Machine learning analysis reveals tumor stiffness and hypoperfusion as biomarkers predictive of cancer treatment efficacy

Abstract

In the pursuit of advancing cancer therapy, this study explores the predictive power of machine learning in analyzing tumor characteristics, specifically focusing on the effects of tumor stiffness and perfusion (i.e., blood flow) on treatment efficacy. Recent advancements in oncology have highlighted the significance of these physiological properties of the tumor microenvironment in determining treatment outcomes. We delve into the relationship between these tumor attributes and the effectiveness of cancer therapies in preclinical tumor models. Utilizing robust statistical methods and machine learning algorithms, our research analyzes data from 1365 cases of various cancer types, assessing how tumor stiffness and perfusion influence the efficacy of treatment protocols. We also investigate the synergistic potential of combining drugs that modulate tumor stiffness and perfusion with standard cytotoxic treatments. By incorporating these predictors into treatment planning, our study aims to enhance the precision of cancer therapy, tailoring treatment to individual tumor profiles. Our findings demonstrate a significant correlation between stiffness/perfusion and treatment efficacy, highlighting a new way for personalized cancer treatment strategies.

Keywords

• Ultrasound imaging
• Shear wave elastography
• Chemo-immunotherapy
• Precision oncology
• Biomarkers