Lung Tumor T-Stage Classification

Overview

The manual grading of a lung tumor’s T-stage can be a time-consuming and easily botched task. A delayed or misdiagnosis can be fatal, especially for an upper stage tumor. Early-stage tumors are also often missed. Automating this classification process can significantly reduce time between screening and diagnosis, minimize the risk of human error, and make cancer screenings more affordable. Our proposed model, LungVision, aims to automate this process.

Community Benefit

The preprocessing method utilized in LungVision proved to extract tumor features without removing important tumor features, as it performed slightly better than the same neural network trained on the original scans. Upon inspection, the preprocessed CT scans are also noticeably less noisy while still retaining the tumor shape.

Team Members

• Garrett Reardon - greardon2018@fau.edu
• Joshua Lavieri - jlavieri2018@fau.edu
• Danilo Montalvo - dmontalvo2016@fau.edu
• Giovanna Yuen - gyuenorozco2016@fau.edu
• Raymond Budoff-Kingsland - rbudoffkings2018@fau.edu 

Sponsor

Dr. Hanqi Zhuang