Investigating the lack of representation of melanoma diagnosis in people of color

Summary: Abstract Body: Melanoma is one of the deadliest forms of skin cancer and the prognosis of the patient is dependent on time of diagnosis. It was found that the diagnosis of melanoma for people of color (POC) is typically in advanced stages with lower survival rates. This review investigates the underrepresentation of melanoma in POC and its impact on time to diagnosis. We conducted a systematic review of 3 databases according to PRISMA guidelines. Collected data from 7 studies included study characteristics, correct diagnosis, measurement methods, and source-identifying melanoma. The proportion of ethnicities and skin tones that were represented in each online source was disproportionately lower for images representing POC. The LaRosa study reported that only 16.66% of melanoma depicted in YouTube videos were shown on POC. Of this percentage, only African American, Asian, and Native American ethnicities were shown while no videos depicted melanoma on Hispanic patients. A similar search through Sadur’s 2024 study displayed only 13.2% of the images found from online searches of skin cancer were of patients with Pantone C-E or darker skin tones. The diagnostic scores of melanoma diagnosis for skin cancers on darker skin tones displayed that artificial intelligence (AI) has a better probability of correctly identifying skin cancer in patients with Fitzpatrick IV-VI skin tones. In the Lyman study, general practitioners scored 38% and 69% when diagnosing 2 melanoma pictures on “black skin. Schneider’s study reported that AI produced 77.78% and 83.33% scores for malignant neoplastic skin conditions. Scheider’s study also showed that AI produced scores of 69.57 and 82.61 for benign neoplastic skin conditions. The longer time for melanoma diagnosis for POC can be attributed to limited representation in media sites and online searches. AI may be used to close this disparity by increasing the likelihood of correctly identifying forms of skin cancer on all skin types. Further research is needed to establish a clearer relationship between the lack of representation of POC and its effect on skin cancer diagnoses. Chisom J. Nwosu¹, Hannah Chang¹, Jennifer Gullo¹ 1. California Northstate University, Elk Grove, CA, United States. Minoritized Populations and Health Disparities Research