++
SUMMARY
Artificial Intelligence (AI) has demonstrated the potential to augment the practice of dermatology.
The performance of AI has been comparable to board-certified dermatologists in the identification of squamous cell carcinoma, basal cell carcinoma, and melanoma.
++
DON’T FORGET
AI is a new and evolving field in health care and the management of skin cancers.
The focus on development has been placed on augmenting dermatology practice, not replacing dermatologists.
++
CLINICAL PEARLS
AI and augmented intelligence show promise in delivering greater skin cancer detection, democratizing skin cancer detection for patients with limited health care resources, and likely increasing the volume of skin cancer surgeries as a result.
++
PATIENT EDUCATION POINTS
++
Artificial intelligence (AI), a technique that enables a machine to mimic human cognitive functions, has grown tremendously in the last few years across many industries, and the health care industry is no exception. Advances in imaging technology, computational processing, and graphics-processing unit technology have led to significant growth in the industry of AI, augmented intelligence, and machine learning. Machine learning is a technique to achieve AI through algorithms trained with data. Deep learning is a subset of machine learning, both of which are subsets of AI (Figure 64-1). Deep learning utilizes an artificial neural network loosely modeled on the human brain that requires a massive volume of data to train. Deep learning is extremely useful when dealing with unstructured data. Health care and medicine stand to benefit immensely from deep learning due to the sheer volume of data being generated—153 exabytes in the United States alone in 2013, with a growth rate of at least 48% annually,1—as well as the proliferation of medical devices and digital record systems.2
++++
In the future, most clinicians will utilize AI technology, specifically deep learning. This complex pattern recognition that uses deep neural networks (DNNs) can help decipher medical scans, skin lesions, retinal images, faces, electrocardiograms, vital signs, and perhaps more in the future.
++
Dermatology has emerged as a field of medicine well suited for early AI models given the visual nature and pattern recognition of many diagnoses (Figure 64-2). Skin cancer, for example, is usually first suspected based on visual skin examination. Furthermore, skin cancer is the most commonly diagnosed cancer in the United States. Since over 5.4 million new cases are diagnosed annually, one in five Americans will be diagnosed with a cutaneous malignancy in their lifetime.3 Although melanoma represents fewer than ...