Laser mammography continues development

Near-infrared lasers are used to obtain mammography images in a technique being developed by Imaging Diagnostic Systems Inc. (IDSI; Plantation, FL). Its second-generation prototype is now being installed at Strax Clinic (Ft. Lauderdale, FL) for continued clinical evaluation, and at least five other sites are planned for 1996. But it will likely be a while before the technique becomes widespread. Hurdles such as optimum laser/detector acquisition parameters and efficacy need to be overcome before the US Food and Drug Administration (FDA) can approve the technique for general diagnostic use.

Like data from computed tomography (CT) or magnetic resonance imaging (MRI), the images the system acquires are composed of a series of image slices that can be manipulated by a computer. Standard x-ray mammography produces a projection image that cannot be segmented by slice. IDSI refers to its technique as CT laser mammography (CTLM).

IDSI`s first prototype was originally demonstrated acquiring images of breast tissue with silicon implants. While the system was able to successfully see around the prosthesis, a number of deficiencies in the optics, mechanics, and software were noted. As a result, the second prototype was developed. "To optimize imaging of breast tissue and various anomalies, we needed a laser whose parameters could be adjusted," says IDSI CEO Richard Grable. "But we had to build a custom laser because no commercial product could provide the range of adjustments we needed in a single laser." The company developed a argon-ion-laser-pumped modelocked Ti:sapphire laser that can be tuned from 750 to 1100 nm with variable repetition rates from 70 to 100 MHz and a pulse width about 100-200 fs long. "Right now, we typically run the laser at 840 nm, with a 76-MHz rep rate and about a 100-fs pulse width," says Grable. "Ultimately, commercial systems will not feature such parameter tunability, however," he adds.

To acquire data, a woman is positioned face down on a table that contains the imaging chamber and a laser/detector array for scanning breast tissue. The laser output beam projects horizontally onto one breast at a time, and energy is both transmitted and scattered. A cluster of about 300 silicon avalanche photodiode detectors collects the forward scattered energy, and another 300 more detectors further distributed around the breast collect scattered energy at various angles. The laser/detector array is then mechanically scanned around the breast to acquire data from many perspectives. The entire cluster is then moved down about 2 mm, and the acquisition process repeated. The average scanning time is less than five minutes per breast.

The scattering of photons in a dispersive media, such as tissue, reduces detectable signal levels and reduces useful image resolution. To minimize these difficulties, IDSI uses range-gating with fast detector electronics (in the gigahertz range). With this method, the researchers essentially determine when scattered photons will arrive at the detector and collect the signal only in a narrow time slot. This significantly reduces background noise, boosting the signal-to-noise ratio.

Resolution of the system is currently limited by the spot size of the laser and the thickness of the imaging slice. Both can be decreased, but acquisition time would have to be longer. "Currently, we can detect abnormalities in the millimeter range," says Grable, which is about the limit of detection in standard x-ray mammography.

The company plans to negotiate a clinical protocol with the FDA very soon, but it is not yet clear whether the FDA will require a 510k or a much tougher PMA protocol. "The next step," says Grable, "is to stop talking and demonstrate we can produce diagnostically relevant images that are comparable to standard x-ray mammography." Other potential applications for the technology include imaging of the testicles, newborns, and young children.