What's New In Digital Health

Artificial intelligence (AI) is driving the development of many industries through a revolution in machine learning technology.¹ Within dermatology practices, beyond its proven capacity for assisting in diagnostic procedures, these deep learning algorithms are also posing questions regarding the extent of automation in doctor-patient relationships along with the associated ethical implications.2 This article will introduce some of the latest advances of AI in dermatology and discuss their impact on the healthcare environment.

1. What is the status of AI in dermatology?
The rapid progress in machine learning, modern algorithms have been designed with the ability to choose their own “learning strategies”, much like a human being would.² In the field of dermatology, machines trained with a wide range and number of dermoscopic images have proven successful in identifying skin cancer lesions. This specially designed algorithm was shown to provide similar accuracy to a group of selected physicians, some of them with years of experience.³

The increased accuracy in the tasks performed are a great leap forward in the early detection of skin cancer and other conditions. Nonetheless, these impressive results have sparked regulatory question relating to the extent of machine interaction in patient/doctor relationships.⁴ Could machines completely displace human doctors from their roles? Given that values like respect, altruism or compassion are inherent human traits, the need for human interaction remains an inflexion point in offering best care practices.

2. AI-powered dermatology software and the future of practices
Intelligent dermatology software such as DermEngine are leading the trend in the application of AI tools to complement the work of dermatologists. As discussed before, the utilization of advanced searching functionalities has transformed the way image analysis can assist in diagnosis. DermEngine’s Visual Search tool is based on a deep learning strategy known as content based image retrieval (CBIR), a powerful algorithm that compares a given image to thousands of previously identified ones. The ease and swiftness with which machines can accurately provide similar images can be used to a doctor’s advantage.

Despite the positive impact that machines have shown in improving healthcare practices skepticism among professionals still exists. Unavoidably, the automation of many processes may imply a shifting in the way tasks are assigned and executed, involving several stakeholders in the process. The ultimate outcome, however, is to offer effective workflows which translate into:

1- Improved patient well-being: by applying these new technologies to daily practices, better, faster and more accurate diagnoses can be achieved. Assisting with some of the most time-consuming tasks can leave precious time dedicated to patients, developing bonds that turn doctors into trusted advocates.

2- Renewed sense of progress: throughout history, the advancements of new technologies always imply change. These evolutions can mean shifts in positions and responsibilities of every participant in the healthcare cycle. However, it is important to recognize the advantages that AI can bring and embrace them as complementary and not a replacement.

3- Focused approach on human interaction: the automation created by AI-powered tools can have the unexpected advantage to provide doctors with the time and resources needed for investing in their natural abilities. A good portion of healthcare is based on developing proper interpersonal skills. Focusing on the very essence of the care system, that is fellow human beings, can only bring dermatology practices forward.

Conclusion
The quest to improve healthcare practices with technology has seen the rise of AI as a revolutionary force. The potential that machine learning has to transform the current practices in fields like dermatology are proving both positive and challenging. While the benefits cannot be overlooked, there is still much to be defined in the years to come. However, beyond the many changes arising from AI’s new applications, one thing seems clear: only the complementation of human activity with the new technologies will offer patients the highest quality of care.

1. https://bmjleader.bmj.com/content/2/2/59
2. https://www.nature.com/articles/s41746-017-0012-2
3. https://www.nature.com/articles/nature21056
4. https://www.nature.com/news/there-is-a-blind-spot-in-ai-research-1.20805