TIDaC is an R package that allows to build automatically two types of labelled datasets containing digital images, which are retrieved using the REST APIs provided by two sources: TCGA (The Cancer Genome Atlas) and TCIA (The Cancer Imaging Archive).
The tar.gz R package can be downloaded here.
- httr
- readr
- jsonlite
The figure below represents the two main components of the TIDaC architecture. For each component the flow diagram is provided, i.e. the sequence of functions call the allows the creation of the datasets. So far TIDaC allows to create two types of labelled dataset:
- Using the component that interfaces with TCGA it is possible to create a dataset containing histopathological images, labelled based on the mutation occurred;
- Using the component that interfaces with TCIA it is possible to create a dataset containing medical images (obtained in a non-invasive way), labelled using several attributes specified through the "groupby" parameter.
- Be aware that in order to use the TCIA Component is required to own a TCIA Api Key. The instruction on how to obtain access to the TCIA REST API can be found here.
Here I report some papers in which these types of images are used in the oncology field:
- Gillies, R. J., Kinahan, P. E., & Hricak, H. (2016). Radiomics: Images Are More than Pictures, They Are Data. Radiology, 278(2), 563–577. doi:10.1148/radiol.2015151169;
- Coudray, N., Ocampo, P. S., Sakellaropoulos, T., Narula, N., Snuderl, M., Fenyö, D., Moreira, A. L., Razavian, N., Tsirigos, A. (2018). Classification and mutation prediction from non–small cell lung cancer histopathology images using deep learning. Nature Medicine. doi:10.1038/s41591-018-0177-5;
- Bychkov, D., Linder, N., Turkki, R., Nordling, S., Kovanen, P. E., Verrill, C., Walliander, M., Lundin, M., Haglund, C., Lundin, J. (2018). Deep learning based tissue analysis predicts outcome in colorectal cancer. Scientific Reports, 8(1). doi:10.1038/s41598-018-21758-3;