This Week in Dermatology — Jun 13, 2026

Welcome to This Week in Dermatology. This week we're covering 7 notable papers spanning the rise and real-world performance of artificial intelligence, new approaches to prognostication in inflammatory disease and skin cancer, and the complex landscape of genodermatoses from molecular mechanisms to patient counseling. Let's dive in.

We begin with a major theme this week: artificial intelligence. As AI tools enter our clinics, critical evaluation is paramount. A paper in Nature Medicine directly addresses this by comparing specialized clinical AI tools against general-purpose, frontier large language models [1]. The researchers tested two clinical AI tools, OpenEvidence and UpToDate Expert AI, against models like GPT-5.2, Gemini 3.1 Pro, and Claude Opus 4.6. The evaluation was rigorous, spanning 500 medical knowledge questions, 500 items measuring alignment with clinician judgment, and a novel benchmark of 100 real, de-identified clinical queries submitted by physicians. The results were striking: across all three evaluations, the general-purpose frontier models outperformed the specialized clinical AI tools. In fact, on the real-world clinical queries, the specialized tools performed comparably to the auto-enabled Google Search AI Overview. This finding sends a clear message: specialized branding does not guarantee superior performance, and independent, real-world evaluation is essential before we integrate these tools into our workflow. But performance isn't the only metric to consider. A commentary in The Journal of Investigative Dermatology raises a crucial and often overlooked issue: the environmental impact of AI [6]. The authors remind us that training large AI models is an energy-intensive process, increasing electricity and water consumption for data centers. These environmental costs can disproportionately affect resource-constrained communities and run counter to our specialty's commitment to environmental stewardship, as outlined in the American Dermatological Association's policy statement on climate change. The paper advocates for a sustainable approach to AI integration. This includes selecting computationally efficient models, sharing datasets to avoid redundant training, and partnering with vendors who are transparent about their environmental footprint. Together, these two papers suggest a path forward for AI in dermatology that is both clinically effective and environmentally responsible.

Next, we turn to new approaches for understanding and risk-stratifying common skin diseases. In psoriasis research, a study in The Journal of Investigative Dermatology explores the potential of microdialysis as a minimally invasive tool for biomarker discovery [5]. Instead of disruptive skin biopsies, researchers used microdialysis to sample soluble mediators from intact skin in patients with psoriasis and healthy controls. Proteomic and metabolomic profiling revealed distinct molecular signatures in psoriatic lesions that correlated with disease severity. Specifically, they found elevated levels of proinflammatory proteins like SERPINB3 and FABP5, as well as metabolites involved in hyperproliferation, such as purines and polyamines. Crucially, these lesional profiles normalized in patients who responded to systemic therapy. This establishes microdialysate profiling as a promising platform for tracking therapeutic efficacy and personalizing treatment in a minimally invasive way. Shifting to cutaneous oncology, a paper in the British Journal of Dermatology questions the value of adding more granular histopathological details to predict metastasis in cutaneous squamous cell carcinoma [3]. Investigators conducted two large, nested case-control studies using the Netherlands Cancer Registry. They evaluated whether refined variables like morphological subtype, tumor budding, and mitosis count improved risk stratification beyond conventional factors like invasion depth and differentiation grade. In a population-based analysis, none of these refined variables showed a significant association with metastasis. In a second, risk-matched analysis, one variable—solar elastosis—was actually associated with a lower risk of metastasis. Severe and moderate solar elastosis were linked to a roughly 70% lower risk. The key takeaway for practicing dermatologists and pathologists is that expanding the histopathological assessment with these specific refined variables may provide limited additional prognostic value, and the surprising inverse association with solar elastosis warrants further investigation.

