A new application is being developed by researchers at the University of Washington that could use software to help users with the early detection of one of the most dangerous types of cancer. Bilirubin build-up in the blood stream is a key early indicator for Pancreatic cancer – which has a five-year survival rate of only 9 percent. Pancreatic cancer’s high fatality rate is primarily the result of how difficult it is to catch. The application, which is currently called BiliScreen, works using an algorithm, a smartphone camera, and machine learning to “detect increased bilirubin levels” by examining coloration changes in the white of a human eye. According to researchers, the application is able to catch changes in coloration long before the human eye can detect them and, better still, users only need to take a selfie to use BiliScreen.
What makes BiliScreen so exciting to the people involved in the project, is how accurate it has been shown to be so far. Although the app won’t officially be presented until September 13 at Ubicomp 2017, the Association for Computing Machinery’s International Joint Conference on Pervasive and Ubiquitous Computing, initial clinical studies have correctly identified cases of concern 89.7 percent of the time. Unfortunately, as of this writing, accuracy is still dependent on a 3-D printed box – which controls light exposure – but the results are promising nonetheless since the findings are backed up by subsequent blood work confirmation. Another study, conducted earlier with BiliScreen’s predecessor BiliCam, showed comparatively accurate estimates of bilirubin levels in 530 infants. Moreover, changes in eye coloration are more consistent than jaundice-related skin changes across all age groups and ethnicities. Discoloration is also not solely linked to pancreatic cancer, so any advancement in measuring bilirubin through examination of the eye is going to be hugely beneficial to fields of medicine.
At very least, if BiliScreen is successful, users will no longer be burdened by the costs associated with visiting a doctor to have blood drawn and the extensive tests tied to that for many patients. Instead, anybody concerned with their individual risk-level for pancreatic cancer will be able to use the application to decide if more tests are needed. However, it will also help those who are already experiencing health problems related to bilirubin concentration in the blood stream to track the progress of their condition with high accuracy and with more frequency. Aside from the implied cost savings associated with that, doctors and medical researchers could eventually find more efficient ways to treat disorders on a more personal level using the data gained from those measurements.