Extraction of Follow up Parameters of Bone Density Microwave Sensor from Post Craniotomy and Lower Extremity Trauma Rehabilitation Measurements

Longitudinal microwave based sensor systems facilitates frequent follow ups in scenarios where healing information is largely missing. An example is neonatalcraniotomy where Computerized Tomography (CT) information is available mostly before surgery and up to three years after that. In such case, frequent CT’s cannot betaken due to multitude of reasons ranging from dosage concerns to sheer cost. In this context, the use of a follow-up modality could substantially improve the quality of life. Bone Density Measurement Analysis (BDAS) and Complex Fracture Orthopaedic Rehabilitation (COMFORT) are two such projects dealing with collecting vital information that will help in addressing the unknown physiological changes. Compliant to ethical approvals 200 low extremity trauma patients from Holland and23 craniosynostosis patients from Sweden, were enrolled in clinical trials for theCOMFORT and BDAS projects respectively. For COMFORT study, itself, it involves200 (patients) x 3 (low extremity locations) x 5 (Repetition) x 9 (time points) =27000 data sets. Similarly, the BDAS projects deals with 966 data sets. Microwave Sensors measure how the signal reflected from target area for a given set offrequency (1GHz to 3GHz). As can be seen, there is a big volume of data that is prone to error during repeated measurements and useful information in terms ofmutual variability between test subjects, targets, time points etc. In this study the follow-up parameters to monitor the physiological changes are identified and are extracted from the large volume of raw data. This is done by delimiting the initial data between 2.3 GHz to 2.6 GHz. It was seen from simulation, error estimation and previous works that the above-mentioned frequency range contains the needed information. Then the delimited data is averaged for its magnitude and phase with respect to frequency. An algorithm for finding the minimum value of the averaged delimited data (resonance) is implemented for the dB magnitude and compared withrespect to time points. A sub function is created to derive the polar coordinates(absolute magnitude, phase in radiance) and the Cartesian coordinates (in thecomplex plane). A preliminary analysis was performed on the processed data and some basic postulations were made. This work segregates the follow up parameters from raw data which can be used in future in depth analysis of clinical outcomes