Key Points

Question  How does artificial intelligence (AI) diagnostic performance compare to human dermatologists of varying experience for skin cancer detection in realistic clinical settings?

Findings  In this diagnostic study of 652 physicians and 3 AI models evaluating 1117 cases, expert dermatologists (>10 years of experience) achieved the highest accuracy (74.2%), considerably outperforming a modern unimodal foundation model (72.2%), which exceeded dermatologists with less than 1 year of experience (59.1%), while the first-generation convolutional neural network underperformed all readers (56.7%).

Meaning  Future practice should integrate human-AI collaboration, with AI supporting less experienced clinicians and providing expert triage assistance and help to minimize fatigue-related diagnostic errors.

To read further on this article please click here.

Malignant skin cancers impose an escalating and heterogeneous health burden worldwide.¹ Using the Global Burden of Disease (GBD) 2023 database,² we summarize these cancers’ epidemiology, subgroup patterns, and projections to 2050.

We analyzed GBD 2023 estimates (1990-2023) for malignant melanoma, cutaneous squamous cell carcinoma, and basal cell carcinoma. Outcomes included prevalence and disability-adjusted life-years (DALYs; years of life lost due to premature death plus years lived with disability). Subgroup analyses were conducted by sex, age group, and Sociodemographic Index (SDI; range, 0-1), defined as the geometric mean of indices of fertility in those younger than 25 years, education among those aged 15 years and older, and lag-distributed income per capita. Projections to 2050 used a bayesian age-period-cohort (BAPC) model, a bayesian hierarchical framework that jointly estimates age, period, and cohort effects and provides uncertainty intervals. For basal cell carcinoma, we excluded the 2005 to 2009 surveillance-artifact period and fit projections using 2010 to 2023 data. Additional methods are provided in the eMethods in Supplement 1. This study was deemed to be not human participant research by Shenzhen People’s Hospital; therefore, institutional review board approval and informed consent were not required.

To read this article in full please click here.

Yin Li, PhD; Robert A. Swerlick, MD

Giuseppe Argenziano, Giulia Briatico, Eugenia Veronica Di Brizzi, Camila Scharf, Gabriella Brancaccio, Elvira Moscarella, Maria Maddalena Nicoletti, Pasquale Verolino, Aimilios Lallas, Harald Kittler

ABSTRACT

Introduction: The recent WHO classification of melanocytic tumors introduces a refined molecular and histopathological framework suggesting distinct pathways and precursor lesions for all melanoma subtypes. While conceptually appealing, its clinical applicability is increasingly questioned.

Objectives: This review critically examines the transformation theory from benign nevi to melanoma, highlighting inconsistencies between the proposed models and real-life practice.

Methods: Through illustrative cases and key epidemiological evidence, we evaluated the validity of current models proposing intermediate lesions in melanoma development.

Results: We argue that most melanomas arise de novo and that the so-called intermediate lesions, such as dysplastic nevi and atypical Spitz tumors, may mimic melanoma but are not true biological precursors.

Conclusions: We propose a simplified, clinically oriented reclassification of melanocytic lesions based on morphologic ambiguity and actual behavior, aiming to guide therapeutic decisions and reduce di-agnostic overinterpretation.

To access the full article please click here.

Maya Mundada, BS; Xiaochen Zhong, BA; Alexandra So, BS; et al

Key Points

Question  What demographic, clinical, and pathological characteristics are associated with the time to recurrence of localized melanomas?

Findings  In this cohort study of 1092 individuals diagnosed with stage IA to IIC melanomas, tumor ulceration, thickness, location on the face or scalp or neck compared with the arms, neurotropism, lymphovascular invasion, and presence of mitoses were associated with time to melanoma recurrence in multivariable analysis.

Meaning  Results of this study suggest that factors in addition to melanoma ulceration and thickness provide an important guide for patient surveillance and counseling about potential recurrence.

To read more on this article please click here.

Alexandre Ladet, Sandra Lawton, Michael J Boffa

ABSTRACT

Introduction: In many Western countries, access to a dermatologist can be difficult, while the incidence of skin cancer has risen steadily over the past 50 years.

Objective: We reviewed the published literature to determine whether training primary care practitioners (PCPs) in dermoscopy through brief interventions based on diagnostic algorithms could improve patient care by improving their diagnostic accuracy of suspect lesions.

Methods: A scoping review of the literature was conducted, focusing on studies published in the period 2003–2023 that assessed the ability of low-experienced PCPs to triage suspicious dermatological lesions using dermoscopic diagnostic algorithms. Regarding outcomes, we focused on quantitative variables relevant to screening practice in general practice, including sensitivity, specificity, referrals to specialists, and unnecessary lesion excisions.

Results: Of the 926 studies initially identified, 13 were eventually selected: 10 cross-sectional observational studies and three randomized controlled trials. The studies were carried out in North America (N=6), Western Europe (N=4), and Australia (N=3). There was heterogeneity in the training interventions and the criteria used to assess diagnostic accuracy of PCPs after training; however, all studies showed an improvement in this parameter. The preferred algorithms for training PCPs were the 3-point checklist, the 7-point checklist, and the Triage Amalgamated Dermoscopy Algorithm.

Patricija Girštautė, Melita Virpšaitė, Iveta Gylienė, Jorinta Jokubaitė, Tadas Raudonis

Introduction

Pyogenic granuloma is a common benign vascular lesion that can clinically and dermoscopically imitate melanoma and other tumors [1]. In our case, pyogenic granuloma and nodular melanoma exhibited dermoscopic similarities. Some dermoscopic patterns, which include characteristics such as a yellow collarette, a reddish homogenous area, and the presence of white rail lines, can aid in diagnosing pyogenic granuloma [12]. Even though pyogenic granuloma is benign, histology is still necessary, especially in cases where vas-cular structures are present, as melanoma cannot be ruled out [1,3].

To read this article please click here.

Alexa S. Podolsky, BS ∙ Rachel Manci, MD ∙ Samantha S. Sattler, MD ∙ Daniel Lozeau, MD ∙ Jordan B. Slutsky, MD

Introduction

Noa Kremer, Isabella N. Dana, Emmanouil Chousakos, Larissa M. Pastore, Allan C. Halpern, Stephen W. Dusza, Jochen Weber; Ofer Reiter, Aimilios Lallas, Cristian Navarrete-Dechent, Ralph Braun, Harold S. Rabinovitz, Gustavo Carvalho, Rashek Kazi, Rozina B. Zeidan, Shirin Bajaj, Nicholas R. Kurtansky, Ashfaq A. Marghoob