Design of an impedance imaging system for orthopaedic use
This thesis describes the development of a microcomputer controlled 16 electrode Impedance Imaging System which can be used by the orthopaedic clinician, to measure resistivity changes associated with diaphyseal fractures of the human limb. The system is also designed to reconstruct two dimensional images of the approximate distribution of these changes. Electrical Impedance Tomography is a relatively new technique. It has attracted a lot of clinical interest since the technique is inexpensive, repeatable, portable and as far as is known, not harmful. Though spatial resolution is poor, temporal resolution is excellent. Research has been directed towards a number of clinical applications such as gastric emptying, dynamic cardiac imaging and monitoring hyperthermia treatment. Since bone is highly resistive, the occurrence of a diaphyseal fracture and the associated oedema should result in a decrease in resistivity in the region of the fracture. This particular application of the impedance technique was investigated here. An impedance measuring instrument was built in the first instance. It was designed to be able to inject a current of 1 milliampere at a frequency of 10 kHz by a chosen pair of electrodes into a region surrounded by 16 electrodes. The resulting voltages on the boundary were then measured. It was interfaced with a BBC Master microcomputer which was programmed to collect these boundary measurements, analyse the data and reconstruct images of the distribution of log(resistivity) or resistivity in the region. The instrument was tested for its linearity, stability and accuracy. Subsequently phantom tests were carried out to assess its performance. Phantom experiments showed that this system was capable of both measuring resistivity changes and imaging phantom objects adequately. The system could image a resistive object of ~ 15% of the diameter of the phantom. It could resolve two resistive objects spaced one phantom radius apart in the central region. The battery of tests performed on the phantom indicated that greater changes in peripheral voltage gradient measurements would occur if the background resistivity changed as compared to changes in a small region of the phantom. The experiments also underlined the fact that the shape of the region was of great importance in the boundary voltage gradient profile. The property of bilateral symmetry of human limbs was utilised to compare resistivity measurements and images of one limb with that of the other. In the normal individual they underlined the similarity of the limbs. Resistivity measurements in volunteer patients with limb fractures confirmed the results of phantom tests. Hence soon after a fracture the oedema of the surrounding tissues causes a large drop in resistivity superimposed on a smaller drop due to the fracture. Similar measurements in an un-united fracture, where there is virtually no oedema, showed that there is significantly lower resistivity in the injured limb compared to the normal limb. Measurements on a volunteer with a united fracture showed greater resistivity in the injured limb compared to the normal limb. Three volunteers with upper arm fractures at different levels showed increasing resistivities towards normal levels during the healing process. Static impedance images of the distribution of log(resistivity) were found to contain a number of artefacts because the geometry of the circular reference and the upper arm did not match. However since bilaterally symmetrical limbs were being compared these artefacts appeared to be duplicated in the two and changes associated with the healing process were still evident in the images. Differential images appeared to reduce the problem of inexact geometry resulting in fewer artefacts. Accuracy of electrode positioning was still found to be critical and the images therefore were less than ideal. Though the number of subjects was limited, the results of this study were promising. They indicated that electrical impedance measurements comparing the fractured limb with the contralateral normal limb could offer the orthopaedic clinician unique information about the changing electrical characteristics of the fracture region. Given the possibility of better electrode positioning and correction for geometric shape, the improvement in impedance images would be a useful adjunct to clinical monitoring of fracture healing.