Launching the OSIC Data Repository has been a collaboration in its truest sense. By bringing together the world’s “best in class,” we’ve been able to build something truly remarkable — the largest and most diverse global database for are fibrotic lung diseases. All open source. All for the benefit of patients.
We spoke with OSIC Board Member Eric S. White, MD, MS, senior clinical program lead, ILD, at Boehinger Ingelheim, and asked him what accomplishment within OSIC he is most proud of so far.
We are very excited to announce the winners of our $55,000 OSIC Pulmonary Fibrosis Progression Challenge, the first-ever computational challenge for interstitial lung diseases (ILDs). Our AI-focused competition was administered by Kaggle, the world’s largest data science community platform, and we asked participants from around the globe to use machine learning to predict lung function decline in people living with pulmonary fibrosis.
After two years of work and through a true team effort, we are very excited to announce the launch of the OSIC Pulmonary Fibrosis Progression Challenge, hosted by our partner, Kaggle, the world’s largest data science challenge platform. This $55,000 challenge will be the first of its kind for Pulmonary Fibrosis and will run until October 6, 2020.
Knowing Pulmonary Fibrosis is primarily diagnosed by images of the lungs and because we have seen advances in pattern recognition/deep learning for other disease states, OSIC work with our many collaborators and under the guidance of full GDPR/HIPAA compliance to create a database of HRCT images to be used for the challenge and for future OSIC efforts.
When you talk to Pulmonary Fibrosis experts, the one thing they all agree on is that the disease is heterogeneous, meaning no patient experience is the same in disease progression. Currently, there is no way to help patients understand the path of their disease or what is to come. It’s beyond difficult for the patients, caregivers AND clinicians. The Kaggle data scientists who will compete have a true opportunity to make a meaningful difference by creating a deep-learning algorithm based on data and image analysis. It’s been done successfully in other disease states, and we believe it can and should be done for Fibrosing Lung Diseases.
One interesting thing about this challenge is whether the perfect algorithm is found or not, the challenge itself will be a valuable source of learning. As author Frans Johansson said in his book The Medici Effect, “When you step into an intersection of fields, disciplines or cultures, you can combine existing concepts into a large number of extraordinary ideas.” We believe that broadening the awareness of Interstitial Lung Diseases to data scientists from a variety of backgrounds may indeed help us progress in ways we hadn’t yet considered.
The other part of this challenge that is powerful is the increased awareness that it will bring this often-ignored disease state. Remarkably, as many people die each year of Pulmonary Fibrosis as die of breast cancer, but most people have never heard of it. This challenge will raise the level of dialog about the disease and hopefully, in time, bring new meaningful clinical innovations for those fighting the disease and for their families. Our number one motivation has always been, and will always be, to make rapid advances in the fight against IPF and other ILDs.
If you have the time or inclination to help us, please share the challenge information on your social pages with the following hashtags: #OSICchallenge #CureIPF
Thanks so much!
