As part of extracurricular work I carried out in my undergraduate program, I designed a non-invasive system for screening for breast cancer in humans. First, we start out with an electrical impedance measuring device connecting to a multiplexing and control circuit I designed to work with a PC via a graphical user interface.

System overview.

The multiplexing system then connects to an array of silver-silver-chloride electrodes injected into a breast phantom with a small deposit of lipid, calcium, and salt injected at a random location in the phantom in order to simulate a tumor.

System implementation.

From there, we perform impedance calculations between all possible pairs of electrodes using a second circuit to inject a current into the phantom using a separate pair of electrodes. For each pair of electrodes, the measurement is actually a set of measurements while sweeping the frequency controls of the impedance analyzer, and generateing Cole-Cole plots for each set of measurements.

Generated plots.

Finally, using the data generated from our sweeps, we use the EIDORS toolkit and some of our own MATLAB scripts to approximate the location of the artificial “tumor” we injected into the phantom, and generating a render of the final results, giving a visual representation corresponding to its expected location. This was perhaps the hardest part of designing the whole problem, giving the nature of solving inverse boundary value problems.

Final render