Published Paper Details:

Oral Lichen Planus (OLP) is a chronic inflammatory disorder affecting the oral mucosa, with diverse clinical presentations and a risk of malignant transformation. Early and precise assessment of lesions, along with prediction of immunotherapy outcomes, is critical for personalized patient care. Manual evaluation is time-intensive and prone to variability among clinicians. In this study, we introduce a hybrid deep learning framework that combines Convolutional Neural Networks (CNNs) for detailed spatial feature extraction with Transformer modules for capturing long-range contextual information. A multimodal fusion strategy further integrates lesion imaging features with patient-specific clinical data to predict individualized therapy responses. Experiments on a curated oral lesion dataset demonstrate that our approach achieves superior segmentation accuracy and predictive performance compared to existing baseline models. The framework offers a reproducible, interpretable, and automated tool to support precision oral healthcare and data-driven treatment planning.