Abstract 4021: Precision diagnostics for early melanoma detection using spatial biology and AI guided image analysis

Key Points

• The research aims to improve melanoma diagnosis by defining transcriptional states that distinguish benign from malignant lesions.
• Conducted spatially resolved whole transcriptome profiling on lesions ranging from healthy skin to melanoma.
• Performed single-cell RNA sequencing to identify marker genes and define cell types.
• Utilized Xenium and Orion platforms to validate gene expression and characterize immune cells in the tumor microenvironment.
• Analyzed 361 cancer genes to understand the mutational profile associated with melanoma progression.
• Developed a deep learning model integrating molecular data with spatial expression analysis for diagnostics.
• Identified changes in melanocyte cell states as melanoma progresses from benign to malignant.
• Established benign and malignant gene signatures that can be spatially resolved.
• Developed a deep learning model significantly enhancing diagnostic precision for melanocytic lesions.
• Characterized immune cell populations within the tumor microenvironment, revealing critical interactions.