Dataset Information

Nondestructive Detection of Targeted Microbubbles Using Dual-Mode Data and Deep Learning for Real-Time Ultrasound Molecular Imaging.

ABSTRACT: Ultrasound molecular imaging (UMI) is enabled by targeted microbubbles (MBs), which are highly reflective ultrasound contrast agents that bind to specific biomarkers. Distinguishing between adherent MBs and background signals can be challenging in vivo. The preferred preclinical technique is differential targeted enhancement (DTE), wherein a strong acoustic pulse is used to destroy MBs to verify their locations. However, DTE intrinsically cannot be used for real-time imaging and may cause undesirable bioeffects. In this work, we propose a simple 4-layer convolutional neural network to nondestructively detect adherent MB signatures. We investigated several types of input data to the network: "anatomy-mode" (fundamental frequency), "contrast-mode" (pulse-inversion harmonic frequency), or both, i.e., "dual-mode", using IQ channel signals, the channel sum, or the channel sum magnitude. Training and evaluation were performed on in vivo mouse tumor data and microvessel phantoms. The dual-mode channel signals yielded optimal performance, achieving a soft Dice coefficient of 0.45 and AUC of 0.91 in two test images. In a volumetric acquisition, the network best detected a breast cancer tumor, resulting in a generalized contrast-to-noise ratio (GCNR) of 0.93 and Kolmogorov-Smirnov statistic (KSS) of 0.86, outperforming both regular contrast mode imaging (GCNR = 0.76, KSS = 0.53) and DTE imaging (GCNR = 0.81, KSS = 0.62). Further development of the methodology is necessary to distinguish free from adherent MBs. These results demonstrate that neural networks can be trained to detect targeted MBs with DTE-like quality using nondestructive dual-mode data, and can be used to facilitate the safe and real-time translation of UMI to clinical applications.

SUBMITTER: Hyun D

PROVIDER: S-EPMC7793556 | biostudies-literature |

REPOSITORIES: biostudies-literature

ACCESS DATA

Similar Datasets

Project description:BackgroundThere is growing interest in limb contrast-enhanced ultrasound (CEU) perfusion imaging for the evaluation of peripheral artery disease. Because of low resting microvascular blood flow in skeletal muscle, signal enhancement during limb CEU is prohibitively low for real-time imaging. The aim of this study was to test the hypothesis that this obstacle can be overcome by intermediate- rather than low-power CEU when performed with an acoustically resilient microbubble agent.MethodsViscoelastic properties of Definity and Sonazoid were assessed by measuring bulk modulus during incremental increases in ambient pressure to 200 mm Hg. Comparison of in vivo microbubble destruction and signal enhancement at a mechanical index (MI) of 0.1 to 0.4 was performed by sequential reduction in pulsing interval from 10 to 0.05 sec during limb CEU at 7 MHz in mice and 1.8 MHz in dogs. Destruction was also assessed by broadband signal generation during passive cavitation detection. Real-time CEU perfusion imaging with destruction-replenishment was then performed at 1.8 MHz in dogs using an MI of 0.1, 0.2, or 0.3.ResultsSonazoid had a higher bulk modulus than Definity (66 ± 12 vs 29 ± 2 kPa, P = .02) and exhibited less inertial cavitation (destruction) at MIs ≥ 0.2. On in vivo CEU, maximal signal intensity increased incrementally with MI for both agents and was equivalent between agents except at an MI of 0.1 (60% and 85% lower for Sonazoid at 7 and 1.8 MHz, respectively, P < .05). However, on progressive shortening of the pulsing interval, Definity was nearly completely destroyed at MIs ≥ 0.2 at 1.8 and 7 MHz, whereas Sonazoid was destroyed only at 1.8 MHz at MIs ≥ 0.3. As a result, real-time CEU perfusion imaging demonstrated approximately fourfold greater enhancement for Sonazoid at an MI of 0.3 to 0.4.ConclusionsRobust signal enhancement during real-time CEU perfusion imaging of the limb is possible when using intermediate-power imaging coupled with a durable microbubble contrast agent.

Project description:Image-guided and target-selective modulation of drug delivery by external physical triggers at the site of pathology has the potential to enable tailored control of drug targeting. Magnetic microbubbles that are responsive to magnetic and acoustic modulation and visible to ultrasonography have been proposed as a means to realize this drug targeting strategy. To comply with this strategy in vivo, magnetic microbubbles must circulate systemically and evade deposition in pulmonary capillaries, while also preserving magnetic and acoustic activities in circulation over time. Unfortunately, challenges in fabricating magnetic microbubbles with such characteristics have limited progress in this field. In this report, we develop magnetic microbubbles (MagMB) that display strong magnetic and acoustic activities, while also preserving the ability to circulate systemically and evade pulmonary entrapment. METHODS:We systematically evaluated the characteristics of MagMB including their pharmacokinetics, biodistribution, visibility to ultrasonography and amenability to magneto-acoustic modulation in tumor-bearing mice. We further assessed the applicability of MagMB for ultrasonography-guided control of drug targeting. RESULTS:Following intravenous injection, MagMB exhibited a 17- to 90-fold lower pulmonary entrapment compared to previously reported magnetic microbubbles and mimicked circulation persistence of the clinically utilized Definity microbubbles (>10 min). In addition, MagMB could be accumulated in tumor vasculature by magnetic targeting, monitored by ultrasonography and collapsed by focused ultrasound on demand to activate drug deposition at the target. Furthermore, drug delivery to target tumors could be enhanced by adjusting the magneto-acoustic modulation based on ultrasonographic monitoring of MagMB in real-time. CONCLUSIONS:Circulating MagMB in conjunction with ultrasonography-guided magneto-acoustic modulation may provide a strategy for tailored minimally-invasive control over drug delivery to target tissues.

Dataset Information

Nondestructive Detection of Targeted Microbubbles Using Dual-Mode Data and Deep Learning for Real-Time Ultrasound Molecular Imaging.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets