Artificial Intelligence and Skin Cancer Diagnosis

Original article can be found here (source): Artificial Intelligence on Medium

The Research Study

The study involved 48 patients at Harvard’s Brigham and Women’s Hospital, and melanoma clinics at Dana-Farber Cancer Institute over two months in 2019. The goal of the study, according to the authors, was to “explore how patients conceptualize [artificial intelligence] and view the use of direct-to-patient and clinician decision support AI tools for skin cancer screening.”

Results of the study were recently published in JAMA Dermatology.

Patients enrolled in the study were split into three groups: 1) those with a history of melanoma; 2) those with a history of non-melanoma skin cancer; and 3) those with no history of skin cancer.

When asked about the possible benefits of using AI to diagnose skin cancer, 60% said they believed the diagnostic process would be faster, while the same percentage said health care access would be increased.

Photo by Foto Sushi on Unsplash

Perceived Risks

Patients also some risks. Forty percent said patient anxiety could be an issue, especially in the direct-to-patient group. Other risks patients identified included loss of social interaction, patient privacy issues, and data collected by the AI being used for “nefarious” purposes.

One area that patients could not agree on was the accuracy of a diagnosis by AI. The authors noted that 69% said that the diagnosis would be the greatest strength, basing that on the ability to draw on more data and more experience. However, 41 patients viewed the accuracy of diagnosis as the greatest weakness, based on the possibility of false negatives, lack of context, and lack of physical examination among the answers given.