Seminar: Anna Mrukwa
On March 18, 2026, PhD student Anna Mrukwa gave a presentation at our department’s seminar entitled “Preliminary assessment of early lung cancer possibility based on whole lung low‑dose CT image”, focusing on decision‑support for lung cancer screening.
The main goal of this work was to estimate a patient’s risk of early lung cancer directly from a single low‑dose CT scan, without prior manual nodule annotation, by analysing the entire lung parenchyma and bronchovascular structures.
The presentation introduced a complete processing pipeline, including segmentation of lungs, lobes, mediastinum, airways and vessels, followed by experiments with modern CNN architectures (DenseNet, SEResNet, EfficientNet) for healthy vs. suspicious patient‑level classification across multi‑centre screening datasets.
This research is a step toward scalable tools that can preselect potentially healthy individuals and support radiologists in large‑scale lung cancer screening programmes based on low‑dose CT imaging.
