Tools
Tools
Cruz Sánchez, José Alberto (2025) Hybrid CNN-RNN Models and Explainability in Early Lung Cancer Detection from Chest Radiographs: A Study on Enhancing Radiologist Confidence and Diagnostic Accuracy. Masters thesis, Dublin, National College of Ireland.
Preview |
PDF (Master of Science)
Download (847kB) | Preview |
Preview |
PDF (Configuration Manual)
Download (540kB) | Preview |
Abstract
Lung cancer causes over a million deaths each year, making it one of the most serious health problems worldwide. When the disease is found early, people have a much better chance of survival. But in practice, spotting early signs is not easy. Chest X-rays are common in routine screenings, yet the first indicators of cancer are often faint, and even trained radiologists can overlook them.
This work explores how artificial intelligence might support early detection. It introduces a model that brings together two kinds of deep learning systems: one that looks at details in an image and another that considers how those details fit together. The first part, a convolutional network, scans the image to find anything unusual. The second, a recurrent network, tries to understand how these findings relate to one another across the lungs.
Just giving a result is not enough. In medicine, it is important to know how and why a decision is made. That’s why this model also includes explainability tools. These tools allow the system to highlight which parts of the image influenced the result the most. This kind of feedback can help doctors better understand and trust what the AI is saying.
The model is trained using the ChestX-ray14 dataset. While this project focuses on technical development and evaluation, future work could include studies involving radiologists to explore how explainable results affect clinical decision-making and trust.
The goal is simple: build something that works and make sure people can trust it. With better tools that are both reliable and easy to understand, it may be possible to catch lung cancer earlier and support doctors in making faster, clearer decisions.
| Item Type: | Thesis (Masters) |
|---|---|
| Supervisors: | Name Email Anand, Devanshu UNSPECIFIED |
| Subjects: | Q Science > QH Natural history > QH301 Biology > Methods of research. Technique. Experimental biology > Data processing. Bioinformatics > Artificial intelligence Q Science > Q Science (General) > Self-organizing systems. Conscious automata > Artificial intelligence Q Science > Life sciences > Medical sciences > Pathology > Tumors > Cancer R Medicine > Healthcare Industry |
| Divisions: | School of Computing > Master of Science in Artificial Intelligence |
| Depositing User: | Ciara O'Brien |
| Date Deposited: | 12 Aug 2026 08:41 |
| Last Modified: | 12 Aug 2026 08:41 |
| URI: | https://norma.ncirl.ie/id/eprint/9502 |
Actions (login required)
 |
View Item |