Research
My current research includes following projects:
Real-time motion tracking for medical procedures
Real-time motion tracking of a target (patient body, organ, tumour, endoscope, catheter etc.) has very wide applications in medical care. For example, during external beam radiation therapy on pulmonary and abdominal tumours, delivering accurate radiation therapy is limited by the motion of the tumour as the patient breathes. The solution is to track the tumour’s location during radiotherapy, then compensate for the motion through respiratory gating or even adjust the beam to follow the tumour in real-time. We proposed to use implanted positron emission markers for real-time tumor tracking (PeTrack). By implanting positron emission markers into the tumor, and using pairs of position-sensitive detectors to detect the resulting annihilation gamma rays, the position of the tumor can be tracked in real-time with high accuracy. We are also applying this technique to image guided surgery. By integrating the PeTrack with surgical x-ray c-arm, we will be able to provide real-time feed-back of the position of the surgical instruments without constant x-ray imaging, thus reduce the radiation dose to patient and surgeons. We recently have successfuly applied PeTrack in motion correction during Positron Emission Tomography.
GPU based fast Monte Carlo Simulation
Monte Carlo simulation has become the most precise radiation dose calculation method in radiation therapy. However, it is a very computation intensive. Usually computer cluster is required to achieve a practical calculation time, which comes with significant infrastructure cost. Thanks to the development in the gaming industry, Graphics Processing Unit (GPU) becomes a highly cost effective solution for parallel computing. We have recently developed a GPU based Monte Carlo simulation system adopted from the well accepted Monte Carlo simulation package (EGSnrc). We would like to further develop the potential of this GPU based system and eventually apply it for clinical radiation therapy.
Dynamic Dual-energy x-ray imaging (dDEXI)
While x-ray fluoroscopy can be used to observe internal organ motion, automatic and accurate evaluation of the motion is usually hindered by the interference between bone and soft tissue signals. For example, during breathing, the regional lung density variation can be observed on the x-ray images, which may be used to diagnosis emphysema and chronic obstructive pulmonary diseases. However, due to the overlapping rib signal in the image, the accuracy of these analyses is questionable. Dual-Energy x-ray imaging can overcome this issue by separating bone and soft-tissue images, taking advantage of their different attenuation coefficients as a function of x-ray energy. We have developed the dDEXI technique that acquires two image sequences simultaneously, thus, it is free of motion artifacts. We also propose using dDEXI for lung tumour motion assessment. The oncologist can use this information to determine whether motion management is required during radiotherapy. dDEXI can be a faster, cheaper, low dose alternative to current motion assessment methods, such as MRI or 4D CT.