The gamification of data analysis in cancer increases citizen contribution and reduces research time
Individuals diagnosed with muscle-invasive bladder cancer face a difficult treatment decision – intensive radiotherapy or complete surgical removal of their bladder. Each option has benefits and draw backs, and there are limited data available to patients and physicians to help predict which treatment might provide the best outcome.
Dr. Anne Kiltie, Associate Professor of Radiation Oncology at Oxford University and Clinical Group Leader at Cancer Research UK (CRUK) is trying to improve that decision making process. She is investigating whether proteins involved in DNA damage signaling and repair might serve as biological indicators, or ‘biomarkers’, predictive of a patient’s response to treatment.
In 2010 her team published data showing that higher levels of the DNA repair protein MRE11 correlated with better survival rates in bladder cancer patients who had undergone radiotherapy. This was a critical finding suggesting that MRE11 could be a treatment predictive biomarker. Unfortunately the finding relied on time consuming pathology analysis. Bladder tumor samples are sliced, labeled for each protein of interest and photographed. While the staining and imaging of these slices can be automated, each image must be manually quantified for the level of protein present. Computer algorithms are not yet as reliable as the human eye. So to study just one protein, Kiltie’s team must individually score hundreds of images, a major time sink.
Of course if there was a way to get more people involved, the research could proceed much faster.
And this is the basis of Reverse the Odds, a mobile app game in which citizens help with real data analysis. The game was the result of CRUK’s partnership with British television broadcaster Channel 4 and developed in collaboration with Chunk, Maverick Television, and Zooniverse.
“Developing this game was a new experience for both our researchers and the game developers,” says Rupesh Robinson-Vyas, Science Engagement and Operations Officer at CRUK. The development team had some key concerns. Could a game that dealt with a serious topic like cancer still be approachable and fun? And could the data analysis be gamified without compromising the quality of analysis?
Yes and yes.
The final product is a puzzle game in which players help whimsical creatures, ‘Odds’ reclaim their world. Rather than hide the science, the game developers put the science front and center. To level up, players visit the ‘lab’ where they learn to quantify protein expression in real bladder tumor samples. All the samples are from patients who have already undergone treatment. Players are not diagnosing patients.
Each image is reviewed by multiple players, with each player’s response being compared to other players’ responses. In this way discrepancies in analysis are weeded out. After the images are quantified, the data are sent back to Kiltie’s team who compare the level of protein expression to the patient’s known outcome.
Not only has the game cut down on data processing time, Kiltie’s team can concurrently evaluate the expression level of multiple proteins. This means the relationships between proteins, how they might work together to affect treatment outcome can also be studied.
As for the seriousness of the subject matter, rather than a deterrent, the science and the potential to make a significant contribution to cancer research is a strong motivator for game play.
“It was an unexpected outcome. The game allowed us to engage a younger age group, individuals who might not be able to contribute financially, could instead contribute their time,” says Robinson-Vyas. The game has been a huge success, translated into five languages, with an international reach of 150 countries.
CRUK is testing new methods of training and presenting pathology data to users. “Rather than asking our users just one question, the new interactive will teach us how to ask questions and identify what kind of data our users can help us collect,” says Robinson-Vyas. Results with citizen scientists matching pathologist’s accuracy in data analysis are encouraging; CRUK is writing a paper on these methods and plans to test further iterations.
As for Reverse the Odds, Kiltie’s team hopes to complete data collection this spring. Her team intends to make the results as well as background on game development openly available.
Visit the project page to learn more about Reverse the Odds and join the effort help Kiltie’s team analyze the final 500,000 slides by March.
Learn about CRUK’s other project Play to Cure: Genes in Space