We’ve shared before about how dogs can be trained to sniff out certain cancers, but now, researchers have discovered a way to use AI and machine learning to detect cancers in much the same way.
Researchers from the University of Pennsylvania are using AI to decipher the mixture of volatile organic compounds (VOCs) coming off cells in blood plasma samples of cancer patients. The ”e-nose” system uses nanosensors to detect VOCs and in trials, the device was able to detect early and late-stage ovarian cancer with 95 percent accuracy and pancreatic cancer with 90 percent accuracy.
What’s most impressive is that the machine was trained to differentiate between healthy and cancerous samples in just 20 minutes. The researchers hope that if scaled up, the technology could aid in rapid cancer detection even in remote areas. The efficacy of the system with ovarian and pancreatic cancer is significant as these are some of the most difficult cancers to detect.
Source study: ASCO, Penn