MLPI Research
Clinical research
In this Christian Doppler Laboratory, radiologists are conducting research together with AI specialists to further improve the informative value of CT images for the assessment of lung carcinomas. Currently, radiologists assess tumors and metastases primarily in terms of their size, density and location and their relationship to surrounding structures. By assisting with AI, tumors can be analyzed in a much more automated manner, thereby significantly improving the estimation of prognosis. With this project, we also expect to be able to support therapy decisions and ultimately improve the prognosis.
AI Research
We are developing machine learning methods for the assessment and prediction of treatment response in lung cancer. We are specifically focussing on methodology that bridges imaging phenotype data, and molecular characteristics of the tumor, and the biological processes of disease progression. To enable wide and sustainable applicability of the methods, we will investigate approaches to cope with heterogeneity of data, and the continual change of acquisition technology and treatment options.
Legal Research
In addition to ensuring respect for individual patients' privacy rights the general legal compliance of project activities, we are interested in broader legal questions related to data sharing multi-party medical research projects. Specifically, we are interested in the impact of existing and upcoming legislation on fostering and/or hindering data-rich, multi-party medical research projects such as CD-Laboratory for Machine Learning Driven Precision Imaging. We also hope to develop best practices on how to navigate the legal landscape for partners involved in such projects.
Publications
- Watzenboeck, Martin L., et al. "Contrast Agent Dynamics Determine Radiomics Profiles in Oncologic Imaging." Cancers 16.8 (2024): 1519.
- Langs, G. "Artificial intelligence in medical imaging is a tool for clinical routine and scientific discovery." Seminars in Arthritis and Rheumatism. Vol. 64. WB Saunders, 2024.
- Langs, G. 2023 Innovations in Deep Learning to Predict Individual Risk and Treatment Outcome. Radiology, 307:5
- Röhrich, S., Hofmanninger, J., Prayer, F., Müller, H., Prosch, H. and Langs, G., 2020. Prospects and Challenges of Radiomics by Using Nononcologic Routine Chest CT. Radiology: Cardiothoracic Imaging, 2(4), p.e190190.
- Kifjak, D., Hochmair, M., Sobotka, D., Haug, A., Ambros, R., Prayer, F., Heidinger, D., Roehrich, S., Milos, R.-I., Wadsak, W., Fuereder, T., Krenbek, D., Fazekas, A., Meilinger, M., Mayerhoefer, M., Langs, G., Herold, C., Prosch, H., Beer, L. Metabolic tumor volume and sites of organ involvement predict outcome in NSCLC immune-checkpoint inhibitor therapy. European Journal of Radiology, 170 (2024), 111198