Research Project
9103032
DESCRIPTION (provided by applicant): Melanoma, the fastest growing cancer worldwide, kills more than one person every hour in the United States and costs more than $2.4 billion per year. Yet, if detected at an early stage, it can be easily cured with a tenfold improvement in survival rate and a hundredfold reduction in treatment cost. Current methods have not yielded the breakthroughs needed, either because they are too simple (visual examination of skin, with or with- out digital imaging) or because the required high sensitivity comes at the expense of lowered specificity, even in emerging advanced methods. What is necessary to address these unmet needs is a more sophisticated approach. We propose such an approach, based on analysis of the current state-of-the-art in melanoma diagnosis and feedback from clinicians, as well as concepts from the well-established field of risk management. We concluded that a combination of several imaging methods is needed, brought together onto the same instrument, in a synergetic multimode set. Under other funding, we developed a research prototype multimode dermoscope, with (powerfully diagnostic) hyperspectral imaging as its central method. This proposal outlines how we plan to turn this device (SkinSpect) into a useful clinical tool for better (earlier and more reliable) diagnosis of melanoma. The entire Spectral Molecular Imaging Inc. (SMI) team has the extensive experience needed, and will be joined by prestigious groups at University of California Irvine - Beckman Laser Institute and Memorial Sloan-Kettering Cancer Center, bringing valuable expertise/contributions in key areas. Our diagnostic instrument combines hyperspectral, polarization, and fluorescence imaging of pigmented skin lesions (nevi), integrated in a high-resolution, non-invasive dermoscopic imaging system. We can extract three-dimensional physiological and morphological diagnostic measures including lesion depth (Breslow thickness) and the concentration and distribution of skin/nevus components such as melanin, collagen and oxy/deoxy-hemoglobin. We will focus on the accuracy with which we can quantitate features important for lesion modeling and diagnosis, with emphasis on improving specificity. Building on our previously developed research prototype, we will manufacture and technically validate an improved clinical prototype (Phase I) and will evaluate this system in three clinical trials (Phase II). We will (1) characterize the system with tissue phantoms, test algorithms for compositional mapping and human volunteers; (2) build SkinSpect clinical prototypes under FDA-type tight design controls; (3) optimize comprehensive tissue analysis software using clinical trial data; (4) demonstrate relative efficacy of each of the imaging modes (including some competing ones) versus our full multimode combination. The desired overall outcome is a truly advanced yet cost-effective instrument to be adopted by clinicians fighting melanoma for saving lives, productivity and costs.
