Related Experiment Video
Updated: Aug 5, 2026
08:05
Lung CT Segmentation to Identify Consolidations and Ground Glass Areas for Quantitative Assesment of SARS-CoV Pneumonia
Published on: December 19, 2020
Deep Learning-Based Lung Cancer Segmentation and Volumetric Analysis Using CT Imaging: A Comprehensive Survey
Tugce Gulseren Tezel1, Mehmet Turkan2, Ebru Sayilgan3
1Department of Electrical and Electronics Engineering, Izmir University of Economics, Izmir, Turkey.
Annals of Biomedical Engineering
|July 24, 2026
Summary
Deep learning improves lung cancer volumetric analysis on CT scans, offering more accurate tumor burden assessment than traditional methods. Challenges remain in clinical translation, but new AI models show promise for robust, interpretable lung cancer imaging.
Area of Science:
- Medical Imaging
- Artificial Intelligence
- Oncology
Background:
- Lung cancer is a leading cause of cancer death, necessitating precise tumor burden and treatment response evaluation.
- Traditional methods like RECIST lack 3D and longitudinal tumor change assessment, creating a clinical need for advanced volumetric analysis.
- Deep learning (DL) offers automated, reproducible tumor segmentation in CT scans, aligning with this clinical demand.
Purpose of the Study:
- To systematically review deep learning methods for lung cancer segmentation and volumetric analysis on CT scans.
- To analyze the relationship between segmentation performance, volumetric reliability, and clinical utility for treatment response and monitoring.
- To identify challenges and future research directions for AI-driven lung cancer imaging.
Main Methods:
- Categorized existing research by DL architecture (e.g., U-Net, CNN-Transformer), learning strategy, and dataset characteristics.
- Examined publicly available datasets (LIDC-IDRI, RIDER Lung CT, NSCLC-Radiomics) for nodule detection, reproducibility, and prognostic modeling.
- Focused on the link between segmentation accuracy, volumetric biomarker reliability, and clinical application.
Main Results:
- DL models, including U-Net variants and multimodal PET/CT strategies, have advanced lung cancer segmentation.
- A gap persists between segmentation accuracy and clinically meaningful volumetric assessment models.
- Public datasets are utilized for evaluating DL models in nodule detection and prognostic volumetric modeling.
Conclusions:
- DL-based volumetric analysis shows potential to overcome limitations of linear tumor assessment in lung cancer.
- Clinical implementation faces hurdles: limited data, annotation variability, scanner sensitivity, interpretability, and multimodal integration.
- Future directions include transformer models, self-supervised learning, generative modeling, and uncertainty-aware systems for improved AI in lung cancer imaging.
