About this session
The Open Source Imaging Consortium (OSIC) is a global collaboration platform that leverages imaging technologies and clinical data to develop machine learning biomarkers for diagnosing, staging, and assessing therapy responses in rare lung diseases. OSIC, a not-for-profit entity, is funded by competitors in the pharmaceutical and machine learning sectors and is supported by academia and patient advocacy groups. In this session, we explore the long-term vision for imaging in diagnosing fibrosing lung diseases, including the use of lung cancer screening for risk assessment of progressive ILAs. We review preliminary machine learning findings from lung cancer screening data related to ILAs and ILDs and learn about the anonymized, diverse, multi-ethnic, and multi-center data within the OSIC Cloud Data Repository. Additionally, we discuss the importance of normalizing and curating imaging data to create effective algorithms and review recent machine learning results presented using OSIC Cloud data. Lastly, Dr. Bruno Hochhegger presents findings on classifying progressive pulmonary fibrosis using algorithms derived from imaging and clinical data.