Conventional skin imaging modalities are often bulky, expensive, and impractical for routine dermatology practice. There is a need for a portable, multimodal imaging tool that integrates high-resolution surface and subsurface visualization at the point of care. The aim of this study was to describe the design, technical capabilities, and clinical application of dermoscopy-guided high-frequency ultrasound and to evaluate its performance across a range of dermatologic conditions. Dermoscopy-guided high-frequency ultrasound was applied to a total of 130 lesions from 122 patients at the Department of Dermatology, Semmelweis University (Budapest, Hungary); Université Libre de Bruxelles (Brussels, Belgium); and Hospital Clínic, Universidad de Barcelona (Barcelona, Spain). The examined cases included malignant skin cancers and inflammatory disorders. Dermoscopy-guided high-frequency ultrasound enabled simultaneous visualization and correlation of surface dermoscopic patterns with underlying structural alterations in real time. The device identified disease-specific imaging features for both malignant and inflammatory lesions. Artificial intelligence-based segmentation improved image interpretability. Dermoscopy-guided high-frequency ultrasound bridges a critical gap between surface and subsurface dermatologic imaging, offering a practical, portable, and cost-effective solution that could enhance noninvasive diagnosis and management in dermatologic care.

Extramammary Paget disease (EMPD) is an uncommon adenocarcinoma of apocrine gland-rich areas, presenting significant diagnostic challenges due to its nonspecific clinical appearance and frequent misidentification as benign, inflammatory skin conditions. Traditional diagnostic methods such as biopsy are invasive and uncomfortable, often required repeatedly due to high recurrence rates. Dermoscopy and non-invasive imaging techniques have been used but provide limited diagnostic accuracy due to their constraints in depth penetration and resolution. Recent advancements in imaging technologies, such as line-field confocal optical coherence tomography (LC-OCT), show promise in enhancing diagnostic precision while minimizing invasive procedures. LC-OCT merges high-resolution imaging with deep penetration capabilities, capturing detailed horizontal and vertical skin images akin to histopathology. This study evaluated the diagnostic performance of LC-OCT in detecting EMPD and its recurrence in 17 clinically suspicious anogenital regions, belonging to six patients. Data were collected prospectively at the patient’s bedside by an LC-OCT expert with poor training for EMPD, and, then, reviewed retrospectively by an independent LC-OCT expert with adequate training for EMPD and no concerns about time. The prospective examination yielded 64.7% accuracy (11 true results out of 17 total cases), 71.4% sensitivity (10 true positives out of 14 actual positives), and 33.3% specificity (1 true negative out of 3 actual negatives). The retrospective analysis achieved 94.1% accuracy (16 true results out of 17 total cases), 100% sensitivity (14 true positives out of 14 actual positives), and 66.7% specificity (2 true positives out of 3 actual positives), with the only false positive case being a difficult-to-diagnose concomitant presentation of a lichen sclerosus et atrophicus. Despite the need for specialized training, our results suggest that LC-OCT represents a valuable tool for accurately identifying EMPD and improving its management by reducing unnecessary biopsies. Further studies are needed to standardize its clinical application.

Background: Line-field confocal optical coherence tomography (LC-OCT) is an innovative non-invasive imaging technique recently introduced in dermatology. Its use has been established particularly in the diagnosis of skin cancer. The LC-OCT description of basal cell carcinoma (BCC) criteria has served for the conduction of retrospective evaluations on its diagnostic performance, which demonstrated its utility in diagnosis and subtyping BCC. However, prospective studies performed at patients’ bedside are currently lacking in this field. Objective: The objective of the study was to provide parameters of LC-OCT diagnostic performance for BCC derived from a prospective study performed at patients’ bedside. Methods: Lesions clinically equivocal for BCC were prospectively included. Dermoscopic and LC-OCT diagnoses were obtained at the patients’ bedside by an expert observer prior to surgical excision or biopsy. Results: A total of 214 lesions (163 BCCs and 51 BCC imitators) belonging to 119 patients were included. For the differentiation of BCC from BCC-imitators, LC-OCT had the same sensitivity as dermoscopy (98%) but a better specificity (90% vs. 37%). For the discrimination of superficial BCC from other BCC subtypes, LC-OCT had an increased sensitivity and specificity compared to dermoscopy (72% vs. 62%, and 97% vs. 84%, respectively). Conclusions: This prospective study showed that the diagnostic performance for BCC diagnosis and subtyping can be remarkedly increased by LC-OCT compared to dermoscopic examination alone. Our data encourages the inclusion of LC-OCT in the diagnostic process and management of equivocal lesions for BCC.

Basal cell carcinoma (BCC) is the most common skin cancer. Off-the-shelf multimodal large language models are widely accessible, yet their performance for BCC remains unclear. The aim of this study was to assess BCC detection (BCC vs non-BCC) and BCC subtype classification from clinical and dermoscopic images using 3 web-based large language models (ChatGPT-5, Gemini 2.5 Flash, Claude Sonnet 4). We evaluated 772 images: 402 from 290 histopathology-confirmed BCCs (290 clinical, 112 dermoscopic) and 370 from an independent BCC-mimicker cohort (250 clinical, 120 dermoscopic). Standardized prompts were used. Primary outcome was BCC detection accuracy; secondary outcomes were subtype-classification accuracy and performance by lesion features. For clinical images, ChatGPT-5 achieved the highest detection accuracy (75%), followed by Claude (64.3%) and Gemini (50.7%). For dermoscopy, Claude performed best (69.8%), compared with ChatGPT-5 (55.2%) and Gemini (50.9%). Accuracy was lower in crusted and flat lesions and higher in exophytic lesions; pigmentation effects were model dependent. Subtype-classification accuracy was modest across models. Images were primarily from European centers with limited skin-type diversity; several subgroups were small. Current web-based large language models are not clinically suitable for BCC detection or subtyping. Dermatology-specific training, transparent reporting, and rigorous prospective validation are required before any clinical use.

Background/Objectives: Basal cell carcinoma (BCC) is the most frequent skin cancer in Caucasian populations. While dermoscopy supports diagnosis, accurate subtype classification requires histopathology. Line-field confocal optical coherence tomography (LC-OCT) offers high resolution, adequate penetration, and three-dimensional imaging, bridging the gap between dermoscopy and histopathology. This study assessed the concordance between LC-OCT and histopathology for BCC criteria and subtypes. Methods: We retrospectively analyzed 127 histopathologically confirmed BCCs from the Departments of Dermatology and Pathology, Hôpital Erasme, Brussels. LC-OCT images and corresponding histopathological slides were evaluated. Objective analysis used a predefined checklist of LC-OCT criteria compared with histopathology. Subjective analysis consisted of independent side-by-side assessments of global resemblance by three observers with varying expertise. Concordance rates and κ statistics were calculated. Results: The objective analysis showed the highest concordance (≥80%) for lobules, blood vessels, bright cells, lobule location, and dermal-epidermal junction disruption. Intermediate concordance (50–80%) was found for hemispheric morphology, outer bright rim, stromal stretching, and parakeratosis. Inner dark rim and palisading showed low concordance (<50%). Subjective evaluations demonstrated strong resemblance between LC-OCT and histopathology (overall concordance 81.1%), ranging from 86.6% to 98.4% across observers. Interobserver agreement was slight overall (κ = 0.10, p = 0.02), with one moderate pairwise κ (0.41). Conclusions: LC-OCT demonstrates good concordance with histopathology for key diagnostic and subtype-discriminating BCC features. Despite variability in subtle criteria, the findings support LC-OCT as a clinically relevant tool for non-invasive diagnosis and management of BCC.