LungSight

LungSight Write Up

Problem Statement/Inspiration

While AI has made a lot of progress in medical imaging, it still feels like it’s missing the big picture. Most systems are built on limited or fragmented data, and they don’t generalize well across different types of patients or scans. On top of that, many tools don’t integrate smoothly into clinical workflows, which makes them harder to actually use in real settings. The idea behind LungSight came from noticing this gap. We wanted to build something that makes AI in medical imaging more interactive and practical. Using HOPPR’s Medical Imaging Foundation Models, our goal was to format AI output in a way that successfully explains the results of the x-ray scan.

Our Solution In order to combat these issues in modern medical imaging, we decided to build a project with the primary focus of analyzing medical images and providing information in a way that is easily readable and extremely useful. The image processing will detect abnormalities and highlight them for the medical professionals. It will also give the medical professionals the possibilities that are causing these issues.

How we built it (Tech Stack) Front-end - We utilized HTML, CSS, as well as some parts of JavaScript to build a beautiful and simple user interface that is easy and quick to use. Back-end - We used Node.js and Express for the backend. We used the .env file to connect the hoppr api and connected this backend to work with the front end.

Challenges we ran into

During the development on the problem we ran into multiple problems Integration with the Hopper API. We ran into issues when integrating the Hoppr API with our project. We kept encountering errors when we tried to use the features of the API. One of our biggest issues was that when the API was called we kept getting a 403 error where permission was denied. To fix this: We added mock data that was in the similar format Hoppr API would create, and it would simulate an actual formatting process with the mock data to properly format it for the user Connecting the backend with the frontend. We were struggling with getting the front end connected with the backend. The front end was struggling to communicate with the back end of the code. When making requests they would not go through.

What's next for LungSight:

Utilize Successful API connections: LungSight is currently using mock data that is meant to be in a similar format as the Hoppr API data when scanning chest x-ray images. Once a successful API connection is utilized, LungSight will have the opportunity to use other APIs that specialize in other x-rays. Expanded Diagnosis: Our project right now is limited with the medical images it can diagnose. It currently only works with X-Rays in the chest area. In the future we can expand it so it can work with all medical images in all different parts of the body. History: We can add a component in the project which would store the previous X-Rays for the patient. When the doctor looks at a new X-Ray he should have the ability to see the past medical history for his patient. Patient Portal: We can expand on the app where there is a separate platform for their results. They would have the ability to see their X-Rays, their results and what the doctor has to say.