The unique characteristics of light dictate how it interacts with human tissue. Harness this interaction efficiently and light becomes a potentially powerful imaging tool to aid in the early diagnosis of cancer and other diseases. Diagnosis at this stage can be either a mixed blessing or a blessing in disguise, depending on a variety of issues, including the rate of abnormal-cell growth. Some types of cancer are potentially lethal within months, others not for 20 or 30 years.
The abnormal cluster of cells lies hidden from probing hands by the dense tissue surrounding it. Routine mammograms will not indicate the presence of microcalcifications for several years, and when they do, still may not spot the lesion. By then, if the cells are hungry enough, they will have been feeding voraciously off blood shed by the nearby vascular system for some time. They may be ready to multiply.
The race is on to develop new noninvasive diagnostic devices to hunt down such cells in the breast and throughout the human body at the earliest stages of the cancer life cycle. Refining our hunting techniques, however, is a two-step process. First, we must first understand exactly what is going on, a task that once seemed insurmountable. Now, however, research efforts such as video-microscopic techniques developed by researchers from Duke University Medical Center and Duke Comprehensive Cancer Center (Durham, NC) are bringing us closer than ever before to understanding why some abnormal cells never grow and why others begin a feeding frenzy.
Second, we must fine-tune the weapons used to hunt down the disease. X-ray mammography, perhaps the most widely used of cancer diagnostic tests, has its problems, which range from false-positive results to missed lesions. While it still remains the frontline diagnostic technique in the fight against breast cancer, it may soon have help. The special properties of light could help optical imaging "see" what other diagnostic methods, including conventional x-ray mammography, may miss.
After skin cancer, breast cancer is the most frequently diagnosed cancer in women. Second only to lung cancer in cancer-related deaths, it is projected to kill more than 40,000 women in 2000 alone.1 These statistics may explain why breast cancer appears to be one of the more popular "diseases" targeted by optical imaging researchers—or perhaps the reason is simply that the breast is easily accessible compared to other body parts. There also is a wealth of data from techniques such as x-ray mammography and nuclear magnetic resonance (NMR) for baseline data comparisons.
Whatever the reason, any woman who has ever had a conventional mammogram looks forward to the day she is offered a "comfortable" optical imaging alternative. And if that day never comes, most women will be satisfied with what may be a more realistic alternative—an optical biopsy.
Men also will benefit from the research, in part because they are not immune to breast cancer. Equally important, however, is the fact that many of the optical techniques currently under evaluation as breast-cancer diagnosis aids also could help in the diagnosis of other conditions (see "OCT reveals tissue microstructure").
For example, according to professors Robert Alfano and Swapan Gayen at the Institute for Ultrafast Spectroscopy and Lasers, City University of NewYork (CUNY; New York, NY), the body parts with potential to be investigated by just one technique—optical spectroscopy imaging—extend beyond the breast to the prostate, brain, bladder, bone, cervix, colon, eye, digestive and gynecological tracts, skin, and teeth.
The power of light
Simply put, light is a flexible design tool. Alfano and Gayen believe that various properties of light, along with the ways it interacts with tissue, may provide multiple windows to peer inside the human body.2
"The purpose of an optical body-scanning modality is to identify, locate, and diagnose a lesion inside the body," says Alfano. "You illuminate the relevant area, such as the breast or the prostate, with bright light of appropriate wavelength and then search for indications of pathology in the emergent light—the differences between the interaction of light with the lesion and the surrounding tissues."
In many cases, laser light is a common component of biomedical optical imaging experiments because of its built-in flexibility. Each of the enabling attributes—wavelength, coherence, polarization, intensity, directionality, and pulse duration in the case of modelocked and Q-switched lasers—can change as a result of light/tissue interaction that depends on the wavelength of light. According to Alfano, light between 700 and 1300 nm is less absorbed by tissue constituents than near-ultraviolet, visible, or infrared light and is commonly used for breast tissue three-dimensional (3-D) imaging.
Also important is the interaction of light with biological tissue in terms of scattering, absorption, specular and diffuse reflectance, and diffuse transmission. "Light transmitted through a scattering medium breaks up into ballistic, snake, and diffuse components," says Gayen. "The early-arriving forward-propagating ballistic and snake photons carry image information, while the scattered diffuse light blurs the image and in extreme cases buries it in the background noise. The difference in transit time, as well as changes in light characteristics, are exploited for selecting the image-bearing light to form direct transillumination images or shadowgrams." One problem is that the image-bearing light can become too weak to form a shadow image in a highly scattering, thick medium. Researchers must then reconstruct an image using inverse methods based on the scattered light-intensity patterns measured around the object, a mathematical model for description of light propagation in scattering media, and sophisticated computer algorithms.
