FIND-MEL: Developing an Application for Following Images of Nevi to Detect Melanoma
Abstract
Dr. Rotemberg’s goals in melanoma research are to become an independently funded physician-scientist who focuses on automatic ways to find melanoma. Dr. Rotemberg has been an advocate for high quality studies and public data for public use, and her goal is to continue to develop new technologies that can benefit patients in the most collaborative way that inspires future research all over the world. The research and career development plans in the attached proposal support Dr. Rotemberg’s research development as a leader in melanoma research as well as her technical development in technology research. She has partnered with Dr. Susan Swetter as a career guide, a stellar melanoma researcher, leader, and mentor. Dr. Dy is also a co-investigator on the award, who will continue to support Dr. Rotemberg’s development toward her own research laboratory. Automatic methodologies, such as artificial intelligence (AI), for detection of skin cancer is already better than many dermatologists, but it has mostly been tested outside of real-world conditions. Most of the research in this area has focused on dermoscopic images, which uses a specialized device that magnifies a skin lesion and uses polarized light. This proposal addresses the Melanoma Research Program (MRP) Focus Area of: Develop new tools for the detection and diagnosis of melanoma, which includes easily accessible technology (beyond the dermoscope) for primary care physicians and dermatologists. Melanoma is the deadliest form of skin cancer, but it can be cured with surgery if caught at an early stage. We see AI to use patient photos to help find melanomas as soon as possible. We have designed a study that will use all different types of photographs to improve automatic detection of melanoma on photographs captured by any smartphone or digital camera. The study addresses the fiscal year 2022 MRP Challenge Statement by focusing on both improved detection and AI-enabled lesion monitoring to detect melanoma as early as possible. This will target diagnosing melanoma prior to invasion and preventing the development of metastases. This research is very applicable for improving melanoma detection in melanoma patients and those who are at high risk for developing melanoma. We know patients who have had melanoma and those with other risk factors are at higher risk for developing melanoma. Through this project, we will take advantage of photographs that have been collected in dermatology clinics to improve the ability of AI to find melanomas remotely through a cell phone or digital camera photograph. The potential benefits, especially for patients who don’t have access to a dermatologist nearby (including Veterans and deployed military personnel) are especially valuable, because most spots on a patient s body will end up being benign. Having an accurate way to rule out melanoma from a cellular phone without going to a dermatologist in person would be very valuable. However, it means these technologies must be very accurate before we use them directly. We hope that the techniques developed in this proposal will be used for rigorous clinical trials in the future so they can be comprehensively studied and deployed. In the short term, our goals are to develop and validate a method for monitoring photographs of skin lesions without in-person dermatologist examination. We will then use this data to support development in the community by making the algorithm open source as well as use this work as preliminary data for a prospective clinical trial via additional funding. Our long-term goals are that this type of approach to monitor lesions over time reduces unnecessary biopsies and improves our remote ability to find melanoma. We hope that this will especially help patients in remote and austere areas, especially our Veterans and active-duty personnel, by reducing their need to find a dermatologist when they are doing well, and making triage of dangerous lesion
Document Details
- Document Type
- DoD Grant Award
- Publication Date
- Jan 04, 2024
- Source ID
- HT94252310848
Entities
People
- Veronica Rotemberg
Organizations
- Sloan-Kettering Institute
- United States Army