My name is Bence Lukácsy and I created this website to help people assess their risk for cancer and encourage them to get a screening test. The website provides questionnaires that ask users a series of questions about their medical history, lifestyle, and other factors that can influence their risk for cancer. Based on their answers, the website provides a recommendation on whether or not they should consider getting a screening test. Currently, my focus is colorectal (colon) cancer.
Regarding colorectal cancer, the CDC reports that approximately “1 in 3 adults (23 million) between 50 and 75 years old [are] not getting tested as recommended.” Screening tests are among the most effective methods for early cancer detection, but several reasons discourage people from getting tested. For example, colonoscopies are both invasive and expensive, and even the non-invasive and cheaper alternatives such as stool tests can interfere with one’s diet and produce false-positive results. This website provides a completely free, accessible, and accurate method for analyzing colon cancer risk. My goal is to encourage people to get the appropriate screening test for them.
Upon completing a questionnaire, the answers will be inputted into a machine learning model that I have created. The model will predict a user’s cancer risk and the website will display the percentage and subsequent recommendation. Please remember that although the model is accurate on its training and testing data, it is NOT a substitute for cancer diagnosis. The model’s purpose is solely to encourage users to get an appropriate screening test, not claim that they have cancer.
An alternative to the questionnaires is cancerprediction.org’s API. If using the website is not ideal for a user, then the API Documentation will make clear how to create an HTTP POST request to receive the same results that the website would provide.
The code for both the website and deployed machine learning model is open source and available here.
Email: info@cancerprediction.org