AI-mediated immunotherapeutics in adenoid cystic carcinoma: Challenges and current perspectives

Adenoid Cystic Carcinoma (AdCC) is an uncommon, aggressive, and incurable head and neck cancer. Using cutting-edge machine learning (ML) and deep learning (DL) techniques, artificial intelligence (AI) has become a game-changing tool for the diagnosis, comprehension, and treatment of AdCC. In order to provide precision medicine, models like Support Vector Machines (SVM), Random Forest (RF), K-Nearest Neighbours (KNN), Logistic Regression (LR), Gradient Boosting Machines (GBM), Convolutional Neural Networks (CNNs), and Artificial Neural Networks (ANNs) combine imaging, histopathological, and genomic data. Few studies reveal that Convolutional Neural Networks (CNNs) have achieved higher accuracy in feature extraction and tumor classification, whereas Random Forest (RF) models demonstrated considerable specificity in detecting the disease associated with genetic mutations. AI also enables targeted immunotherapy by identifying molecular markers and optimising drug responses to individual patients. SVM and RF classify AdCC subtypes based on MYB::NFIB fusion, NOTCH, and WNT pathways. DL models analyse imaging and histology to assess immune infiltration and predict response to checkpoint inhibitors. GBM models group patients by PI3K/AKT and NF-κB pathway alterations for tailored immunotherapies. AI further enhances CAR-T cell therapy by predicting the presentation of neoantigens and engineering T-cell receptors. Real-time treatment monitoring is made possible by liquid biopsies that use AI-driven ctDNA and immune profiling. Notwithstanding advancements, issues with clinical integration, model interpretability, and data quality still exist. Future directions highlight federated learning models, explainable AI, and large-scale clinical validation as key to integrating AI into precision oncology for patients with AdCC.