Our third theme delves into the world of genodermatoses, bridging molecular science with patient-centered care. From The Journal of Investigative Dermatology, a study provides a beautiful example of how a modifier gene can influence clinical phenotype [2]. The paper focuses on ILNEB, a severe subtype of junctional epidermolysis bullosa caused by biallelic ITGA3 variants, which is known for its clinical heterogeneity. Researchers studied two siblings with the same homozygous ITGA3 variant but strikingly different disease severity. Whole exome sequencing revealed that each sibling also carried a different heterozygous variant in ITGB4. The variant in the more severely affected sibling, p.Arg977Cys, caused a greater reduction in protein localization to the membrane and more significantly compromised a newly identified direct interaction between the ITGA3 and ITGB4 proteins. This work elegantly demonstrates that monoallelic ITGB4 variants can modulate the severity of ITGA3-associated disease, providing a molecular explanation for the observed clinical variability. Complementing this molecular focus, a qualitative study in Acta Dermato-Venereologica explores the real-world impact of genodermatoses on reproductive decision-making [7]. Through semi-structured interviews with 30 participants, the study found that the desire to avoid passing on their condition heavily complicates family planning. This preference was shaped by their own negative experiences with the disease, leading to fear and uncertainty about the potential severity in their offspring. A critical finding was a perceived gap in clinical care: participants highlighted a lack of routine reproductive counseling as a standard part of their management. The study underscores the substantial emotional burden of these conditions and calls for integrating reproductive counseling into the care of all patients affected by or at-risk for genodermatoses.

Finally, in procedural dermatology, a large retrospective study in the Journal of the American Academy of Dermatology compares two common treatments for port-wine stains on the face and neck: pulsed dye laser, or PDL, and hematoporphyrin monomethyl ether-mediated photodynamic therapy, or HMME-PDT [4]. The study included over 700 patients treated with HMME-PDT and over 650 with PDL. After adjusting for confounders, the results showed that HMME-PDT was significantly more effective. Patients in the HMME-PDT group had nearly triple the odds of achieving excellent improvement, defined as 75% or greater lesion clearance. However, this superior efficacy came at a cost. The incidence of hyperpigmentation was significantly higher with HMME-PDT at 5.4% versus 1.8% for PDL, and the risk of scarring was also higher, at 3.3% versus just 0.5% with PDL. This study provides important data for counseling patients, presenting a clear trade-off between higher efficacy and a greater risk of adverse events when choosing between these two modalities.

If you only have time for one paper this week, make it the evaluation of large language models in Nature Medicine [1]. It's a crucial, independent reality check on the performance of clinical AI tools being marketed to us, demonstrating that general-purpose models often outperform them and reinforcing the absolute necessity of rigorous, real-world validation before we adopt any AI into clinical practice.

Here are the key takeaways from this week in Dermatology: First, be skeptical of specialized clinical AI tools. Independent evaluation shows that general-purpose large language models may perform better, and rigorous, real-world validation is essential before clinical adoption [1]. Second, when considering AI, we must also weigh its environmental impact. Advocate for computationally efficient models and partner with vendors committed to sustainability to align technological innovation with our specialty's climate commitments [6]. Third, for high-risk cutaneous squamous cell carcinoma, adding refined histopathological variables like tumor budding may offer limited prognostic value over conventional factors. Keep an eye on the emerging data suggesting solar elastosis may be linked to a lower metastatic risk [3]. Fourth, proactively discuss reproductive options and offer counseling to patients with genodermatoses. They face significant dilemmas and have clear unmet needs for support in family planning [7]. And finally, for facial and neck port-wine stains, HMME-PDT offers superior clearance compared to pulsed dye laser, but this efficacy must be balanced against a higher risk of hyperpigmentation and scarring, which is a key part of the patient consent discussion [4].

That's your roundup for This Week in Dermatology. The full transcript and references are available on the episode page in your AudioScholar library. This is an AI-curated summary — for clinical decisions, always consult primary sources and current guidelines. See you next week.

If this weekly briefing is useful, follow the show in your podcast app so new episodes arrive automatically. And for audio briefings on your own clinical questions and papers, visit audioscholar dot C C.

This is an automated summary generated by artificial intelligence, which can make mistakes. Always review the original source materials.