Integrating Imaging and Circulating Tumor DNA Features for Predicting Patient Outcomes

Area of Science:

• Oncology
• Biomarker Discovery
• Translational Medicine

Background:

• Accurate tumor response evaluation and relapse risk prediction are critical clinical needs.
• Current methods include imaging (PET/CT, MRI) and minimally invasive liquid biopsies (circulating tumor DNA).
• Combining imaging and circulating tumor DNA (ctDNA) biomarkers may enhance patient outcome prediction.

Purpose of the Study:

• To review original research combining quantitative imaging and ctDNA biomarkers for predictive model development.
• To assess the feasibility and current state of integrated imaging-ctDNA predictive models.
• To identify needs for future research in this emerging field.

Main Methods:

• Systematic literature search for studies integrating quantitative imaging and ctDNA biomarkers.
• Analysis of studies that developed prognostic or response-predictive models.
• Focus on regression and machine learning approaches for model building.

Main Results:

• Seven studies developed prognostic models for survival outcomes (recurrence-free, progression-free, overall survival).
• Three studies focused on treatment response prediction using endpoints like tumor volume and response rates.
• The field is nascent, with limited studies and modest cohort sizes.

Conclusions:

• Combining imaging and ctDNA biomarkers is feasible for developing multivariable predictive models.
• Current models show promise but require validation in larger, diverse patient populations.
• Future research should focus on larger studies to improve model accuracy and generalizability.