Alfano and Gayen are part of a research team at CUNY working to understand how light interacts with human tissue. Their research ranges from optical imaging to spectroscopic techniques to develop a molecular footprint of diseased tissue.
Once all the components are in place—the right wavelengths, image-processing algorithms, and so on—researchers investigating optical techniques for biomedical applications, such as cancer detection, face yet another hurdle—clearance of the resulting device by the Food and Drug Administration (FDA; Washington, DC). Not many optical imaging devices designed as aids for breast-cancer diagnosis have made it to clinical trials, partly because so few are ready for market and the cost is substantial. Of the optical and nonoptical imaging devices just approved for market, the FDA does not suggest that they are alternatives to x-ray mammography. Instead, these techniques so far appear to have been approved as adjunct devices—optical biopsy techniques.
The optical biopsy
In December 1999, OmniCorder Technologies (Stony Brook, NY), received clearance from the FDA for an infrared detector system designed to locate blood flow affected by nitric oxide emissions from a cancerous lesion. The BioScan system—which may be the first optical technique to provide feedback as both a digital image and a single assessment number indicative of positive or negative results—uses a detector tuned to 8-10 µm to locate changes in blood flow in tissue surrounding cancerous cells. The system is sensitive to temperature changes of less than 0.027°F.
"Early in their life cycle, cancer cells go through three basic stages," says OmniCorder Technologies CEO Mark Fauci. "First, the abnormal cells cluster near vascular tissue, which sheds blood that they feed on. In the second stage, angiogenesis occurs. As the cancerous lesion grows, it emits nitric oxide and the vascular tissue surrounding it begins to relax. Eventually, the lesion will begin growing its own vascular tissue, and the surrounding tissue will become necrotic." Fauci adds that researchers believe the emission of nitric oxide can be detected and precisely measured several years before the presence of a detectable calcified mass (see "The end of the road for cancer?").
The BioScan camera uses a large focal-plane array of gallium arsenide quantum-well infrared photodetectors (QWIPs) developed at the Jet Propulsion Laboratory (Pasadena, CA) to tune to the required wavelength. The device also contains a cooler or closed-cycle refrigerator about the size of a fist. The small motor cycles cooling gas millions of times and cools the camera from room temperature to very low temperatures (-343°F) in about 10 minutes.
In combination with dynamic area telethermometry, a computerized data-analysis technique, the unit "images" the target area and provides the physician with immediate diagnostic information. Unlike conventional mammography, the device is not affected by breast-tissue density and so is equally effective on patients of any age. The patient experiences no discomfort and does not receive any ionizing radiation.
In preliminary studies of the detector, which is still in clinical trials, each subject sat with both hands on top of her head, while the camera was positioned about 50 cm from a breast to obtain medial, frontal, and lateral images. After the camera was positioned and focused, the subject was asked to take a deep breath and sit still. The computerized camera allowed the accumulation of 2048 consecutive digital thermal images in 20.5 seconds.
Another device, a computed tomography laser mammography (CTLM) system from Imaging Diagnostic Systems (Fort Lauderdale, FL), may have the potential to eventually supplant conventional mammography, but this would be some way off. The equipment still has a cost that limits its widespread use and it is still in clinical trials.
Laser mammography works much like x-ray spiral computed tomography (which builds a 3-D x-ray-based image), except that the technique uses a 360° laser scan to map the local changes in optical scattering and absorption throughout a thin cross section of breast tissue. Multiple sections are imaged and the results combined to produce a 3-D image. The system's diode laser emits between 750 and 800 nm, and a double-row detector array can image both the absorption and fluorescence of breast tissue.
One major advantage of this device over x-ray mammography involves its efficiency at seeing beyond microcalcifications, even in dense breasts. The device also eliminates the need for breast compression. The patient lies flat on a table with the breast to be scanned dangling through an opening in the table top that holds the detector array. The 3-D image is completed in about the time it would take for a conventional mammogram.
X-ray still useful
Both Omnicorder Technologies and Imaging Diagnostic Systems offer promising noninvasive diagnostic techniques that allow the woman or man to retain his or her dignity during an examination. The techniques face stiff competition from other imaging methods, though. For example, although magnetic resonance imaging has some flaws related to cancer diagnosis, such as an inability to spot microcalcifications, there is evidence of its potential benefits as a backup to conventional mammography. A project sponsored by the National Institutes of Health is currently evaluating the effectiveness of the technique for breast-cancer diagnosis.
