Derivative Texture Analysis of Tumor Histology for Evaluating Ultrasound Nanobubble-Driven Radiation Enhancement

Microbubbles (MBs) have been used as a radiosensitizer, and their combination with ultrasound (US) has emerged as a promising strategy to improve radiotherapy outcomes in tumor treatment. Nanobubbles (NBs), which are about 1000 times smaller than MB and have enhanced stability, have been considered to have the potential for further radiosensitization enhancement. In this study, tumor-bearing models were treated with and without nanobubble-ultrasound (NBUS) treatment before radiotherapy (XRT) to evaluate the radiosensitization. The time intervals between the NB injection and US exposure, and between US and XRT, were optimized to maximize the therapeutic efficacy. A derivative texture analysis was applied to extract microstructural features from haematoxylin and eosin (H&E)-stained images. A one-way ANOVA test combined with a k-NN classifier and Tukey's Honestly Significant Difference (HSD) test was used to identify the best features for assessing treatment outcomes. These were applied to H&E-stained tumor sections for evaluating microstructural alterations associated with NB-driven radiosensitization. The results indicated that both the 2 Gy and 8 Gy radiation groups showed enhanced treatment outcomes when NBUS therapy was administered before radiotherapy. Notably, NBUS + 2 Gy could achieve outcomes comparable to 8 Gy only. Additionally, a derivative texture analysis was demonstrated to be a promising and powerful technique for analyzing H&E images. This study provides a quantitative framework for assessing nanobubble-enhanced radiotherapy.