mHealth, powered by machine learning technology, has made large inroads to healthcare, and has numerous useful applications in the area of diagnostics, especially for diseases with alarmingly low survival rates. In recent years, a host of smartphone apps have been made available to help those that require early diagnosis of diseases by regularly monitoring symptoms, thereby facilitating timely treatments. Pancreatic cancer, marred by one of the lowest survival rates, is hard to diagnose due to the late surfacing of its symptoms, especially jaundice.
A team of researchers at the University of Washington are developing a smartphone app known as BiliScreen that will reliably screen pancreatic disorders by monitoring the symptoms of jaundice. These apps will work while taking selfies and will regularly check for increased levels of bilirubin in patients by detecting the discoloration of sclera in the eye.
Detecting Bilirubin Buildup in Blood Before It Reaches Alarming Level
Jaundice, one of the most common symptoms of pancreatic cancers, is characterized by buildup of bilirubin that evades detection till it reaches an elevated level.
BiliScreen can be useful in diagnosing milder forms of jaundice and other pancreatic disorders. The app uses phone’s camera, machine learning technologies, and computer vision algorithms to detect the alarming buildup of bilirubin in the blood, much before they are visible to the naked eye. The details of the app functioning will be tentatively described in the upcoming Ubicomp 2017 in September.
Early Diagnosis of Pancreatic Cancer May Reduce Mortality
The team of researchers claims that the app is non-invasive and facilitates the early diagnosis of pancreatic cancer triggering early treatments, even before patients may need surgery. This may help in bringing down the mortality rate in patients. Furthermore, BiliScreen can be used for the constant monitoring of bilirubin levels in infants and toddlers, for which the blood test is the only way.
The investigators in the clinical study included 70 people and established that the app when used with a 3-D printed box can correctly detect the cases of concern in 89.7% of times. Though detection of bilirubin may still call for confirmation by conventional blood tests, the results are markedly better and may prove promising for developing potential mHealth technologies.