Detecting pancreatic cancer: Web searches could hold the key

Scientists in Microsoft and University of Columbia have found that studying the past searches of pancreatic cancer patients could go a long way in early detection.

Of the all the cancers, pancreatic cancer is the deadliest, we are told. Now, scientists have found that web searches could help detect pancreatic cancer and aid future diagnosis.

Scientists at Microsoft have teamed up with researchers from Columbia University in the US to study how web searches can be used to detect future diagnosis of pancreatic cancer.

Researchers were able to identify between 5 and 15 per cent of cases with low false-positive rates.

Read: Study reveals new strategy for treating pancreatic cancer

Scientists collected data from the search engine Bing and found that web searches can be used under certain circumstances to detect a future diagnosis of pancreatic cancer. They identified recent search queries indicative of pancreatic cancer, such as the term, “I was told I have pancreatic cancer, what to expect.”

Past searches like ‘yellowing skin’, ‘dark urine’ and ‘abdominal swelling’ of a pancreatic cancer patients could become valuable leads to early detection, say scientists.

Researchers then found and analysed patterns to identify cases. “Finding low cost, low risk, high coverage health surveillance systems is an important challenge,” said Robert Grossman from University of Chicago who was not involved in the study.

Read: Slow poison | Alcohol harms not just your liver, but pancreas too

According to researchers, there is a lot of potential in analysing health data because there is a wider range of access points now. Aside from making web searches, individuals can take advantage of wearables that record health data specific to them, researchers said.

Gaining access to these data will give not just the current researchers but others as well a trove of information that can be used to improve health outcomes for everyone, they said. The findings were published in the Journal of Oncology Practice.