The following is a summary of “Advancement and independent validation of a deep learning-based tool for automated scoring of nail psoriasis severity using the modified nail psoriasis severity index,” published in the April 2025 issue of Frontiers in Medicine by Kemenes et al. 

Researchers conducted a retrospective study to enhance and validate a convolutional neural network (CNN)-based model for automated scoring of nail psoriasis severity using the modified Nail Psoriasis Severity Index (mNAPSI) with consistent accuracy across all severity levels and without reliance on standardized conditions.  

They included individuals with psoriasis (PsO), psoriatic arthritis (PsA), and non-psoriatic controls, including healthy individuals and those with rheumatoid arthritis, for model training, while validation used an independent individual with psoriasis. Nail photographs were pre-processed, segmented, and assigned mNAPSI scores by 5 expert readers. A CNN based on the Bidirectional Encoder Representation from Image Transformers (BEiT) architecture, pre-trained on ImageNet-22k, was fine-tuned for mNAPSI classification. Model performance was evaluated using the area under the receiver operating characteristic curve (AUROC) and other metrics, comparing results with human annotations. A reader study assessed inter-rater variability.  

The results showed that 460 individuals contributed 4,400 nail photographs for training, while the independent validation dataset included 118 individuals with 929 images. The CNN achieved high classification accuracy, with a mean (SD) area under AUROC of 86% ± 7% across mNAPSI classes in the training dataset. Performance remained strong in the validation dataset, with a mean AUROC of 80% ± 9%, despite imaging variability. Compared to human annotations, the CNN demonstrated a Pearson correlation of 0.94 at the individual level, maintaining consistency in the validation dataset.  

Investigators concluded that they had developed and validated a CNN capable of automated, objective scoring of nail psoriasis severity based on mNAPSI with high reliability and without image standardization. 

Source: frontiersin.org/journals/medicine/articles/10.3389/fmed.2025.1574413/full