Artificial intelligence was as good or better than doctors at detecting small lung cancers on CT scans in a study by researchers from Google and several medical centers, according to The New York Times. In a process known as deep learning, researchers set to train computers to recognize patterns linked to specific conditions such as pneumonia by feeding them large amounts of data and used that information for CT scans in lung cancer. After testing 6716 cases, and pitted against 6 expert radiologists, the system was 94% accurate, according to the article.
 
Investigators at the Université de Genève, Switzerland, have devised a cell co-culture platform that reproduces a patient’s tumor structure in 3D, to be used to test several drugs or combinations at different stages in the tumor’s development, according to a press release through the American Association for the Advancement of Science. Scientists now need only 5 days to identify which treatment will be the most effective for a particular case. The new platform opens the door to personalized medicine, according to the release.
 
A new white paper from the Friends of Cancer Research and the Parker Institute for Cancer Immunotherapy calls on the FDA to ease clinical and manufacturing requirements in order to speed early development of anti-cancer cell therapies, according to the Regulatory Affairs Professionals Society. The paper calls for FDA to revise its 2006 guidance on exploratory investigational new drug studies to provide recommendations for exploratory studies of cell therapies. It are these regulations that the groups believe are hindering new developments for treatment therapies and are leading to higher costs.