Melanoma AI
Melanoma Prognostication
ML models for primary melanoma recurrence using clinical and histopathologic features, achieving AUROC 0.845 internal and 0.812 external validation.
Learn moreThe Semenov Lab develops machine learning models for melanoma prognosis, immunotherapy toxicity prediction, and digital pathology — translating computational breakthroughs into clinical impact.
Our lab sits at the intersection of clinical medicine, artificial intelligence, and genomics — building precision dermatology tools that improve outcomes for patients with melanoma, skin cancer, and inflammatory skin conditions arising from cancer therapy.
ML models for primary melanoma recurrence using clinical and histopathologic features, achieving AUROC 0.845 internal and 0.812 external validation.
Learn more
Multiplex imaging and spatial transcriptomics of melanoma tissue reveal immune cell interactions, stromal features, and multiomic biomarkers of treatment response and disease progression.
Learn more
Seminal epidemiologic studies on cutaneous immune-related adverse events (irAEs) — skin toxicities from cancer therapy including rash, pruritus (itch), and inflammatory skin conditions from immune checkpoint inhibitors — and their role as biomarkers of immunotherapy response.
Learn moreOur mission is to harness the power of artificial intelligence and multi-modal clinical data to predict, prevent, and personalize treatment for patients with skin disease.
Dr. Semenov is a dermatologist and clinician-scientist with expertise in applied mathematics, biostatistics, epidemiology, and clinical informatics. He is co-Director of the MGH Oncodermatology Program and Assistant Professor of Dermatology at Harvard Medical School, where he has built a nationally and internationally recognized research program at the interface of dermatology, oncology, and AI.
He co-chairs data governance for ASPIRE, a national irAE registry, and serves on the Scientific Advisory Board of the RARE Registry for rare melanoma. He leads a federally funded dermatology informatics research program and serves as Statistics Editor for JAAD, Associate Editor for the International Journal of Dermatology, and Associate Editor for npj Digital Medicine. He holds active leadership roles including: DataDerm Oversight Committee, American Academy of Dermatology; Co-Chair of Research, Oncodermatology Society; Co-Chair of Data Governance, ASPIRE (Alliance for Support and Prevention of Immune-Related Events); and member of the RARE Registry Oversight Committee through the Melanoma Research Alliance.
From MGH on Fruit Street to Harvard Medical School, our work spans the most innovative biomedical research ecosystem in the world.
The lab operates across multiple Harvard-affiliated institutions in the greater Boston area, enabling deep collaboration across clinical care, cancer research, and computational biology.
Dr. Semenov is a board-certified dermatologist and internationally recognized expert in cutaneous oncology and oncodermatology. He provides comprehensive, multidisciplinary care for patients with melanoma and skin cancer across all stages of disease, patients experiencing skin side effects from immunotherapy and other cancer treatments, and patients with complex immune-mediated skin disorders. He has contributed to the development of clinical guidelines for the diagnosis and management of immunotherapy-related skin toxicities and receives referrals from across the United States. Since its launch in 2022, the MGH Oncodermatology Program has grown to 600+ referrals annually and serves as a national model — adopted by institutions across the US and internationally.
