How AI Is Transforming Skin Cancer Diagnosis

In an engaging session at the SDPA Fall Meeting in San Antonio, Texas, attendees were invited to imagine the near future, one where artificial intelligence (AI) becomes an everyday partner in clinical care, according to presenter, Renata Block, DMSc, MMS, PA-C, board-certified PA from Chicago, Illinois, and instructor at Rush University Medical Center.¹

“I'm going to be talking about your future,” she began. “And we all need to be prepared, because it is coming, whether we like it or not.”

From Artificial to Augmented Intelligence

While “artificial intelligence” is the more popular term, Block argued that health care should instead focus on augmented intelligence, AI designed to support rather than replace clinicians. In this model, human expertise remains central while the machine provides additional insights that improve accuracy, efficiency, and access to care.

Augmented intelligence, Block explained, enhances human decision-making rather than automating it. It’s a digital assistant that never tires, scans thousands of images in seconds, and helps clinicians focus on what matters most, which is the patient.

The Tools Behind the Technology

The talk walked the audience through the key branches of AI already shaping dermatology:

• Natural Language Processing (NLP): Used to summarize and interpret clinical documentation and patient histories.
• Machine Learning: Supervised models are trained on labeled data to predict outcomes, while unsupervised models identify hidden patterns in data.
• Deep Learning: The same technology behind facial recognition and voice assistants is now used to interpret dermoscopic and histologic images with remarkable precision.
• Generative AI: Powering note-writing and report drafting tools, though clinicians must stay alert for “hallucinations” and ensure data privacy and accuracy.
• Robotics and Automation: This includes hospital robots to dermatologic instruments that move specimens or assist with imaging to improve workflow efficiency.

Each of these technologies, Block noted, contributes to a larger goal of faster, more objective, and more equitable dermatologic care.

Promise and Pitfalls of AI

According to Block, the opportunities are enormous—earlier detection, shorter wait times, and broader access, especially for underserved patients. But she cautioned that data bias remains a serious concern. Many AI models have been trained predominantly on lighter skin tones, which can lead to errors and inequities for patients with Fitzpatrick skin types V and VI. Ensuring diversity in training data and conducting fairness testing must remain top priorities. Transparency, accountability, and patient privacy are equally critical.

“We have to conceptualize AI that focuses on an assistant role so it enhances our intelligence, right? It's not going to replace us, and it shouldn't replace us.” Block said.

Devices on the Horizon

Several FDA-cleared and emerging technologies were highlighted in the session. These tools aim to streamline the pathway from suspicious lesion to diagnosis, potentially cutting months off the traditional referral timeline.

• DermaSensor and Nevisense use light-based spectroscopy to analyze lesions and produce a risk score for malignancy.
• Reflectance Confocal Microscopy (RCM) and Multiphoton Microscopy can visualize skin at the single-cell level, though currently confined to academic settings.
• Digital dermoscopy platforms are integrating AI for triage, enabling primary-care physicians to flag lesions for dermatology referral.
• Future devices, such as SKLIP awaiting FDA Approval, will integrate both AI and Augmented AI by providing stand-alone devices to hone in on the three-point check list in diagnosing melanoma and telling the clinician why it is suspicious. Original research found it increased the specificity and sensitivity when using the device compared to traditional dermoscopy alone.

Navigating the Regulatory Landscape

The session also decoded the FDA’s device-classification framework of Class I (hardware without AI output), Class II (AI-assisted triage), and Class III (standalone diagnostic systems). Many new dermatologic devices follow the De Novo pathway, which allows novel but moderate-risk devices to reach the market once validated for safety and effectiveness. The FDA now demands multi-center trials, demographic diversity, reproducibility testing, and human-factors analysis to ensure the tools work across populations and clinical settings.

Adoption and Affordability

Beyond performance, adoption depends on cost and reimbursement. Subscription-based software models, uncertain CPT coding, and unclear insurance coverage can make otherwise promising tools hard to sustain. As Block put it, “If it’s not affordable, it’s not sustainable.”

Still, with proper validation and equitable design, AI offers tremendous potential to augment the dermatology workforce, especially amid growing demand for skin-cancer screening and limited specialist availability.

The Human Element Remains Central

The session concluded with a reminder that while AI could transform dermatology, it will never replace humans. Patients still want compassion, communication, and accuracy. The future, Block said, belongs not to machines alone but to clinicians who learn to work with them.

“We all want to be treated with empathy and a human touch,” she concluded.

Reference

1. Block, R. AI Utilization and Management Tools for Dermatology. Presented at: Society of Dermatology Physician Associates Fall 2025 Conference; November 5-9, 2025; San Antonio, Texas