What is there to learn from your internet search history? According to a new study, a lot more than simply reminding you of that random curiosity at that random moment.

Researchers from Microsoft have shown an ability to identify internet users with pancreatic cancer by analyzing large amounts of search engine entries, even before they received a diagnosis.

Published in the Journal of Oncology Practice, the study could be a step forward in the early diagnosis and treatment of cancer.

“We asked ourselves, ‘If we heard the whispers of people online, would it provide strong evidence or a clue that something’s going on?’” said Eric Horvitz, MD, PhD, a Microsoft researcher, in an interview with the New York Times.

The study examined online search activity that suggested a recent diagnosis of pancreatic adenocarcinoma. From there, researchers looked at previous searches to connect users’ questions with patterns of symptoms. Statistical classifiers were developed that could predict future appearances of search entries related to a diagnosis.

Horvitz, along with co-authors Ryen White, PhD, and John Paparrizos, MSc, claimed to identify between 5 and 15 percent of cases, while minimizing false-positives, which can lead to unnecessary anxiety and medical costs.

But the availability of such data may lead to improvements in cancer screenings and diagnoses. The researchers claimed their work could be used to improve early detection for pancreatic cancer, raising the five-year survival rate from 3 percent to 5 or 7 percent.