Each year oral cancer kills about 10,000 people in the US alone. Another 20,000 are diagnosed with it. Of those who survive, many suffer painful surgeries to remove tumors and restore lost oral and facial tissue that otherwise would have left them disfigured. As with most cancers, early detection holds the key to successful treatment.In 96 percent of oral cancer cases, the tumors start as small lesions in the mouth. These lesions are often visible, but only some actually become cancer. At present, the only way for a dentist or doctor to know whether an oral lesion is cancerous is to perform a biopsy. That, however, poses its own problems. When a painful biopsy turns out to be noncancerous, patients are less likely to cooperate with future biopsies.Researchers at Beckman Laser Institute at the University of California, Irvine, have developed two methods that could help. They use optical coherence tomography (OCT) and polarimetry to distinguish accurately between cancerous and noncancerous oral lesions. The OCT system is already in the early stages of human clinical studies. In March, the scientists presented this work at the 85th general session of the International Association for Dental Research in New Orleans.Polarimetry imaging based on the Mueller matrix eventually may allow dentists to distinguish between normal and dysplastic tissue. After only 14 days, the application of DMBA to hamster cheek pouches begins to create dysplasia. At this stage, the dysplasia cannot be distinguished with the naked eye, but it creates retardance two to three times greater than that of normal tissue picked up with polarimetry imaging (a). By day 28, the dysplastic tissue has a retardance four to five times that of normal tissue (b,c). Courtesy of Jungrae Chung (SCC = squamous cell carcinoma).A noninvasive diagnostic modality would permit regular monitoring of these lesions, which are visible to the naked eye before their transformation into cancer, according to Beckman researchers Jun Zhang and Jungrae Chung. Because the risk factors for oral cancer are clearly defined — tobacco use, alcohol abuse, living in an urban environment, poor diet, frequent exposure to sunlight or belonging to certain ethnic groups — a noninvasive system could be used effectively to screen people with these risk factors to prevent many more serious cases of cancer.Using optical coherence tomography (OCT), researchers at Beckman Laser Institute at the University of California, Irvine, can distinguish healthy oral mucosa (a) from dysplastic oral mucosa (b) and malignant oral mucosa (c), seen here as a three-dimensional reconstruction. OCT images of dysplastic mucosa compare well with standard histology (d). Courtesy of Jun Zhang.To study the ability of OCT and polarimetry to distinguish accurately between oral cancer and noncancerous lesions, the researchers used the Golden Syrian hamster cheek pouch model, which is standard for oral cancer, applying 0.5 percent 9,10-dimethyl-1,2-benzanthracene (DMBA) in mineral oil, which induces squamous cell carcinoma in the treated area after 10 to 12 weeks. They tested the ability of the methods to accurately detect the cancer as compared with standard biopsy.For several years, OCT has been employed in ophthalmology for studying and detecting retinal diseases. However, Petra Wilder-Smith, a dentist and lead researcher at Beckman, explained that it has taken time to find other uses for OCT because ophthalmic systems use shorter wavelengths, and the optical probes have different specific needs. Based on a Michelson interferometer, OCT is particularly well-adapted for assessing small areas because it can image features from 1 to 15 μm. It is noninvasive and could be applied when using surgical scopes or catheters in blood vessels or in internal organs.Zhang said that the group uses three-dimensional OCT instead of the more common two-dimensional method because 3-D better identifies multiple lesions and provides superior mapping of lesion margins. However, it requires a high-speed system and a 2-D scanning probe. The image processing also must be fast enough to process the high volume of data. “We developed our OCT system, scanning probe and image processing for real-time 3-D imaging,” he said. OCT relies on splitting a beam of broadband laser light into a reference and a sample beam. The reference beam strikes a mirror and returns, and the sample beam interacts with the sample, is reflected and returns. The two split beams are recombined and analyzed. The interference between them creates the image of the sample. To build the OCT system, the Wilder-Smith group worked with a research team led by Zhongping Chen. The scientists used a swept source Fourier domain OCT system. For illumination, they used an HSL-2000 swept light source from Santec Corp. operating at 1310 nm with a FWHM bandwidth of 100 nm and output power of 5 mW. The light source was operated at the sweeping rate of 20,000 Hz. For detection, the investigators used an amplified InGaAs photodetector from Thorlabs Inc. and ran the signal through a high-speed analog-to-digital converter from National Instruments. Lastly, Zhang said, they used three kinds of sampling probes: a linear scanning microscope probe made with a motorized linear stage from Newport Corp., a galvo scanning probe with a galvanometer scanner from Cambridge Technology and an endoscopic microprobe developed and assembled by the researchers themselves.When compared with standard histology methods, the OCT images clearly showed moderate to severe dysplasia and squamous cell carcinoma, as evidenced by a noticeable loss of normal epithelial stratification with invasion of underlying layers. PolarimetryThe researchers also worked with a technique that relies on Mueller matrix imaging polarimetry. This matrix relates the Stokes vector of the light impinging on a sample to the Stokes vector leaving sample models. “Using the method with the known input and output polarization states, the 4 × 4 Mueller matrix can be used to describe the polarization properties of a sample,” Chung said. She explained that, in tissue, collagen is linearly birefringent. Squamous cell cancers tend to destroy the natural collagen structure, making it depolarize the light less than healthy tissue does.Chung studied the same animal subjects as Zhang. The polarimetry study showed that sites with squamous cell carcinoma had a four to five times greater retardance of the polarization than healthy sites. Likewise, sites with dysplasia had two to three times greater retardance. She said that the polarimetry images were particularly useful for mapping areas of cancer and lesion margins.Chung explained that the polarimetry system, also built with Zhongping Chen’s group, consisted of a fiber optic illuminator from Dolan-Jenner Industries. The light is passed through a laser line interference filter with wavelength transmittance from 400 to 700 nm. For detection, the team used a 12-bit CCD camera with a chip size of 1024 × 1024 pixels and a pixel resolution of 14 × 14 μm from Kodak. The detection system also included a high-speed data acquisition board from National Instruments and a four-channel analog voltage generator to control the electro-optic variable retarders to achieve a given polarization state.Taken together, these two techniques could provide useful information for screening patients for oral cancer. Wilder-Smith explained that, at first, she expected that a clinical system would be used mostly by oral surgeons or other oral cancer specialists, who are in the best position both to screen for cancer and to remove the lesions for treatment or biopsy.At present, Zhang said that the group has not combined the OCT and polarimetry systems into the same probe because the two systems use different wavelengths and setups. Zhang said that the fiber-based polarimetry system he plans to build could be combined with the OCT system. “In addition, we have developed a polarization-sensitive OCT system that combines the polarization detection and OCT techniques for measurement of tissue birefringence,” he said.The group continues to work on developing the system further. Although human trials are already under way, Wilder-Smith said that the group wants to develop faster 3-D probes and a clinical system with standard probes for the oral cavity.