Optimizing electrical impedance myography outcomes through data mining

Period of Performance: 09/20/2017 - 02/28/2018

$150K

Phase 1 SBIR

Recipient Firm

Convergence Medical Devices, Inc.
SAN FRANCISCO, CA 94107
Principal Investigator
Principal Investigator

Abstract

Project summary Electrical impedance myography (EIM) is a non-invasive technology for the assessment of muscle that is based on the application of a weak, high frequency electrical current to a muscle and the measurement of the resulting surface voltages. The further development and application of EIM remains the main business focus of Skulpt, Inc, a small business concern based in Boston and San Francisco (Specific Aims just say San Francisco). Alterations to the condition of the muscle, including myocyte atrophy, fat and connective tissue deposition, and inflammation all alter the EIM data in predictable and consistent ways. To date, through Skulpt, EIM has been applied as a potential biomarker for assessing disease progression and response to therapy in a wide variety of neuromuscular disorders, including amyotrophic lateral sclerosis, Duchenne muscular dystrophy, and spinal muscular atrophy, as well as other disorders that impact muscle condition, such as disuse atrophy and sarcopenia (age related muscle loss); over 1000 people have been studied with Skulpt?s EIM technology. Whereas the results of these applications are promising, the analytic approaches taken to the data sets have been fairly basic, utilizing only simple single frequency or simplistic multifrequency values. However, with every single muscle measurement, over 240 individual data points are acquired at different frequencies, different depths of muscle penetration, and at different angles to the major muscle fiber direction. Moreover, each of the above studies has been done in isolation, and thus how results differ between diseases is unknown. Given the plethora of data, applying more sophisticated analytic approaches has the potential of yielding improved EIM measures. Moreover, collaborators have already collected an associated wealth of animal EIM data that will help further inform this analysis. Thus, in this proposed Phase 1 SBIR, we plan to apply a variety of data mining techniques to the vast set of data already accumulated at Skulpt, Inc such that improved EIM outcomes can be developed and implemented. In Specific Aim 1, we will study human data across all disease types evaluated to determine which data sets are most effective at discriminating diseased from healthy muscle as well as distinguishing between diseases. In Specific Aim 2, we will focus on finding the metrics that are most sensitive to the degree of muscle pathology in a specific disease. In both of these aims, we will evaluate how these new metrics are mirrored in already obtained animal data. In Specific Aim 3, we will study these metrics in a new set of data (a test set) that was not used to develop the analytical paradigms so as to ensure their robustness. With the conclusion of this work, we will plan to pursue a Phase 2 SBIR that will focus on the development of a software suite to assist in EIM data interpretation based upon these results followed by a prospective observational clinical study to evaluate the efficacy of these newly developed metrics for disease diagnosis and tracking of progression/response to therapy.