The FDA also has just approved a hand-held impedance-measuring imaging device designed to help radiologists determine "whether a woman should be evaluated further when the results of her mammograms are ambiguous." The T-Scan 2000 developed by TransScan Medical (Ramsey, NJ) uses a hand-held scan probe placed on the breast to evaluate certain suspicious areas detected on the mammogram. The images, produced when a small electrical signal is passed through the body, are based on differences in the electrical impedance between malignant tumor tissue and surrounding normal tissue. Based on sensor feedback, the computer produces an image that contains bright spots where the impedance values are consistent with a possible malignancy. As a condition of FDA approval, TransScan is conducting a post-market study on the effects of hormonal changes during a woman's menstrual cycle on the device's ability to detect and distinguish among breast abnormalities.
X-ray technology also is advancing (see "X-ray vision improves"). One example comes from a research team led by Etta Pisano, professor of radiology at the University of North Carolina at Chapel Hill, and Dale Sayers, a professor of physics at North Carolina State University (Raleigh, NC). The scientists are developing a diffraction-enhanced imaging (DEI) technique that improves dramatically on traditional x-ray image clarity. Other project participants are Brookhaven National Laboratory's National Synchrotron Light Source (Upton, NY), the Illinois Institute of Technology (Chicago, IL), and the European Synchrotron Radiation Facility (Grenoble, France).
According to Sayers, the DEI technique differs from traditional x-ray radiography in that an "analyzing" crystal is placed in the x-ray beam between the object being studied and an image-creating medium such as film, x-ray plate, or digital detector. The silicon crystal diffracts a specific wavelength of x-ray because of Bragg's law. When the crystal is adjusted and two images taken and processed, the result is a single image based on x-ray absorption that is similar to a standard x-ray and a new image based on reflection.
Pisano admits the team still has a long way to go before it has a practical machine that can be used clinically. "Obviously we can't bring everybody to the world's few synchrotrons for routine mammography, and so we have to make this technology clinically portable. Right now, we think that is possible," Pisano says.
A word of caution
Regardless of the publicized problems with x-ray mammography, from false positives to missed lesions, the technique still provides the most-readily-available front-line imaging option to aid clinicians in breast-cancer diagnosis. There is no question, even among the technique's naysayers, that the widespread implementation of x-ray mammography that began in the 1980s has been a contributing factor in the steady decline in deaths from breast cancer.
Unlike x-ray mammography, the optical diagnostic techniques discussed in this feature are still either in clinical trials and/or not readily accessible to the general public.—Paula Noaker Powell
REFERENCES
- NCI, "The Nation's Investment in Cancer Research: A Budget Proposal for fiscal year 2001," www. nci.nhi.gov (Jan 2000).
- S. K. Gayen and R. R. Alfano, Optics Express 4, 475 (May 24, 1999).
OCT reveals tissue microstructure
Although human tissue transmits light relatively well in the red and near-infrared, scattered photons muddy the view enough that subsurface features remain hidden to most imaging techniques. Optical coherence tomography (OCT) takes advantage of the short coherence length of a broadband source to skirt this problem. In an OCT scanner, infrared light is sent down a single-mode optical fiber and into tissue. The imaged spot is scanned laterally and the backscattered light collected by the same fiber. The fiber is part of an interferometer in which the reference arm can be changed in length. The interferometer allows the OCT scanner to select only singly scattered light having a path length equal to that of the reference arm. Varying the reference arm is equivalent to an axial scan through the tissue; a two-dimensional image can be built up from the gathered information.
The great advantage of OCT is that it allows images to be created of live tissue. Researchers at Massachusetts General Hospital (Boston, MA) and Harvard University (Cambridge, MA) have built an endoscopic OCT system that allows them to see the microstructure of a patient's gastrointestinal (GI) tract. Such information can potentially help a doctor diagnose cancers or precancerous conditions.
The light source for the OCT instrument is a multiple-quantum-well semiconductor optical amplifier that emits light at a 1.3-µm center wavelength with a bandwidth of 80 nm. The short coherence length of the source results in an axial resolution for OCT imaging of 9 µm. Visible light emitted by a 630-nm diode laser aids in aiming the infrared light. The instrument was built at Massachusetts General Hospital and has been used for imaging in gastroenterology, gynecology, urology, otolaryngology, and cardiology, according to Brett Bouma, one of the researchers. Specific probes have been developed for each of these fields.
The group has imaged the GI tracts of 170 patients, including the esophagus, stomach, and colon. The studies to date are pilot studies, notes Bouma. "The intent is to demonstrate that it is possible to visualize the microstructure of the GI tract and to establish the correlation between the observed structures and the presence of pathology," he says. The first step is to show that OCT can aid biopsy effectiveness in the GI tract by finding the locations of most-severe pathology.
