An AI diagnostic tool developed for doctors to classify gastrointestinal cancer tumours into MSS or MSI.
MSI (microsatellite instability) and MSS (microsatellite stability) biomarkers indicate the stability of the DNA in the tumour. Determining the stability status is extremely important as it dictates the type of treatment needed for the tumour.
The system is aimed towards doctors and other medical professionals to assist in diagnosing and classifying gastrointestinal cancer tumours.
-
Achieved 85.77% accuracy and 88.85% F1-Score by training a DenseNet CNN model using TensorFlow.
-
Used Flask framework and SQLAlchemy to develop the platform.
-
Implemented a user authentication and patient management system for secure storage of diagnosis history.
-
Deployed the website to Heroku using a PostgreSQL database.
Dataset of histological images