STTR Phase I: Intelligent Scoring of Inflammatory Skin Disease Progression

The broader impact and commercial potential of this Small Business Technology Transfer (STTR) Phase I project is to empower providers with the unprecedented ability to treat patients with chronic inflammatory skin disease effectively and at a lower cost. Today, primary care physicians often refer patients with these conditions to dermatologists for the initial diagnosis and multiple follow-up appointments, especially when patients are prescribed high cost, high touch therapies like biologics or procedures, such as phototherapy. By introducing this automated severity scoring system, these providers will have the expertise of dermatologists at their fingertips, giving them the ability to assess disease severity and provide systemic treatments previously available only through specialists. It will also establish more objective assessments, evaluating patient progress to adjust dosage and treatments as necessary. For pharmaceutical companies, the ability to have non-specialists make these severity assessments during clinical trials reduces expensive and restrictive staffing requirements. Less expensive trials can help expedite new drugs to market. This innovation will enhance scientific and technological understanding by applying known machine learning techniques in a novel manner to solve a difficult visual problem. This Small Business Technology Transfer (STTR) Phase I project will prototype a clinical tool that can assess images of a chronic skin inflammation, determine its severity, and suggest treatment. This system will be able to guide users with minimal training through the image collection process, determine if the information is sufficient, and request additional information if required via novel machine learning techniques. Currently, clinicians trying to assess the severity of chronic inflammatory skin diseases rely on a series of estimations and manual weighted averages, a time consuming and biased process. This Phase I research will explore what information is necessary for an algorithm to determine the severity of a psoriasis case. It will determine if it is possible for a machine to guide an imager through imaging regions of interest at higher detail, rather than the entire body at large. It will also explore how quickly and practically such a calculation can be performed. The end goal of the project is a working prototype that can consistently score a few psoriasis cases and serve as a proof of concept for expansion of the prototype in future work. This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.