A 5.5 x 2.5-mm section of a normal esophagus imaged by OCT depicts its tissue layers (top image). A second 14 x 2.5-mm image, of a gastroesophageal junction, shows normal gastric cardia at the left and normal esophageal epithelium on the right. Between the two sections, the OCT image indicates Barrett's metaplasia—a condition that, although benign, can turn into cancer. Subsequent biopsy of the area has shown this evaluation to be correct.—John Wallace
The end of the road for cancer?
Like a thief in the night, an abnormal cell can eventually steal enough nourishment from surrounding tissue to kill it, but not in the mouse in this photo. With help from researchers at New York's Memorial Sloan Kettering Cancer Center, this mouse was born resistant to a variety of cancers.1 Robert Benezra and colleagues accomplished this by breeding him and other mice so that they lack the genes responsible for angiogenesis. The goal of the scientists' research, which is still in its infancy, is not to someday breed humans without the genes. That would be too Orwellian and probably impossible because the human fetus requires the genes to develop blood vessels. Instead, the researchers are searching for some way to turn these genes off in adults.
Until such time as we can simply "switch off" cancer development in humans, mice continue to play another role in cancer research—helping researchers both to image cancer development in real time and to evaluate the impact of engineered anti-angiogenesis drugs. Using video-microscopic techniques, researchers from Duke University Medical Center and Duke Comprehensive Cancer Center (Durham, NC) have documented for the first time the earliest steps of the organization of cancer cells into tumors in mice and rats. To visualize the tumors' inception, the researchers made the cancer cells glow by modifying them to express a molecule known as green fluorescence protein. The glowing cells were then injected into a special window chamber on each rodent's skin that allows the tumor cells to receive needed nutrients and the researchers to monitor tumor growth without sacrificing the animal.
Three or four implanted cancer cells usually survived the first few days and, as they multiplied, they reached toward nearby existing blood vessels. By the eighth day, the cancer cells, now numbering 100 to 300, had created visible, fully functioning new blood vessels, said principal investigator Mark Dewhirst, a Duke professor of radiation oncology.
Even though up to eight days were needed for the growing tumor to form new blood vessels, a second experiment showed that existing blood vessels and cancer cells must be able to communicate somehow for the implanted cancer cells to survive. When a drug that inhibited a specific angiogenesis pathway was injected along with the cancer cells, all of the cells died without reaching out toward existing vessels. Without the inhibitor, at least a few cells always survived to begin forming a tumor.
According to the Duke researchers, the fluorescence technique has perhaps one major advantage over other optical imaging techniques for imaging cancer cells. Even with the latest imaging equipment, tumors can still be hard to distinguish from other tissue. With the Duke technique, only a blue light is necessary to illuminate the cancer cells for viewing with a video camera and the glowing target tumor becomes the source of the light.—Paula Noaker Powell
REFERENCE
- S. Salvatore, "Researchers breed cancer-resistant mice," CNN.com (Oct. 14, 1999).
X-ray vision improves
With the advent of digital techniques that allow computer enhancement of x-ray images and methods such as spiral computed tomography (CT), which in essence builds a 3-D x-ray-based image, the x-ray can provide more image detail then ever before. One result—which should make even the optical imaging scientists mentioned in the main article happy—is the virtual colonoscopy, a safe and reportedly "comfortable" colon screening technology that uses computer graphics and computed tomography to develop a 3-D virtual image inside a patient's colon. One eventual outcome could be the elimination of the need to use conventional colonoscopy and barium enema to detect polyps or carcinoma of the colon.
Scientists at the Center for Visual Computing at the State University of New York at Stony Brook (Stony Brook, NY) have used advanced visualization techniques for imaging the mucosal surface of the colon, their goal being to develop a method for automatically generating images of the inner structures of the colon, as well as a flythrough animation along the inside of the colon. After the patient's colon has been cleansed and distended with air, a spiral CT scanner obtains a sequence of 2-D thin axial slices from the top of the splenic flexure of the colon to the rectum based on landmarks obtained from a scout image. These data are then reconstructed into a 3-D volume and visualized with a volume visualization software program developed at Stony Brook.
During experiments, the researchers also successfully applied their automatic animation technique to a data set from the Visible Human Project. This project through the National Library of Medicine is in the process of creating complete anatomically detailed 3-D representations of normal male and female human bodies based on acquisition of transverse CT, MR, and cryo images of representative male and female cadavers.
According to project director Arie Kaufman, unlike other approaches to 3-D virtual colonoscopy that extract the colon surface and represent it as a series of triangles, the Stony Brook approach is volume-based. The user can both visualize the mucosal surface of the colon and explore the data beyond the surfacea virtual surgery cut can be applied to a polyp or the outer layers of the polyp can be rendered translucently to reveal the information from inside the polyp.—Paula Noaker Powell
