Improved Control System for Digital Imaging Elasto-Tomography Breast Cancer Screening

Abstract

Breast cancer is the most recognised form of cancer in New Zealand women. Early treatment is a key factor in the long-term outcome of the disease. Robust screening methods are an essential component to help early diagnoses. Common screening methods include mammography and physical examination. Both methods have significant downsides including discomfort and exposure to radiation. Digital Imaging Elasto-Tomography (DIET) is a new breast cancer screening technology. Screening involves women lying prone on the machine, while sinewaves of known amplitude and frequency are applied to the breast at the nipple. Images are taken at multiple phases of the sinewave as the breast is vibrated. These images are processed to determine the surface motion and consequently the tissue properties of the breast, which is then used to distinguish tumours from healthy tissue. A prototype of this machine has been designed and used on breast phantoms and in clinical trials.A known issue with the prototype is noise in the actuation system. This creates vibration in the rest of the machine and impacts the quality of the oscillation imparted to the breast. A possible solution isolates the actuator from the machine by using damping feet. The motion of the feet was measured using an extra Linear Variable Differential Transformer (LVDT) sensor. Feedback control for the actuation was provided by the combination of this sensor and the original one. The results of these changes were compared to using a single sensor for feedback.The vibration of the feet reduced the overall amplitude of the actuator when in operation. This reduction was accounted for when two sensors were used for feedback. The Signal to Noise Ratio (SNR) was used to quantify the quality of the actuation. At 20 Hz the SNR was 22.7 and 22.0 using one and two sensors respectively. At 50 Hz the SNR improved to 39.9 and 38.6 for one and two sensors respectively. The impact of the reduced range of motion could be seen in the surface motion reconstruction of the breast phantom. The new control scheme reduced the vibration transferred to the rest of the machine and improved the amplitude accuracy. Potential changes to the software implementation of the control system may further improve the quality of the output.