Canadian Dermatology AI in 2025: Year-End Industry Retrospective

Q: What is the trend in dermatology in 2025?

The defining trend in dermatology in 2025 was AI moving from academic research into early clinical consideration. Clinics across Canada began evaluating AI-assisted triage tools, teledermatology expanded into underserved communities, and Health Canada published clearer frameworks for AI as a medical device. The Canadian Dermatology Association highlighted AI-assisted lesion classification at its 2025 annual meeting, signalling that the field now treats AI as a practical clinical tool rather than a future prospect. DermaDex is among the companies working in this space, building toward production deployment in Canada.

Q: How will technology change dermatology in the future?

Technology will change dermatology primarily by improving how fast a patient reaches the right level of care. AI triage allows primary care physicians to confidently handle more dermatology cases independently, reserving specialist time for complex presentations. Federated learning lets AI models improve across clinic networks without moving raw patient data, which matters under Canadian privacy law — PHIPA (Ontario) and PIPEDA (federal). The next major shift will likely be multimodal AI that combines imaging data with patient history from EHR systems. The CIHI has documented wait times exceeding 26 weeks for dermatology in some provinces — that structural problem is what drives development work across the Canadian healthtech industry.

Q: What's new in dermatology in 2025?

In 2025, the notable developments in dermatology technology included AI classifiers trained across diverse skin tones to address documented gaps in historical datasets, federated learning research protecting patient privacy across multi-clinic networks, and tighter integrations between AI diagnostic tools and existing EMR platforms. Health Canada published updated Software as a Medical Device (SaMD) guidance that created a clearer regulatory path for AI tools in Canada. Research published in journals like JAMA Dermatology continued to validate AI classifier performance, particularly for basal cell carcinoma and melanocytic lesions.

Q: Is AI going to replace dermatologists?

No. AI tools are designed to support dermatologists and primary care physicians, not replace them. AI classifiers can flag lesions, assign confidence scores, and prioritize cases by urgency, but all clinical decisions — biopsy, treatment, patient communication — remain with the physician. Canada's dermatologist shortage is a workforce distribution problem documented by CIHI and Statistics Canada. AI helps primary care physicians handle more dermatology cases confidently and makes specialist referrals more efficient. Responsible AI tools in this space should always include confidence scores and explicit recommendations to consult a physician for any ambiguous result.

As of December 16, 2025.

Canadian dermatology care ran into a familiar bottleneck in 2025: too few specialists, too many patients, and wait times that stretched well past six months in several provinces. Across the industry, healthtech companies and academic research groups spent the year building tools designed to close that gap — not by replacing dermatologists, but by giving clinicians better information sooner. This article looks at what happened in Canadian dermatology artificial intelligence (AI) broadly in 2025, what regulators and professional bodies published, and where the field is heading.

What is the trend in dermatology in 2025?

Short answer: The dominant trend in dermatology in 2025 was artificial intelligence (AI) moving from research settings into early clinical workflows. Canadian clinics began evaluating AI-assisted triage tools, teledermatology platforms expanded coverage into rural communities, and regulatory bodies like Health Canada published clearer guidance on software as a medical device (SaMD). Professional conferences began treating AI not as a theoretical concept but as a practical clinical tool under active evaluation, and federal working groups turned attention toward defining performance benchmarks for AI-assisted dermatology in a Canadian context.

At the Canadian Dermatology Association (CDA) annual meeting in June 2025, multiple presentations highlighted machine learning (ML) models for lesion classification — a shift from prior years, when AI in dermatology was still largely theoretical at Canadian conferences. The Canadian Institute for Health Information (CIHI)'s specialist wait-time data has documented specialist shortages in every province, and AI triage is now part of the mitigation conversation at a federal level.

The table below summarizes specialist wait-time data from CIHI's most recent reporting, which covers the 2024 reference period and provides the access-gap context motivating AI investment across Canadian healthtech.

Province	Median specialist wait (weeks)	Context
Prince Edward Island	40+ weeks	Highest reported median in CIHI data
New Brunswick	35+ weeks	Limited specialist density outside Fredericton/Moncton
Nova Scotia	26–30 weeks	Rural communities face longest waits
Manitoba	24–28 weeks	Northern communities severely underserved
Ontario	18–24 weeks	Urban centres shorter; Northern Ontario significantly longer
British Columbia	20–26 weeks	Interior and Northern BC longest
Alberta	18–22 weeks	Rural-urban gap documented

Source: CIHI Wait Times for Priority Procedures in Canada. Figures are approximate ranges based on CIHI's reported provincial medians for specialist (non-emergency) referrals. Dermatology-specific data varies; these figures reflect specialist care broadly.

