We would like to establish an artificial intelligence (AI) working group in ITMIG. We envision a multidisciplinary ITMIG effort that is centered around imaging and image-based histologic slide analysis in thymic malignancies, primarily focusing on, but not exclusive to histopathology and radiology.
This analysis could include more conventional thresholding techniques/statistical modeling, as well as using the images for interobserver studies. A major objective would be the application of artificial intelligence (AI) methodology in both domains, aimed at improving classification, prognostication and prediction, as well as generating new biological insights through ‘explainable AI’ approaches in a more unsupervised setting.
This would require:
- obtaining approval for use, and following this, the collection, standardization, storage and preprocessing of data from previous studies;
- prospective case collection;
- generating ideas for studies using this material;
- setting up pipelines (ideally using existing infrastructure) for AI analysis of data;
- providing information and advice to researchers for setting up this type of studies.
We would welcome the participation of all interested ITMIG members in this effort, and ideally the group should be co-led by a histopathologist and a radiologist with some experience in the field of computational pathology/radiology. If you are interested in this group please let us know by emailing myself (Jan von der Thüsen) and Pam Bruce.
Jan von der Thüsen, Erasmus Medical Center, Rotterdam, The Netherlands