Microsoft researchers have begun conducting around the topic of lung cancer risks. Their secret weapon is search engines.

Smoking is the leader when it comes to lung cancer, which is the most fatal cancer in the world, causes. BUT, 20 percent of people that are diagnosed with cancer are non-smokers. So, how can this be explained? There are other factors such as genetic, geographic, and demographic factors that are determinants of lung cancer.

Researchers from Microsoft labs decided to conduct a project that explores the likelihood of learning about lung cancer risks through anonymous web search logs. The reason behind doing this project is to be able to present a forewarning to people who are candidates for the disease.

The machine has a unique learning method which picks up on patterns found in search engines. The important thing about this project is they don’t only look at the text searched; they look at the information behind the text. Locational risks seem to be their main focus. The zip code behind the search ties together data taken from maps from the US Geological Survey to establish how much radon gas is within a given environment, which is a very common lung cancer risk factor. Also, data that comes from the US Census provides important information about how old homes are within an area or region.

This information is relevant because knowing the age of a home can reveal information about ventilation operation intensity which ultimately leads to radon levels. This is because the older the home the more poor the ventilation level which equals to radon being more easily trapped within a home.

Zip codes also provide a view of the socioeconomic status of a region, which give off a better vicinity of the cancer risk information. Some researchers have come to a conclusion that people living below poverty line are more susceptible of smoking, which puts them more at risk for lung cancer.

The machine uses an algorithm to figure out the gender and age of the searcher from the patterns given off by the web search. When a search is made from the same mobile device but from different locations hundreds or thousands of miles away from each other, this may indicate frequent air travel, which may lead to being a cause to lung cancer. Also, the model looks back in time to see if any searches were made that have a relationship to lung cancer, for example any symptoms.

In conclusion, the purpose of this study is the collect research that may or may not lead to future clinical studies regarding lung cancer. These researchers are looking to discover new risk factors that previously have not been taken into consideration or researched. Their findings are merely associations and although very useful, they cannot yet be determined as actual cause for lung cancer.