What's new in dermatology in 2025?

Short answer: In 2025, the notable advances in dermatology technology included multimodal AI classifiers combining dermoscopy images with patient history, federated learning research protecting patient privacy across multi-clinic networks, and deeper integrations between AI tools and electronic medical record (EMR) systems. Health Canada's updated Software as a Medical Device (SaMD) framework gave industry a clearer regulatory path. Research in journals including JAMA Dermatology continued to validate AI classifier performance on common lesion types, and the field moved toward explicit calibration across Fitzpatrick skin tone diversity as a publication standard.

Health Canada's updated Software as a Medical Device guidance provided a clearer regulatory framework for AI diagnostic tools in Canada, which accelerated serious development work across the industry. Research published in journals like JAMA Dermatology and Nature continued to validate AI classifier performance on common conditions including basal cell carcinoma and melanocytic lesions, with particular attention to ensuring calibration across diverse skin tones — a documented gap in earlier training datasets.

Federated learning emerged as the preferred technical architecture for multi-clinic AI research under Canadian privacy law — PHIPA (Personal Health Information Protection Act) in Ontario and PIPEDA (Personal Information Protection and Electronic Documents Act) at the federal level — allowing model improvement across clinic networks without raw patient data leaving each site. This is an area of active research and early-stage work across Canadian healthtech, including by DermaDex as the company works toward production readiness.

How will technology change dermatology in the future?

Short answer: Technology will change dermatology by compressing the time between a patient's first concern and a qualified clinical assessment. AI triage allows primary care physicians to independently manage a larger share of dermatology presentations, routing only the most complex and ambiguous cases to specialist review. Asynchronous teledermatology extends specialist reach into rural and remote communities where in-person access is limited. The next major technical frontier is multimodal AI — combining imaging data with patient history, medications, and prior diagnoses from electronic health record (EHR) systems to improve classifier accuracy and reduce false-positive referrals.

The Canadian Institute for Health Information (CIHI) reported that specialist wait times for dermatology exceeded 26 weeks in some provinces in 2024, a figure the Canadian Medical Association (CMA) has cited as a patient safety concern. AI-assisted triage does not shorten that wait for patients who need a specialist appointment, but it does allow primary care physicians to confidently manage a larger share of cases independently. The cases that do require specialist review arrive with more context, which makes those appointments more efficient.

Research published in 2025 continued to show strong sensitivity and specificity for AI classifiers on common conditions, particularly basal cell carcinoma and melanocytic lesions. The integration of EHR data with imaging AI is broadly viewed as the next technical frontier. DermaDex is among the companies building toward this capability.

Is AI going to replace dermatologists?

Short answer: No. AI tools in dermatology are designed to help dermatologists and primary care physicians work more effectively, not to replace specialist judgment. AI classifiers can flag lesions that warrant closer inspection and prioritize cases by urgency, but the clinical decision — including biopsy, treatment selection, and patient communication — stays with the physician. The evidence base for AI-assisted dermatology consistently frames these tools as a triage and decision-support layer. Canada's dermatologist shortage is a workforce distribution problem; AI helps extend the reach of the specialists who do exist, particularly into rural and remote communities.

Per Statistics Canada data on health workforce distribution, rural and remote communities have the fewest specialists and the longest waits. AI tools help extend the reach of the specialists who do exist. A family physician using an AI-assisted tool can assess a lesion with confidence scoring, document it in their EMR, and make a more informed referral decision — or choose to monitor and reassess — without requiring an immediate specialist consult.

Dermatologists bring pattern recognition built over years of clinical exposure, the ability to ask follow-up questions, and the judgment to weigh factors that do not appear in an image. AI contributes speed and consistency at scale. The two work better together than either does alone. Responsible AI tools in this space should include confidence scores and explicit recommendations to consult a physician for any ambiguous or high-risk result.

Sources

Canadian Institute for Health Information (CIHI). Wait Times for Priority Procedures in Canada. https://www.cihi.ca/en/wait-times-for-priority-procedures-in-canada
Canadian Medical Association (CMA). CMA Health Care Access Reports. https://www.cma.ca
Statistics Canada. Health Workforce Distribution in Canada. https://www.statcan.gc.ca
Health Canada. Software as a Medical Device Guidance. https://www.canada.ca/en/health-canada/services/drugs-health-products/medical-devices/application-information/guidance-documents/software-medical-device-guidance-document.html
Esteva A et al. Dermatologist-level classification of skin cancer with deep neural networks. Nature, 2017. https://pubmed.ncbi.nlm.nih.gov/28117445/

Year in Review: Canadian Dermatology AI in 2025