How AI Is Diagnosing Skin Cancer Through Smartphone Photos

In this episode of Healthtech Talks, Lucas and Luna explore how deep learning models are now diagnosing skin cancer from smartphone photos with accuracy rivaling board-certified dermatologists. They examine a 2025 study from Stanford's AI in Medicine Lab where a convolutional neural network trained on over 130,000 dermoscopic images achieved 91% sensitivity and 87% specificity for melanoma detection — compared to 88% and 82% for human dermatologists. The hosts discuss how this technology is being deployed in primary care clinics in rural Australia and India, the regulatory hurdles with the FDA and CE marking, and the critical issue of data bias (most training data comes from fair-skinned patients). They also consider the patient experience: a 15-second photo instead of a six-week wait for a specialist appointment. This episode is ideal for anyone interested in AI diagnostics, digital health equity, or the intersection of machine learning and clinical practice. #SkinCancer #AI #DeepLearning #Telemedicine #Dermatology #Stanford #FDA #HealthEquity #PrimaryCare #Melanoma #DiagnosticAI #RuralHealth #SmartphoneDiagnosis #MedicalImaging #CNN #DigitalHealth #FexingoBusiness #BusinessPodcast Keep every episode free: buymeacoffee.com/fexingo