Oct 2, 2012
Omni-tomography: the ultimate in multimodality imaging?
Multimodality imaging and targeted imaging agents are enablers of early disease screening, cancer staging, therapeutic assessment, interventional guidance and other aspects of personalized medicine.1,2 PET/CT, SPECT/CT and PET/MRI are powerful examples of the synergy provided by dual-modality imaging, in particular for oncology and cardiology. Indeed, there are no longer any lone PET scanners; today, all are coupled with CT scanners.
In 2009, Simon Cherry of the University of California–Davis raised a question that may still represent the thought processes of experts in this field: "Is the fusion of PET and SPECT with CT the ultimate answer in multimodality imaging, or is it just the first example of a more general trend towards harnessing the complementary nature of the different modalities on integrated imaging platforms?".1
Preclinical and clinical studies both depend on in vivo tomography, often requiring separate evaluation by different imaging modalities. Over the past decade, two approaches have been used for multimodality fusion: software-based registration of images acquired separately, and hybrid scanners such as PET/CT, SPECT/CT and PET/MRI. While there are intrinsic limitations to both approaches, the seamless fusion of multiple imaging modalities has been particularly challenged by the spatial conflicts of these scanners.
To overcome this obstacle, we recently published an article that lays the foundation for the integration of multiple major tomographic scanners into a single gantry.3 This new thinking – known as omni-tomography – represents the next stage of multimodality fusion for biomedical imaging.
The enabling technology for omni-tomography is "interior tomography".4,5 Traditionally, CT theory targets theoretically exact image reconstruction of a whole cross-section or volume, assuming projections are measured without any truncation. On the other hand, many important real-world problems are localized, or at least often observed within a relatively small region-of-interest (ROI).
Conventional wisdom holds that an internal ROI cannot be exactly reconstructed solely from truncated projection data measured through the ROI – known as the interior problem. When a traditional CT algorithm is applied to truncated projections after artificial data extrapolation, artefacts are generated that overlap features in the ROI. In 2007, our group proved that the interior problem can be exactly and stably solved, provided that a sub-region is known within an ROI. We call this new thinking interior tomography, to indicate the theoretically exact nature of such an ROI reconstruction. Similar results have been independently reported by others.
Precise prior knowledge of a sub-region, however, is not always available. For example, blood density can no longer be assumed as a constant when contrast injection is involved. Hence, we were motivated to relax the requirement of prior knowledge. For that purpose, we analytically and experimentally showed that the interior problem indeed permits a unique and stable solution if the ROI is piecewise "smooth"; i.e., each piece can be fitted into a polynomial function6,7. Indeed, this piecewise polynomial assumption is a quite general image model.
In our latest work, we elevated interior tomography from its origin in CT to a general tomographic principle, and demonstrated its validity for different tomographic modalities, including SPECT, MRI and phase-contrast tomography, among others.3 As a result, the relevant tomographic scanners can be made slimmer or more compact, and can be integrated together to allow comprehensive and simultaneous data acquisition of an ROI.
A look ahead
Once established, omni-tomography will prove desirable as a systematic, compact imaging option that offers spatiotemporal synchrony of diversified features. First of all, it is designed to collect complementary tomographic datasets simultaneously, revealing spatiotemporal links that are critical for physiological, pathological and pharmaceutical studies. It also has the potential to improve some important image-guided procedures, such as MRI- and CT-guided interventions. Economically, omni-tomography could be a cost-effective mode relative to a fully-fledged imaging centre, by saving on equipment space, scan time, patient throughput and staffing.
Omni-tomography has many potential clinical applications. For example, a combined CT-MRI system could be used as a screening device for cardiac and stroke applications. An interior CT-MRI scanner can also target the fast-beating heart for registration of functions and structures, delivery of drugs or stem cells, and guidance of complicated procedures such as heart valve replacement.
In our aforementioned omni-tomography paper, we presented a unified interior CT-MRI reconstruction strategy. This unification has the potential to greatly reduce radiation dose when MRI-aided interior CT reconstruction is implemented. On the other hand, CT-aided interior MRI reconstruction can generate high-resolution details. Omni-tomography is an emerging frontier, with many other possible research and commercial opportunities.
So when will omni-tomography come into reality? With sufficient resources, it is estimated that the interior CT-MRI scanner could be prototyped in a few years. Conceptually, it is not difficult to achieve omni-tomography. We recognize that the rotating X-ray source and detector can interfere with MR imaging. One solution is to use a stationary CT architecture in which multiple sources are placed around a patient, along with the corresponding detector pieces. All the X-ray beams focus on an ROI, representing a few-view imaging setup. Since the CT subsystem is not moving, the electromagnetic shielding for the MRI subsystem can be simplified.
Currently, for satisfactory image reconstruction, we need a large number of projections. However, in the fused CT-MRI scanner, the synergy between CT and MRI data can be utilized to greatly reduce the number of X-ray projections. A comprehensive cost-performance analysis will be needed to get an accurate estimate of how much an omni-tomography scanner will cost, but we expect a cost that will not differ significantly from that of current scanners.
Our team welcomes feedback and collaboration.
About the author
Ge Wang is Pritchard Professor and Director of the Biomedical Imaging Division at Virginia Tech-Wake Forest University School of Biomedical Engineering & Sciences.
1 S R Cherry 2009 Multimodality imaging: Beyond PET/CT and SPECT/CT Semin. Nucl. Med. 39 348
2 J A Patton et al 2009 Hybrid imaging technology: From dreams and vision to clinical devices Semin. Nucl. Med. 39 247
3 G Wang et al 2012 Towards omni-tomography – Grand fusion of multiple modalities for simultaneous interior tomography PLoS ONE 7 e39700
4 Y B Ye et al 2007 A general local reconstruction approach based on truncated Hilbert transform International Journal of Biomedical Imaging ID: 63634
5 H Yu and G Wang 2009 Compressive sensing based interior tomography Phys. Med. Biol. 54 2791
6 J S Yang et al 2012 High order total variation minimization for interior SPECT Inverse Problems 28 015001
7 E Katsevich et al 2012 Stability of the interior problem with polynomial attenuation in the region of interest Inverse Problems 28 065022