Molecular Imaging is the Next Gen

Single-slice CT scanners for routine clinical use were around for nine years before being replaced by the next generation—4-slice CT. They lasted 5 years, before 16 and 32-slice machines appeared. Eighteen months later, the 64-slice made its debut, and within nine months a 256-slicer was up and running in Japan. Where do we go after 256-slice CT, and how fast can hospitals write the $2 million checks it seems to take to buy the latest CT scanner? Most of the action in CT today seems more concerned with improving the software that runs them, to configure the scanners more efficiently and effectively, to produce better images by extracting more information out of the image data, and to reduce radiation dosage to the patient.

It’s a similar story with MRI and PET. All these technologies have given us increasingly better anatomical and functional imagery, but they provide only millimeter-level resolution and may be approaching their limits in that respect. The next generation of imaging will probably not be a 512-slice or 1,024-slice CT. It will be molecular imaging.

Today, we image the haystack with CT, PET, MRI, ultrasound, etc., to see what is inside. It’s a discovery process. With a new imaging technique able to detect a single cancerous cell that has broken free from a tumor and is traveling through the bloodstream, we can ignore the haystack and directly locate and image a needle we have good reason to believe is inside. It’s a confirmation process.

The new experimental technique, not yet in animal testing, involves a nanoparticle contrast agent that eliminates background noise from the tissues. The 30-nanometer particle consists of an iron-oxide magnetic core with a thin gold shell that surrounds but does not touch the center. The gold shell absorb infrared light, and could also be used for optical imaging, delivering heat therapy, or attaching a biomolecule that would grab on to specific cells.

3-D Images at Nanoscale

Scientists using the Advanced Light Source (ALS) at the US Department of Energy’s Lawrence Berkeley National Laboratory have made high-resolution images, with features as small as 11 nanometers, of freeze-dried whole yeast cells. This means that 3D tomography of whole cells at equivalent resolution is now possible, though it requires “heroic” experimental effort. It currently takes a long time to image a single specimen, and full 3-D imaging of hydrated (watery) cells will take even more work. Nevertheless, this was described as “a big step in the right direction.”

The electron microscope can achieve similar resolution, but it requires very thin samples (a few hundred nanometers or less) so cannot be used to look through a whole cell.

1,000 Times More Sensitive MRI

Even though MRI and CT may never be able to compete with such technologies in terms of molecular resolution, there remain some important ways in which they advance. In speed, accuracy, and cost, for example. Early last year British researchers boosted MRI sensitivity a thousand fold using parahydrogen, the fuel that powers the space shuttle. This should lead to faster, more accurate, and cheaper diagnosis, and the ability to use MRI foe medical conditions currently undetectable by MRI.

The new method can generate in one second the information it would take today’s MRI 100 hours to obtain. Standard MRI usually takes information from the magnetic fields surrounding hydrogen atoms found in the water and fat in tissues. The new technique adds data from carbon-based molecules. Clinical testing of the technology is expected within five years, following animal testing.

Other potential ways to boost the sensitivity of MRI include dynamic nuclear polarization, which uses electron spin and requires hours of preparation at ultra-low temperatures.

4-D Heart MRI

Another way for MRI to remain competitive is to enhance its capabilities, such as improving its ability to image and measure blood flow in the heart. PC VIPR (Phase Contrast Vastly undersampled Isotropic Projection Reconstruction) offers just that: a new and still experimental way to image and measure blood flows in the heart and major arteries. But if and when clinical trials are successfully completed, in an anticipated 3-4 years, doctors will have an unprecedented window into the heart.

PC VIPR images show blood flows as a bundle of long threads or filaments, color-coded to indicate the speed of the flow at various locations. Blue represents the relatively slow flow when the heart is relaxed, green for the faster flow during contraction, and yellow and red for abnormally fast flows in a diseased or damaged heart. The direction of the flow and the effect of any obstructions or deviations on blood flow, such as in patients with Tetralogy of Fallot, are shown. The images are 4D (3D + time, essentially a 3D video.) PC VIPR can also help identify areas of weakening in the vessel wall or areas under increased stress, which can lead to aneurysms or build-up of damaging plaque.

In contrast to the current method of measuring blood flows, which takes between 45-90 minutes in the MRI and may require general anesthesia for younger patients, the new method takes 10 minutes in the MRI with no breath-holding, contrast agent, or (except in rare cases) anesthesia. (Ultrasound is another common technique used to look at blood flow in the heart, but the chest bone and other anatomy can make it hard to see some portions of the heart and vessels surrounding it.)

Predicting Brain Disorders from fMRI

A capability of MRI that was until recently used mainly in research is figuring out how the brain works. The technique, functional magnetic resonance imaging (fMRI), has not hitherto been used in clinical medicine for the individual patient. But now, using a new approach based on earlier fMRI work that mapped how the brain evolves as children age, fMRI can not only diagnose but also prognose brain disorders in individual patients. In a study, the new method was able to determine whether a child at risk for autism would actually develop the disorder, or what treatments might work best for that individual. It could also predict an individual’s age from just a five minute session in the MRI.

The new approach relies heavily on machine learning (a form of artificial intelligence) and on computing power that was not available as recently as five to 10 years ago.

Telepathy with Vegetative State Patients

fMRI has also been used to communicate with a patient diagnosed to be in a vegetative state. The patient, in this state for several months following a car accident, was able to correctly answer a series of yes or no questions asked—and answered—via fMRI. Subsequent research has shown that the case was not a fluke, and that fMRI can indeed be used—albeit at considerable expense and with considerable difficulty—to communicate with patients whose condition is caused by trauma rather than by oxygen deprivation.

