-invasive radiosurgery for uveal melanoma and to compensate for the eyeball’s movement. The Cyberknife system, driven by a robotic arm with 6D freedom of motion and guided by the x-ray stereotactic and infrared cameras, was chosen as the radiosurgery platform to investigate the feasibility of the method designed in this research. The essence of this study demonstrates the non-invasive tumor tracking capability through the data transformation by tracking the pupil in real time. The research was conducted in three stages. We first performed a computer graphical simulation, in which the eyeball, the intraocular tumor, and the pupil were all simulated by mathematical models. We successfully proved that the pupil’s 2D projection on the image plane (captured by a camera) relates to the tumor’s 3D location. We derived a 2D/3D transformation, a linear model linking the pupil’s 2D coordinate and the tumor’s 3D coordinate. This model laid a mathematical foundation for the research. The error prediction analysis was also performed in the simulation. In the second stage of the research, we built a mechanical phantom, which is a prototype to implement the designed method. The mechanical phantom consists of a camera module, an eyeball module, an eyeball holder module, and an eyeball motion module. In the third stage of the research, we performed the validation by using the mechanical phantom under the CT and CyberKnife machines. Under the CT machine, the captured pupil’s positions were associated with the scanned tumor’s positions so that we generated the 2D/3D transformation. We then moved the mechanical phantom under the Cyberknife system. After the re-alignment of the phantom, we used the newly tracked pupil position to predict the tumor position. Based on the promising outcomes, we concluded that the designed non-invasive pupil tracking method for the tumor tracking under the Cyberknife system is a feasible approach and is ready for the dosimetry validation. This dissertation presents, in details, the above mentioned research stages and includes an introduction of the physiology and anatomy of the eye and uveal melanoma (Chapter 1), a review of radiosurgery methods (Chapter 2), descriptions of the optical tracking mechanism and the derivation of the 2D/3D transformation (Chapter 3), descriptions of the mechanical phantom (Chapter 4) and system integration (Chapter 5), discussion and conclusion of the research (Chapter 6 and 7).


Eyeball Phantom, Uveal Melanoma, Image registration, Robotic Radiosurgery, Cyberknife