Hybrid MR/optical scanner examines breast tissue

Dartmouth researchers have combined MR and near-infrared (NIR) imaging to form a hybrid breast scanner. The prototype, which has been tested successfully on volunteers and some breast cancer patients, is being groomed to detect early tumor growth and to stage tumors by characterizing their vascular and cellular makeup. By developing combined NIR/MR imaging, the researchers hope to evolve this hybrid into a mainstream modality for diagnosing breast cancer and following the treatment of patients.

The two modalities complement each other's strengths and offset their weaknesses. MR provides the spatial resolution that NIR imaging lacks, just as NIR-generated images improve the interpretation of MR contrast enhancement profiles. Specifically, the differential scattering of NIR light may indicate and quantify regions of oxygenated and deoxygenated hemoglobin believed to be associated with cancer, said Brian Pogue, Ph.D., an associate professor of engineering at Dartmouth. MR may provide the anatomical references used to guide reconstruction of NIR data.

"This new system means that we may be able to enhance the information that MR provides by allowing us to image breast tumors with a completely different mechanism of contrast, namely hemoglobin, oxygen saturation, water, and optical scattering," Pogue said.

Despite encouraging results, the technology is still in a very early stage of development, according to Pogue.

"We have designed it for imaging in the plane of the tumor, and, because of that, we sometimes miss the cancer," he said.

The next engineering step is to optimize the system for full breast imaging. The means for processing data also need updating, Pogue said, as the current prototype requires two computer systems--one for the MR and the other for NIR imaging. The end result, however, is promising, as the use of gadolinium fiducials to identify the location of the fiber-optic array on MR allows positioning of the two types of images at millimeter-level precision, he said.

The fiber-optic array developed to test NIR/MR imaging was developed in lab space at the Dartmouth-Hitchcock Medical Center. It was integrated with a Philips 3T scanner operating at the Dartmouth Advanced Imaging Center. Philips and Dartmouth engineers are working together to refine the array, which fits into a breast coil.

Recent progress in hybridizing NIR imaging and MR is part of a broader effort at Dartmouth, the long-term goal of which is to develop and evaluate alternatives to current breast imaging techniques. Specific objectives are to increase the frequency and accuracy with which cancer can be detected and diagnosed, and to improve the prediction and monitoring of disease during treatment and follow-up.

NIR imaging is one of four emerging modalities being studied. The other three are MR elastography, electrical impedance spectroscopy, and microwave imaging spectroscopy. Over the next five years, the program, which is being sponsored by the National Cancer Institute, is expected to gather enough evidence to assess the potential clinical value of these breast imaging alternatives for differential diagnosis. The data will come from pilot studies designed to lead to larger clinical trials.

Early on, the researchers focused on developing a fast, multispectral NIR data acquisition system, utilizing a custom-designed photomultiplier tube that combines parallel and multiplexed detection of photons. A key element in developing the prototype was refining the MR-compatible breast interface to support NIR imaging. The goal now is to develop and evaluate methods for incorporating MR information into the NIR reconstruction process. Much of the early work involved phantoms and freshly-excised breast tissues.

NIR images are based on endogenous absorption and scatter of photons resulting from the presence of hemoglobin, oxygen saturation, and water concentration, according to Pogue. Eventually the researchers plan to enhance the images through the use of a luminescent agent.

Recent human studies have shown that combining NIR and MR imaging in a single exam, though challenging, can be done. Now the researchers are looking to expand their studies to demonstrate the feasibility of such breast scans in comparison with conventional breast imaging, contrast-enhanced MR studies, and pathology analyses of biopsied or excised tissues.

"Once we have built the optimized system (for full breast imaging), probably within the year, we will start up a clinical trial," Pogue said.