New or changing skin lesion

In a diagnostic accuracy study of 608 skin lesions measured with a 3D fringe projection scanner across two European hospitals, significant morphological differences were found between melanomas (n=60) and benign nevi (n=81) with p<0.001 for multiple surface parameters including area-to-perimeter ratio and volume-to-perimeter ratio. A supervised machine learning classifier using these 3D morphological features achieved 80.0% sensitivity and 76.7% specificity for distinguishing melanomas from nevi. The study population included 194 analyzable lesions: 42% benign nevi, 31% melanomas, 9% basal cell carcinomas, 9% non-nevi benign lesions, 6% seborrheic keratosis, and 3% squamous cell carcinomas.

View Source →

A community pharmacy mole scanning service in Norway performed 25,836 scans on 15,777 individuals over 3.5 years (2010-2014). Of those scanned, 1% were identified as having melanoma, and 15.4% had another skin condition requiring attention. In 2014, the service identified 4.1% of all melanoma cases registered in the Norwegian Cancer Registry. Only 83.6% of scans were normal, meaning 16.4% of individuals who sought screening had a condition warranting follow-up. Patient satisfaction was high: 88% would use the service again, 99% found the pharmacy a suitable venue, and 95% would recommend it to others.

View Source →

In this matched cohort of 706,037 men followed for a mean of 4.9 years (3.44 million person-years total), 1,315 malignant melanoma cases were diagnosed. Sun exposure indicators were significantly elevated among those who developed melanoma-related conditions: solar keratosis (HR = 1.21, 95% CI 1.17-1.25, p < 0.001) and basal cell carcinoma (HR = 1.15, 95% CI 1.11-1.19, p < 0.001) were both associated with the same lifestyle profile. The strong link between solar keratosis history and subsequent skin cancer diagnoses (OR = 1.28, 95% CI 1.23-1.34) underscores that individuals with sun-damaged skin face elevated melanoma risk and should monitor for suspicious skin changes.

View Source →

An automated image analysis system tested on the PH2 Dermoscopy database of 200 dermoscopy images from Pedro Hispano Hospital distinguished between normal, atypical, and melanoma lesions using feature extraction including asymmetry, border irregularity, color variation, and dermoscopic structures. Classification accuracy reached 96.3% for normal lesions, 95.7% for atypical lesions, and 97.5% for melanoma. The high accuracy in distinguishing atypical from normal and melanoma cases confirms that visible lesion characteristics reliably differentiate benign from potentially malignant growths. Clinical diagnosis remains prone to misdiagnosis due to subjective assessment, reinforcing the importance of prompt professional evaluation when atypical features are observed.

View Source →