Research & development projects

Electro-thermo-mechanical phenotyping of breast tissues for pre-surgical assessment

Role: Ph.D. thesis work; Institute: IISc, Bangalore; Clinical Collaborator: Dr. Gayatri Gogoi, Assam Medical College

Project summary

Breast cancer accounts for 25% to 32% of all female cancers in all major cities of India. Breast pathology is conventionally assessed through biochemical analysis of tissue biopsies using techniques such as histology and immunohistochemistry, which takes anywhere between 48 to 96 hours to give the results. Technologies that can perform a rapid assessment of such biopsy tissues to delineate between benign, malignant, and normal types can provide significant value addition to routine diagnostics and help manage the caseload. My Ph.D. work involved the indigenous design, development, and clinical validation of a semi-automatic electronic system (RapidET system) capable of comprehensively analyzing the biophysical signatures of breast tissue to discern its pathology. Starting from the input requirements of the most probable sample dimensions and diagnostic resolution, a systems thinking approach was used to design and build the various components of the semi-automatic platform. Microsensors capable of measuring the electrical, thermal, and mechanical properties of the sample were integrated and interfaced with electronic modules for control, actuation, and data acquisition. A software application running on a tablet computer controls the entire system. 

The developed platform helped to assess three important aspects of the biophysical nature of breast tissues. The bulk and surface resistivity of tumours showed a significant increase with temperature compared to the normal. A bimodal temperature and frequency-dependent electrical characterization also revealed a significantly larger scaling in tumours. Finally, a combined electro-thermo-mechanical characterization of the tissues coupled with a Gaussian process classifier could delineate samples into three types–benign, malignant, and normal with low root mean square error.