Ultrasound stiffness and perfusion markers correlate with tumor volume responses to immunotherapy

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

“…Pathophysiological properties of the tumor microenvironment (TME) can pose major barriers to the efficacy of cancer therapies [1] . Among these properties, mechanical aspects of the TME, and in particular the abnormal elevation of tumor stiffness and the lack of functional tumor blood vessels to carry blood, known as hypoperfusion (or ischemia), have shown not only to negatively affect cancer therapy but also serve as valuable indicators for predicting treatment outcomes [2] , [3] , [4] , [5] , [6] , [7] . Tumors are understood to be heterogeneous in both perfusion and stiffness, not just across cancer types, but even within a single type and within a single tumor itself during progression.…”

“…Additionally, hypoxia affects the function of T lymphocytes and other immune cells, while low blood flow and the dense/stiff TME create a physical barrier to T cell infiltration into the tumor. Studies have shown that increasing blood flow can reverse these pro-tumor effects and improve the effectiveness of immune checkpoint inhibition [ [4] , [5] , [6] , [16] , [17] , [18] , [19] , [20] , [21] , [22] , [23] , [24] , [25] ]. Therefore, tumor stiffness and the degree of perfusion have been considered as predictive biomarkers of response to therapy not only in preclinical but also in clinical studies [ [5] , [6] , [7] , 9 , [26] , [27] , [28] ].…”

“…Studies have shown that increasing blood flow can reverse these pro-tumor effects and improve the effectiveness of immune checkpoint inhibition [ [4] , [5] , [6] , [16] , [17] , [18] , [19] , [20] , [21] , [22] , [23] , [24] , [25] ]. Therefore, tumor stiffness and the degree of perfusion have been considered as predictive biomarkers of response to therapy not only in preclinical but also in clinical studies [ [5] , [6] , [7] , 9 , [26] , [27] , [28] ]. Predicting patients’ responses to therapy is crucial as therapy can lead to optimized, patient-specific treatments.…”

“…Mechanotherapeutic drugs, such as the tranilast, bosentan and ketotifen, employed in our study, work by targeting the TME to modulate its mechanical properties, particularly tumor stiffness. These drugs exert their effects through various mechanisms, including the inhibition of collagen and hyaluronan synthesis and the reduction of fibroblast activation [ 5 , [20] , [21] , [22] , [23] ]. By altering the ECM composition, these agents decrease the mechanical resistance within the tumor, thus reducing its stiffness.…”

“…Our research aimed to uncover the potential of stiffness and hypoperfusion as indicators of a tumor's response to therapy, proposing that predictive biomarkers based on these two properties warrant further development. [ 5 , [20] , [21] , [22] , [23] ] In our analysis, we utilized the XGBoost [29] algorithm, a leading method within the realm of supervised machine learning. XGBoost is particularly favored for its efficiency, scalability, and performance in predictive accuracy across a wide range of data types and domains [30] , [31] , [32] .…”