Image-based Artificial Intelligence models in the diagnosis and classification of vascular anomalies of the soft tissue in the head and neck

Background: The International Society for the Study of Vascular Anomalies (ISSVA) provides a detailed classification system to facilitate accurate diagnosis and management of these conditions based on clinical criteria, imaging, and histopathological findings. This review aims to systematically present the state of the art in Artificial Intelligence (AI) applications for the diagnosis, classification, and treatment planning of vascular anomalies affecting the soft tissues of the head and neck region.

Methods: The PubMed research identified 86 articles. After the initial screening of titles and abstracts, 75 studies were excluded due to the following reasons: irrelevance to the review's scope, unsuitable patient population, or inappropriate study design. 11 pertinent papers were included for the full-text screening. At the end of the full-text evaluation, 3 studies were considered eligible to be revised.

Results: Concerning the clinical, radiological, and histological diagnosis of vascular malformations, the three models reviewed seem to be promising in the classification of the different subtypes and the lesion's segmentation. The applications of AI achieved in the fields of radiology and dermatology are very promising. Any study did not provide a prospective validation cohort to verify the diagnostic performance but also to assess the usefulness of the model in clinical practice.

Conclusion: We can also affirm that a hybrid model, combining dermatological images, MRI or US images, and histological images, can be developed as a reliable AI tool, useful for clinicians from diagnosis to treatment decision-making and to supervise interventional procedures.