Phase-change nanodroplets (PCNDs) have been used as controllable theranostic agents in diverse therapeutic and diagnostic scenarios over the years. However, there is still a pressing need to monitor the accurate distribution of non-flowing extravascular PCNDs by clinical ultrasound for further quantitative analysis. Here, we propose a spatiotemporally super-resolved ultrasound monitoring technique based on tuned post-activation dynamics of high-bulk-boiling-point PCNDs with customized perfluorocarbon cores. The underlying idea was to make the stochastic recondensation-induced contrast signal dominate the total post-activation signal, thus obtaining detectable and resolvable recondensation signals at physiological temperature. The recondensation signal was extracted by inter-frame subtraction and then processed by a deep learning-based detection algorithm tailored to the recondensation signal patterns. Experiments in tissue-mimicking phantoms demonstrate that co-restricting the concentration of PCNDs and the focus-wave activation pulse duration could help generate recondensation signals with ideal detectability and sparsity for accurate monitoring. Meanwhile, the quantitative analysis of the super-resolved results shows the spatiotemporal sensitivity of the proposed technique under varying concentrations and activation pulse durations, which was consistent with the patterns drawn from the total post-activation signal and existing theories on post-activation dynamics of PCNDs. This technique may be suitable for in-depth extravascular monitoring and dose analysis for PCNDs-involved therapies.
Read full abstract