Radiomics: Current Applications and Future Directions.

Area of Science:

• Oncology
• Medical Imaging
• Data Science

Background:

• Radiomics extracts quantitative imaging features for noninvasive tumor analysis in precision oncology.
• Solid tumors present challenges in reflecting heterogeneity and dynamic evolution via conventional imaging and biopsies.
• Existing radiomics approaches face bottlenecks in pan-cancer generalization, biological interpretability, and clinical translation.

Purpose of the Study:

• To outline a unified radiomics pipeline for solid tumors.
• To summarize standardization strategies for multi-center, multi-scanner, and multi-cancer data.
• To review pan-cancer clinical applications and discuss future directions for radiomics.

Main Methods:

• Systematic review of radiomics literature across various solid tumors (lung, breast, colorectal, liver, glioma, prostate).
• Analysis of radiomics pipeline, standardization strategies, and clinical applications.
• Discussion of multi-omics integration, interpretability, and translational challenges.

Main Results:

• Radiomics has evolved from static to dynamic assessment, showing promise in early detection, molecular characterization, treatment response prediction, and prognosis.
• Standardization is crucial for multi-center, multi-scanner, and multi-cancer data heterogeneity.
• Lung cancer serves as a key example, with evidence integrated from other major solid tumors.

Conclusions:

• Radiomics is a transformative tool for precision oncology, but faces challenges in generalization, interpretability, and clinical translation.
• Addressing bottlenecks like domain shift and reproducibility is essential for clinical adoption.
• Future directions include foundation models, causal inference, and federated learning to enhance radiomics' clinical utility.