Accurate tumour localization is crucial for the diagnosis and treatment of liver cancer, including procedures such as biopsy and tumour ablation.
Due to respiratory motion, the liver moves constantly, and accurate localisation is challenging as the imaging modalities do not have sufficient high-quality and real-time visualisation to compensate for this. Previous research has explored surrogate signals to develop motion estimation models that predict liver and tumour motion. These models have potential applications in procedures such as robotic-assisted biopsy, where a robotic system, guided by hepatic motion feedback, can dynamically adjust the needle position to remain aligned with the tumour. However, existing surrogate signals often require additional equipment, such as external cameras and abdominal markers, making them less practical in a clinical setting.
This master’s assignment aims to investigate whether routinely monitored vital signs correlate with liver motion and can serve as surrogate signals for motion estimation models. Utilizing vital signs for motion estimation could simplify implementation, as these measurements are already integrated into clinical workflows and require no additional setup for patients already connected to monitoring systems. The motion estimation model may be developed using mathematical modelling, machine learning, or deep learning techniques.
The research will involve a literature review of existing liver motion estimation techniques. Additionally, experiments with human subjects will be conducted to analyse correlations between vital signs and respiratory motion, using external abdominal markers as a ground truth reference due to their established accuracy and non-invasive nature.
Finally, a validation experiment will be performed using MRI at TechMed, where vital signs will be recorded during imaging to assess their reliability as surrogate signals.