Faculty Bibliography

Generative AI, particularly large language models (LLMs), is reshaping clinical workflows in dermatology. However, cloud-based commercial models pose persistent challenges to Health Insurance Portability and Accountability Act (HIPAA) compliance, especially in dermatology, where protected health information (PHI) extends beyond text to clinical photographs, dermoscopic images, and total-body photography that may capture identifiable anatomical features and document conditions carrying social stigma. Locally hosted, open-weight LLMs that are run within the institution's own infrastructure offer dermatology practices a pathway to leverage AI capabilities while retaining full control of their data. This viewpoint synthesizes evidence on when locally hosted, open-weight LLMs should be preferred for dermatologic workflows, when cloud deployment may remain preferable, and how multimodal AI fits into a coherent local deployment strategy. We advance 4 arguments. First, model compression techniques (knowledge distillation, structured pruning, and low-bit quantization) together with mixture-of-experts architectures have lowered hardware thresholds enough that 7- to 33-billion-parameter models now run on consumer-grade workstations with modest neural processing units or graphics processing units. Second, dermatology is fundamentally a visual specialty, and a credible local deployment strategy must integrate LLMs with vision models, including convolutional neural networks, vision transformers, vision-language models, and dermatology-specific foundation models such as PanDerm and medical multimodal models such as MedGemma. Third, locally hosted, open-weight models confer specific advantages for dermatology, including complete institutional control of clinical images, freedom from vendor model deprecation that disrupts validated workflows, and the ability to audit and fine-tune models to address well-documented performance gaps in skin of color. Fourth, local deployment is not a panacea; cloud models remain preferable for some tasks, and local deployment introduces governance challenges (heterogeneity across practices, model drift, and quantization-induced accuracy loss) that require structured mitigation through validated reporting frameworks such as CONSORT-AI (Consolidated Standards of Reporting Trials), SPIRIT-AI (Standard Protocol Items: Recommendations for Interventional Trials), DECIDE-AI (Developmental and Exploratory Clinical Investigations of Decision support systems driven by AI), and TRIPOD+AI (Transparent Reporting of a Multivariable Prediction Model for Individual Prognosis or Diagnosis), as well as retrieval-augmented generation and federated learning approaches. We situate these arguments within the international regulatory landscape, including the European Union's General Data Protection Regulation, the European Union AI Act, and Germany's Digitale Gesundheitsanwendungen (DiGA) framework, in addition to HIPAA. We provide quantitative cost examples showing that current consumer hardware capable of running 14- to 33-billion-parameter models can be acquired for roughly the price of 1 to 2 years of enterprise cloud-AI subscriptions. We close by mapping a practical implementation pathway and identifying near-term research priorities. Locally hosted, open-weight LLMs that are deployed thoughtfully and within governance frameworks offer dermatology practices a credible route to harness generative AI while preserving regulatory compliance, equity across skin types, and the dermatologist-patient relationship.

BACKGROUND:Limited access to dermatologic care may pose an obstacle to the early detection and intervention of cutaneous malignancies. The role of artificial intelligence (AI) in skin cancer diagnosis may alleviate potential care gaps. OBJECTIVE:The aim of this systematic review was to offer an in-depth exploration of published AI algorithms trained on dermoscopic and macroscopic clinical images for the diagnosis of melanoma, basal cell carcinoma, and cutaneous squamous cell carcinoma (cSCC). METHODS:Adhering to the Preferred Reporting Items for Systematic Reviews and Meta-Analysis guidelines, a systematic review was conducted on peer-reviewed articles published between January 1, 2000, and January 26, 2023. RESULTS AND DISCUSSION/CONCLUSIONS:Among the 232 studies in this review, the overall accuracy, sensitivity, and specificity of AI for tumor detection averaged 90%, 87%, and 91%, respectively. Model performance improved with time. Despite seemingly impressive performance, the paucity of external validation and limited representation of cSCC and skin of color in the data sets limits the generalizability of the current models. In addition, dermatologists coauthored only 12.9% of all studies included in the review. Moving forward, it is imperative to prioritize robustness in data reporting, inclusivity in data collection, and interdisciplinary collaboration to ensure the development of equitable and effective AI tools.

Cutaneous T cell lymphoma (CTCL) represents a heterogeneous group of extranodal non-Hodgkin lymphomas in which monoclonal T lymphocytes infiltrate the skin. The mechanism of CTCL development is not fully understood, but likely involves dysregulation of various genes and signaling pathways. A variety of treatment modalities are available, and although they can induce remission in most patients, the disease may recur after treatment cessation. Owing to relatively low incidence and significant chronicity of disease, and the high morbidity of some therapeutic regimens, further clinical trials are warranted to better define the ideal treatment option for each subtype of CTCL.