The patient was asked questions relating to tennis and his personal life, giving answers that only he could have known (as confirmed by his family.) He was not asked medical questions, such as whether he was in pain, because of scientific, ethical, and legal concerns. Such concerns obviously now need to be resolved, since this study represents a potential breakthrough in treatment as well as in cognitive science and neurology.

Because of the expense and difficulty of using fMRI, other imaging methods such as electroencephalography (EEG) are being studied as potential alternatives.

The March of Ultrasound

Despite such advances in the “gold standard” imaging modalities, ultrasound remains a strong contender as a key imaging technology. Early last year, a prototype 3D ultrasound helmet provided real-time images of major blood vessels, including the direction of blood flow, in the brain. The helmet could be used on stroke victims to quickly determine whether the stroke was caused by bleeding or clot. If the latter, then (if other criteria were met) the patient could safely be given the life-saving drug tPA. Currently, suspected stroke patients are sent for a CT or MRI scan, which is time-consuming when time is of the essence. It takes an average of four hours for a potential stroke patient to receive a CT scan. The brain helmet scan can be completed in 15 to 30 minutes.

The helmet’s developers at Duke University “foresee a time in the near future when the brain helmet could transmit its images from a remote hospital, or from an ambulance, through cellular networks or the Internet to the neurological team at a stroke center,” according to a press release. Duke University Ultrasound Transducer Group and senior member of the research team. “Speed is important because the only approved medical treatment for stroke must be given within three hours of the first symptoms.”

The helmet prototype was tested on two healthy volunteers to assess its ability to accurately provide images of the major vessels of the brain. As of last year, the team was designing the next generation prototype with an eye toward reducing the size of its components and improving the efficiency of the signal from the transducers.

Sonic Scalpel, Imager

Ultrasound has therapeutic as well as imaging and diagnostic abilities. High-intensity focused ultrasound (HIFU) and lithotripsy have been around for several years and used to destroy precancerous growths and kidney stones. A potential improvement on those technologies is a new acoustic lens made of a metamaterial. The intensity can be adjusted to a level low enough to image without causing damage.

Ultrasound Superlens

Another acoustic metamaterial cloaks sound in the way that Harry Potter’s cloak of invisibility cloaks light. It represents a significant step toward creating higher-resolution ultrasound images (not to mention acoustic cloaks to hide submarines from sonar by channeling the sonar’s pings around the submarine to the other side, so the ping is never returned to the detector.) In ultrasound imaging, the metameterial should provide a sharper image, although actual applications are “a ways off,” a researcher said.

Nanotube X-ray Machines

What about CT? If we have just about reached the limits with regard to the number of slices, how else might it advance? As with MRI, the answer is: By working faster, and in CT’s case, by delivering a lower radiation dose. A new x-ray machine using carbon nanotubes is faster, produces sharper images, and could increase the accuracy of radiotherapy so it doesn’t harm normal tissue. Instead of the single tungsten emitter used in conventional x-ray machines, this one has an array of vertical carbon nanotubes that serve as hundreds of tiny electron guns. Tungsten takes time to warm up, whereas the nanotubes work instantly.

A CT scanner that used the new x-ray technology would be able to take many more images at nearly twice the resolution of existing scanners, allowing real-time imaging that would be a boon for cancer treatment—making it possible to take 3-D images during radiotherapy sessions, so the effect of the radiotherapy could be monitored in real time.

Clinical tests are under way in preparation for FDA clearance. Meanwhile the technology is already being used in biomedical research on laboratory animals.

Portable X-Ray

A flat-panel x-ray source that uses an array of pyroelectric crystal electron emitters, rather than a single tungsten source as in conventional machines, is under development. It could lead to abriefcase-sized x-ray machine powered by a laptop battery. Because each element of the array is controllable, the radiation dose can be more precisely delivered and possibly lowered.

A full-scale prototype could be ready in early 2011. Though that puts it considerably behind the carbon nanotubes technology described above, the pyroelectric crystal technology has the advantage that the panels can be made in volume using methods already employed in the microchip industry.

* * *

Here are some other developments we couldn’t quite squeeze into the themes of the main body of this issue.

Video Camera Captures Cellular Acivity

The Megaframe project has produced an ultra-fast, extremely high-resolution video camera for dozens of medical applications, including one to record “thought” processes as they travel along neurons. The project has already produced a video camera able to detect a single photon a million times a second, and hence record molecular processes in unprecedented detail.

Other ways to image microscopic biomedical processes include Fluorescence Lifetime Imaging Microscopy (FLIM), which makes it possible, for example, to image ion channels inside neurons and measure the amount of calcium present—useful in studying Parkinson’s, Alzheimer’s, epilepsy, and other conditions where calcium is a factor.

Other applications for the camera, currently under exploration, include intracellular DNAsequencing and proteomics for drug discovery, basic scientific research for gene sequencing and protein-folding, cell membrane scanning to discover what bacteria or other material are present.

Used in combination with MRI, the Megaframe camera has been tested successfully in animals for fluorescence imaging in fields up to 9.4 Tesla.

An imaging nanoparticle that can cross the blood-brain barrier specifically targets tumor cells, to help surgeons pinpoint the boundaries of brain tumors. It is made of an iron-oxide sphere coated with a a protein that attaches to tumor cells and a protein that fluoresces when it does so. In mice with brain tumors, the nanoparticle improved contrast in brain imaging scans.

Holographic Advance

A re-writeable photorefractive polymer material enables close-to-high-definition holographs to be created in minutes—much faster than current methods. The experimental material was stable throughout hundreds of write and erase cycles. The material’s developers have automated the process of capturing, writing, and erasing holographic images from MRI, CT, and other sources. Eventually, surgeons could use such holograms to plan procedures, and pharmaceutical companies could use them to study molecular interactions in drug research.