Multimodal Head and Neck cancer dataset (HANCOCK)

Head and neck cancer is common and often associated with poor prognosis. Personalized treatment, leveraging data from multiple modalities, is a promising approach to improve outcomes. The multimodal Head And Neck Cancer dataset (HANCOCK) includes monocentric, real-world data from 763 patients, offering a valuable resource for AI-driven solutions. It contains demographic, pathological, and blood data, along with surgery reports and histologic images. HE-stained whole slide images are available for primary tumor sites and adjacent lymph nodes. Tissue microarrays, stained for CD163, CD3, CD56, CD68, CD8, HE, MHC-I, and PD-L1, focus on tumor centers and invasion fronts. Surgery reports are available in German and English. All laboratory data, pathology reports, and clinical data have been cleaned and standardized.