Radiomic-Based Approaches in the Multi-metastatic Setting: A Quantitative Review

Area of Science:

• Radiomics and Medical Imaging Analysis
• Oncology and Metastasis Research
• Quantitative Medical Imaging

Background:

• Traditional radiomics focuses on single lesions, potentially missing inter-lesion heterogeneity in multi-metastatic disease.
• Current methods for combining multi-lesion radiomic data are diverse and lack standardization.
• This study addresses the need for established best practices in multi-lesion radiomic analysis.

Purpose of the Study:

• To quantitatively review and assess methodologies for integrating radiomic features from multiple lesions.
• To evaluate the replicability and performance of different multi-lesion radiomic analysis methods.
• To provide insights and guide future research toward best practices in multi-lesion radiomic analysis.

Main Methods:

• Conducted a literature search to identify methods for multi-lesion radiomic data integration.
• Replicated identified methods using author code or paper-based reconstruction.
• Applied and compared ten mathematical methods across three distinct metastatic datasets (16,850 lesions, 3,930 patients).

Main Results:

• Observed variable performance among the ten methods across the three datasets.
• No single method consistently outperformed others across all metastatic scenarios.
• Averaging methods performed better in colorectal liver metastases; feature concatenation excelled in soft tissue sarcoma.

Conclusions:

• Radiomic features can be effectively combined to predict patient outcomes in multi-metastatic settings, but optimal approaches vary by metastatic site.
• This research fills a critical gap by examining challenges in multi-lesion radiomic analysis.
• The study offers valuable insights into effective radiomic analysis strategies for diverse metastatic scenarios.