Inspiration
As a 4th year medical student in AI, I was inspired by the potential of machine learning to assist in clinical decision-making. Thoracic surgeries, such as lung resections, carry significant risks, and predicting post-surgical life expectancy could provide valuable insights for both doctors and patients. Our goal was to create a tool that helps clinicians make more informed decisions based on data-driven predictions.
What it does
Our project is a web-based application that predicts life expectancy after thoracic surgery using machine learning. By inputting key clinical parameters, the model estimates a patient's survival probability, offering a data-backed perspective on post-surgical outcomes. This tool aims to support healthcare professionals in risk assessment and patient counseling.
How we built it
We started by analyzing a dataset containing thoracic surgery patient records and their outcomes. Using Python and machine learning libraries like Scikit-Learn, we developed predictive models to estimate survival probabilities. We then built a web application using Flask to provide an intuitive interface for users to input patient data and receive predictions in real time.
Challenges we ran into
One of the biggest challenges was handling imbalanced data, as survival rates varied significantly. Ensuring model accuracy while preventing overfitting required careful feature selection and hyperparameter tuning. Additionally, integrating the machine learning model into a user-friendly web interface required optimizing performance and deployment strategies.
Accomplishments that we're proud of
We successfully built a working application that combines AI and healthcare in a meaningful way. Our model achieved a balanced accuracy that provides clinically relevant insights, and we were able to deploy it as a functional web tool. More importantly, we took a step toward bridging the gap between AI research and real-world medical applications.
What we learned
Throughout this project, we deepened our understanding of both machine learning techniques and their applications in medicine. We gained hands-on experience in handling medical datasets, addressing class imbalance, and deploying AI models in a web environment. Additionally, we learned the importance of making AI-driven tools interpretable and accessible for clinicians.
What's next for Life Expectancy Prediction After Thoracic Surgery
Moving forward, we aim to refine our model by incorporating more diverse datasets and additional clinical variables. We also plan to enhance the web application's user experience and explore potential integrations with electronic health record (EHR) systems. Ultimately, our goal is to make this tool more robust and widely applicable in thoracic surgery risk assessment.
Built With
- bootstrap
- cometml
- css3
- html5
- javascript
- lime
- matplotlib
- numpy
- pandas
- python
- scikit-learn
- scipy
Log in or sign up for Devpost to join the conversation.