Ge Wang, John A. Clark and Edward T. Crossan Chaired Professor of Engineering
Wang’s major field of research, medical imaging, features an array of technologies: CT, MRI, positron emission tomography (PET), and others. Each is invaluable for certain aspects of healthcare. Each has substantial limitations. “If we could combine all these modalities in one unit, and capture all data in one scan,” noted Wang, “we could do much better imaging than what we can today.”
Until recently, that was a big if — too big for serious consideration. But it is precisely what Wang is pursuing.
Cracking the Interior Problem
Breakthroughs are not new for Wang, who came to Rensselaer as the John A. Clark and Edward T. Crossan Chaired Professor of Engineering in 2013. He and his colleagues introduced the field of bioluminescence tomography in 2004. Their resolution of the debate over fan-beam spiral CT technology (vs. traditional “step-and-shoot” CT imaging) drove its rise to prominence as the dominant paradigm. Most notably, in 1991, Wang was the lead author of the first papers on the algorithms for spiral (helical) cone-beam CT imaging — the technology that, today, enables 100 million CT scans every year in hospitals worldwide.
Even with these advances, CT imaging still faced a stiff challenge with the enigmatic name of “the interior problem.” Traditional CT methods could not circumvent the problem, the root of which is data truncation — a loss of desirable global perspectives about local regions of interest (ROI) that are the focus of most medical investigations. The missing nonlocal data often result in images that are far less than perfect. As Wang wrote in a paper with his research partner, Hengyong Yu, in 2012, “When a traditional CT algorithm is applied for local reconstruction [of the image] from truncated data, quantitative accuracy is lost in a reconstructed image, compromising its diagnostic value significantly.”
The interior problem had baffled people for decades. Some researchers tried to solve it with beams targeted to the ROI, but their algorithms fell short in significant ways: providing only approximate reconstructions of the ROI, or capturing only significant boundaries in the image. Wang and his collaborators set to work on an entirely different solution: interior tomography for the theoretically exact, practically accurate CT imaging of an ROI using only data that directly involve the ROI. The breakthrough was made possible via rigorous analysis and novel utilization of general knowledge on images (for instance, the fact that images are piecewise-regular as described by polynomials).
The advantages would be great. “Interior tomography allows an exact reconstruction from less data,” they wrote. “Less data are equivalent to lower radiation dose…. Interior tomography allows a smaller detector size, a faster frame rate and more imaging chains in a gantry space, all of which contribute to an accelerated data acquisition process.” The shorter the data acquisition takes, the better the time resolution will be, which is important for cardiac imaging, image-guided intervention, and other applications. Interior tomography also enables scanning of large objects or samples, which is critically important in such areas as X-ray phase contrast imaging and industrial non-destructive evaluation.
Making the Next Leap
But Wang and his collaborators had even more in mind. In a paper entitled “Towards Omni-tomography,” which appeared on PLoS ONE in 2012, they boldly extended the application of interior tomography from CT to the entire sweep of imaging modalities, introducing the general interior tomography principle and setting the stage for what they called “the holy grail of biomedical imaging”—“the integration of multiple major tomographic scanners into a single gantry.” In this vision, one machine could provide all data to image cardiac and stroke damage, capture elusive functions and structures, and guide complex surgical procedures, among many other applications where traditional imaging modalities are insufficient.
“We leapt from special interior tomography, involving CT only, to general interior tomography,” Wang explained. “It is, in a small way, analogous to Einstein’s special and general theories of relativity. That comparison speaks to how important omni-tomography could potentially be for the field of medical imaging. It is our grand unified theory.”
From Theory to Reality
Since their initial work on interior tomography in 2007, Wang’s team has been striving to turn the theory of omni-tomography into reality. A new goal of his lab is to explore multi-physics interactions in the omni-tomography framework. In one related project, his students are investigating whether X-ray excited nanophosphors could alter parameters observable by an MRI scanner. Wang hopes that omni-tomography will be not only important for simultaneous imaging in complementary mechanisms (such as CT-MRI, CT-PET, and CT-SPECT) but also necessary for novel information from physical interactions.
If Wang has his way, this research will result in a prototype—and he is already working on it. Among other possibilities, he has developed a proposal to build a combined CT-MRI scanner, using the notion of omni-tomography on a small scale. The first use of such a machine will likely capture images of animals.
Wang’s contributions have received a great deal of attention from leaders in the imaging field. “Present-day hybrid or combined imaging modalities such as SPECT/CT, PET/CT, and PET/MR continue to show promise in providing information that influence patient management. Though combined, these imaging tests are acquired sequentially, which poses unique challenges,” noted Mannudeep Kalra, MD, a professor at Harvard Medical School and a radiologist with Massachusetts General Hospital. “As a clinical radiologist, I believe that omni-tomography will revolutionize radiology as a one-stop and one-shop imaging equipment. In my opinion, development of omni-tomography will impact diagnostic workup in both urgent and non-urgent situations, spanning from coronary artery diseases and strokes to liver diseases and musculoskeletal ailments. From a technical perspective, the prospect of substantial radiation dose reduction for CT and significant time reduction for MRI with omni-tomography are groundbreaking.”
In other words, Wang is pursuing a breakthrough. And it may not be far off.
Breakthroughs are not new for Wang…he and his colleagues introduced the field of bioluminescence tomography in 2004.