Post

New research · Dermatology
Journal der Deutschen Dermatologischen Gesellschaft = Journal of the German Society of Dermatology : JDDG · 1w
AI / informaticsJournal der Deutschen Dermatologischen Gesellschaft = Journal of the German Society of Dermatology : JDDG · 2026

AI-assisted diagnosis of nail unit melanoma and melanonychia using a clinical deep learning model.

Yusung Chu, Sejung Yang, Jin-Woong Jung … Byungho Oh
Read paper
DermatologyAI / informatics

Deep learning model separated nail unit melanoma from benign melanonychia on clinical photographs

AI-assisted diagnosis of nail unit melanoma and melanonychia using a clinical deep learning model.

Yusung Chu … Byungho Oh
Journal der Deutschen Dermatologischen Gesellschaft = Journal of the German Society of Dermatology : JDDG · 2026
Background

Nail unit melanoma (nail unit melanoma) is a rare but potentially fatal malignancy often misdiagnosed as melanonychia.

Purpose

This study aimed to develop and validate an artificial intelligence model to distinguish nail unit melanoma from benign melanonychia using clinical images and to assess its diagnostic utility through human comparison and external validation.

Methods

Clinical images from 172 patients with melanonychia and 122 patients with nail unit melanoma were retrospectively collected.

n = 172 patients
Results

how well the AI told nail melanoma (a nail cancer) from harmless nail pigment - higher is better

70%
Before
80.8%
After
More results

ResNet-50 achieved the highest sensitivity (87.2 %).

Convolutional neural networks assistance improved the diagnostic accuracy of all human raters (accuracy increased from 70.0 % to 80.8 %), with dermatology residents showing the largest gain.

“
Conclusion · 1 of 2

The proposed convolutional neural networks-based models demonstrated robust performance in differentiating nail unit melanoma from melanonychia and improved diagnostic accuracy and agreement among human evaluators.

Conclusion · 2 of 2

While not intended to replace clinical judgment, this approach shows promise as a supportive screening tool in clinical settings, warranting further validation in larger, multi-institutional, and multi-ethnic cohorts.

Read paper
0 comments

No comments yet. Be the first.

Related papers

LatestFoundational
Cohort Study
κ = 0.211
confocal and dermoscopy diagnoses agreed only fair, not reliably interchangeable
Cohort Study
48%
nearly half never started treatment their surgeon recommended
Cohort Study
HR 1.06
6% higher rate of non-melanoma skin cancer with GLP-1 drugs versus SGLT2 inhibitors; effect faded after bias correction.
Study
0.65
converts standard dose planning to match this therapy - lower means a smaller adjusted dose
AI / Informatics
82.1%
patients who said they felt comfortable with an artificial intelligence note-taking tool during visits
Cohort Study
18.7 percentage points
more patients needed no further specialist care after photo-based skin checks than in-person clinics
Cohort Study
70.8%
clinical improvement was seen in far more terbinafine-treated nails than fluconazole-treated ones
Systematic Review
“Chronic skin disorders are consistently associated with psychological distress and reduced HRQoL. Causal inferences cannot be drawn given cross-sectional designs; findings support routine psychosocial assessment and family-centered care to improve outcomes.”