From the Journals

Patients accept artificial intelligence in skin cancer screening


 

FROM JAMA DERMATOLOGY

In a small survey, 75% of dermatology patients said they would recommend the use of artificial intelligence (AI) for skin cancer screening to friends and family members, but 94% emphasized the need for symbiosis between doctors, patients, and AI.

AI under investigation in dermatology includes both direct-to-patient and clinician decision-support AI tools for skin cancer screening, but patients’ perceptions of AI in health care remains unclear, Caroline A. Nelson, MD, of Yale University in New Haven, Conn., and colleagues wrote in JAMA Dermatology.

“We sought to elucidate perceived benefits and risks, strengths and weaknesses, implementation, response to conflict between human and AI clinical decision making, and recommendation for or against AI,” the researchers wrote.

They identified 48 patients seen from May 6, 2019, to July 8, 2019, at general dermatology clinics and melanoma clinics. This included 16 patients with a history of melanoma, 16 with a history of nonmelanoma skin cancer, and 16 with no history of skin cancer. The average age of the patients was 53.3 years, 54% were women, and 94% were white.

The researchers interviewed 24 patients about a direct-to-patient AI tool and 24 patients about a clinician decision-support AI tool.

Overall, 36 patients (75%) said they would recommend the AI tool to family and friends, with 17 patients (71%) saying they would recommend the direct-to-patient tool and 19 (79%) saying they would recommend the clinician decision-support tool. Another nine patients (19%) were ambivalent about the AI tools, and three patients (6%) said they would not recommend the tools.

Diagnostic speed and health care access were the most common perceived benefits of AI (by 60% of patients for each), and increased patient anxiety was the most common perceived risk (by 40% of patients). In addition, 69% of patients perceived more accurate diagnosis to be the greatest strength of an AI tool, and 85% perceived less accurate diagnosis to be the greatest weakness.

The study findings were limited by several factors, including the small sample size, qualitative design, use of a hypothetical rather than real-world situation, and a homogeneous study population, the researchers noted. However, the results merit more studies to obtain perspectives from diverse populations, they said.

“This expansion is particularly important in light of concerns raised that AI tools may exacerbate health care disparities in dermatology,” the researchers wrote.

From the patient perspective, the use of AI “may improve health care quality but should be implemented in a manner that preserves the integrity of the human physician-patient relationship,” the authors concluded.

“Although AI technology has not been widely implemented in dermatology yet, it is the pivotal time to assess patients’ views on the subject to understand their knowledge base, as well as values, preferences, and concerns regarding AI,” wrote Carrie L. Kovarik, MD, of the University of Pennsylvania in Philadelphia, in an accompanying editorial.

“Vulnerable patients, including racial and ethnic minorities, the underinsured or uninsured, economically disadvantaged, and those with chronic health conditions, may be at risk for improper consent for or use of AI,” she wrote.

Dr. Kovarik cited the position statement on augmented intelligence from the American Academy of Dermatology, which states that, for both patients and clinicians, “there should be transparency and choice on how their medical information is gathered, utilized, and stored and when, what, and how augmented intelligence technologies are utilized in their care process. There should be clarity in the symbiotic and synergistic roles of augmented intelligence and human judgment so that it is clear to the patient and provider when and how this technology is utilized to augment human judgment and interpretation.”

Clinicians will need to understand the perspectives on AI from patients of a range of backgrounds to achieve this goal, Dr. Kovarik said.

Dr. Nelson had no financial conflicts to disclose, but her colleagues disclosed relationships with pharmaceutical companies, government agencies, and nonprofit organizations. Dr. Kovarik disclosed serving on the artificial intelligence task force for the American Academy of Dermatology.

SOURCES: Nelson CA et al. JAMA Dermatol. 2020 Mar 11. doi: 10.1001/jamadermatol.2019.5014; Kovarik CL. JAMA Dermatol. 2020 Mar 11. doi: 10.1001/jamadermatol.2019.5013.

Recommended Reading

Sharp declines for lung cancer, melanoma deaths fuel record drop in cancer mortality
MDedge Hematology and Oncology
Lenvatinib/pembrolizumab has good activity in advanced RCC, other solid tumors
MDedge Hematology and Oncology
Circulating tumor cells at baseline predict recurrence in stage III melanoma
MDedge Hematology and Oncology
Global project reveals cancer’s genomic playbook
MDedge Hematology and Oncology
FDA: Cell phones still look safe
MDedge Hematology and Oncology
Smartphone apps for suspicious skin lesions unreliable
MDedge Hematology and Oncology
Banning indoor tanning devices could save lives and money
MDedge Hematology and Oncology
Tumor neoantigenicity metric improves prediction of response to immunotherapy
MDedge Hematology and Oncology
Data overwhelmingly support use of dermoscopy in practice
MDedge Hematology and Oncology
Largest meeting on cancer research canceled: AACR
MDedge Hematology and Oncology