Learning the Language of Histopathology Images reveals Prognostic Subgroups in Invasive Lung Adenocarcinoma Patients
The PathRosetta framework treats tissue as a structured biological language, using cells as 'words' and architecture as 'sentences' to predict five-year recurrence in lung adenocarcinoma. By employing multi-scale feature extraction and cell-type specific models, the AI successfully identified prognostic subgroups with high accuracy, outperforming clinical tools.