Noninvasive tumor imaging with high-frequency ultrasound and microPET in small animals

In this study, we used a micro-ultrasound (microUS) system that we developed in-house as an alternative method for tumor growth calipers. In addition, microUS was combined with small-animal positron-emission tomography (microPET) for tumor metastatic assessment. MicroUS provides anatomical information that can be used for tumor volume measurements while microPET is a functional imaging method with positron-emitting radiophamaceuticals, such as 18F-labeled deoxyglucose, [18F]FDG. In this study, microUS and microPET were performed in a mouse tumor longitudinal study (2-8 weeks), both with 3D tumor segmentation and volume measurements. Compared with vernier calipers, microPET generally overestimated tumor volumes during weeks 2-4 due to its inadequate spatial resolution. During weeks 5-8, standard deviations of microPET results were large due to tumor hypoxia or necrosis. On the contrary, microUS tumor volume measurements were more reliable as they were less affected by these factors. Nonetheless, microUS is not able to provide functional information similar to that provided by microPET. Therefore, microUS and microPET are complementary to each other as microUS has superior spatial resolution and microPET provides functional information, such as hypoxia or necrosis in the progression of the tumor. With image registration and fusion, the combination can be a valuable tool for cancer research.

A three-dimensional registration method for microUS/ microPET multimodality small-animal imaging

Small-animal models are used extensively in disease research, genomics research, drug development and developmental biology. The development of noninvasive small-animal imaging techniques with adequate spatial resolution and sensitivity is therefore of prime importance. In particular, multimodality small-animal imaging can provide complementary information. This paper presents a method for registering high-frequency ultrasonic (microUS) images with small-animal positron-emission tomography (microPET) images. Registration is performed using six external multimodality markers, each being a glass bead with a diameter of 0.43-0.60 mm, with 0.1 microl of [18F]FDG placed in each marker holder. A small-animal holder is used to transfer mice between the microPET and microUS systems. Multimodality imaging was performed on C57BL/6J black mice bearing WF-3 ovary cancer cells in the second week after tumor implantation and rigid-body image registration of the six markers was also performed. The average registration error was 0.31 mm when all six markers were used and increased as the number of markers decreased. After image registration, image segmentation and fusion are performed on the tumor. Our multimodality small-animal imaging method allows structural information from microUS to be combined with functional information from microPET, with the preliminary results showing it to be an effective tool for cancer research.