Radiomics and deep learning in lung cancer

Area of Science:

• Medical imaging analysis
• Computational oncology
• Artificial intelligence in medicine

Background:

• Lung malignancies are complex, necessitating advanced characterization beyond traditional methods.
• Radiomics and deep learning offer powerful tools for analyzing medical imaging data in lung cancer.
• Existing methods struggle with nuanced aspects of diagnosis, staging, and treatment response.

Purpose of the Study:

• To provide an update on the current applications and status of radiomics in lung cancer.
• To highlight the integration of radiomics and deep learning in clinical practice.
• To explore the future potential of radiomics in optimizing the lung cancer care pathway.

Main Methods:

• Utilizing radiomic features from computed tomography (CT) and positron-emission tomography (PET) scans.
• Applying deep learning models for segmentation, risk stratification, and treatment response prediction.
• Leveraging data from various imaging modalities including cone beam CT and 4D CT.

Main Results:

• Radiomics and deep learning models successfully detect nodules and distinguish malignant from benign lesions.
• These models aid in characterizing histology, staging, and genotyping lung tumors.
• Applications extend to predicting treatment side effects like pneumonitis and differentiating recurrence from injury.

Conclusions:

• Radiomics and deep learning are significantly impacting lung cancer diagnosis, treatment, and follow-up.
• These technologies offer a comprehensive approach to optimizing the end-to-end patient care chain.
• Future integration promises further advancements in personalized lung cancer management.