Correlation of Invasion Depth and Histopathological Features With Dermoscopic Findings in Cutaneous Squamous Cell Carcinoma

This retrospective study includes patients who were diagnosed with cutaneous squamous cell carcinoma (cSCC) and underwent total excision between 2016 and 2025 at a tertiary dermatology center. Patient data were obtained from the hospital electronic medical record system, including demographic characteristics (age, sex), clinical features (lesion location, size, and morphology), and histopathological parameters such as tumor depth of invasion and histopathological subtype.

Dermoscopic images were acquired using a FotoFinder digital dermoscope as part of routine clinical documentation prior to excision. All dermoscopic images were independently evaluated by two dermatologists who were blinded to the histopathological results. Dermoscopic features, including vascular patterns and keratin-related structures, were recorded according to predefined criteria. Interobserver agreement was assessed, and discordant evaluations were resolved by consensus.

For objective image analysis, dermoscopic images were segmented and annotated using the LabelMe software. Regions of interest corresponding to specific dermoscopic structures were manually labeled, and pixel-based quantitative measurements were extracted using computer-assisted image analysis pipelines developed in the Visual Studio Code environment. The proportional area of each dermoscopic structure relative to the total lesion surface was calculated and recorded as numerical data.

The primary objective of the study is to investigate the relationship between dermoscopic features and histopathological parameters, particularly tumor invasion depth and histopathological subtype. Secondary objectives include assessing whether quantitative dermoscopic measurements can improve the prediction of invasive behavior in cSCC.

Statistical analyses will be performed to evaluate correlations between dermoscopic variables and histopathological outcomes. Appropriate parametric or non-parametric tests will be used depending on data distribution. Multivariate analyses may be conducted to identify independent dermoscopic predictors of deeper tumor invasion. All analyses will be performed using standard statistical software.