So you came up with this great idea for a medical device for cancer imaging, and even found a way to make it work. Now what? Several researchers involved with coherence imaging technologies discuss the challenges of clinical translation. You have probably heard the facts and figures: One of the leading causes of death worldwide, cancer accounted for 7.6 million – roughly 13 percent of all deaths – in 2008, according to the World Health Organization. And deaths from cancer are on the rise: By 2030, we will likely see 13.1 million per year worldwide. The need for new ways of diagnosing cancer has never been greater. Today we are seeing a variety of new diagnostic techniques that leverage the benefits of noninvasive and less-invasive optical technologies – and that are made possible by the development of novel treatments. “Opportunities for these techniques have been advanced by new therapies,” said Adam Wax, professor of biomedical engineering at Duke University in Durham, N.C. “We’ve seen this in a number of cancer models, with radio-frequency ablation and cryospray ablation, for example.” With therapies targeting cancers at earlier stages, we need diagnostics that can detect those early cancers. Take coherence imaging, for example. Wax was one of several researchers who spoke about coherence imaging and cancer during the BiOS Hot Topics session at this year’s Photonics West meeting in San Francisco (many of the Hot Topics talks can be viewed on the SPIE website and on YouTube). In his talk, “Early Cancer Detection with Coherence Imaging,” he described a suite of spectroscopic techniques designed to assess cell structure and diagnose disease using low-coherence interferometry (LCI) to detect scattered light. The techniques combine the advantages of optical coherence tomography and light-scattering approaches. Angle-resolved LCI, for instance, marries the ability of LCI to isolate scattering from subsurface tissue layers to the ability of light-scattering spectroscopy to obtain structural information using angular scattering measurements. Neil Terry (left), Adam Wax (right) and their colleagues at Duke University have described a technique called angle-resolved low-coherence interferometry that can detect dysplasia in patients with Barrett’s esophagus, for example, and have developed it further for clinical application. Courtesy of Adam Wax. The researchers explored the clinical potential of angle-resolved LCI for in vivo depth-resolved nuclear morphology measurements to detect dysplasia in patients with Barrett’s esophagus, who are at increased risk of developing esophageal cancer. The results, reported in the January issue of Gastroenterology, showed that the technology can provide quantitative depth-resolved measurements of nuclear morphology – measurements used by pathologists for cancer diagnosis – without having to rely on image interpretation or use of exogenous contrast agents. Also during this year’s BiOS Hot Topics session, Stephen Boppart, Bliss professor of engineering at the Beckman Institute for Advanced Science and Technology at the University of Illinois at Urbana-Champaign, spoke about his work with coherence imaging and cancer. In his talk, “Coherence Imaging of Cancer with Novel Optical Sources,” he described a technique called nonlinear interferometric vibrational imaging, or NIVI. Researchers at the University of Illinois at Urbana-Champaign have reported a technique called nonlinear interferometric vibrational imaging, or NIVI, and are now developing it for breast cancer detection in the clinic. Shown here is a NIVI image of a tumor, compared to histology. Courtesy of Stephen Boppart. NIVI offers the high spectral resolution of Raman spectroscopy with the high acquisition rates of coherent anti-Stokes Raman scattering microscopy. Boppart and colleagues have shown that they could obtain NIVI spectra with the accuracy of Raman but at speeds 200 to 500 times faster, and thus demonstrated the potential of the technique for rapid tissue imaging, characterization and diagnosis – for diagnosis of cancer, for example. At the same time, they have been developing novel optical sources to use with the technique. “We’ve worked out ways of controlling the phase and generating a supercontinuum that’s completely coherent,” Boppart said in a phone interview just prior to the BiOS portion of Photonics West. The sources have been described in a series of papers – all involving Dr. Haohua Tu, also of the University of Illinois – as well as in Boppart’s recent Hot Topics talk. Getting into the clinic With technologies intended for clinical application, identifying and solving the problem is, of course, only half the battle. Clinical translation involves a variety of challenges, many of which are unique to this stage of technology development. “You labor under the illusion that you’re going to come up with a solution and companies are just going to run to you with bags of money,” Wax said, but there exist any number of hurdles that must be overcome before the technology gets into the clinic. In parallel with his academic efforts, Wax has been seeking to commercialize the technology through a company he started, Oncoscope Inc., and has run up against several of these. “The whole translational pathway is full of challenges,” he continued. “It’s not the most glamorous work. The most glamorous work is really that first paper” describing the breakthrough. Some of the hurdles have little if anything to do with the technology itself. In the past several years, for example, companies developing new clinical devices and techniques have had to contend with the credit crisis. Other times, they are all about the technology. For example, many devices – which in the early development stages might occupy an entire corner of a room, with fibers and assorted incomprehensible add-ons protruding from them, suggesting nothing so much as an evil scientist’s creation – will have to be redesigned before they can be introduced clinically, providing a reliable, compact, robust, turnkey system. “There have been heroic studies where mode-locked lasers have been brought into the clinic,” Boppart said. “For translation, though, you’ve really got to have these systems better designed and better engineered.” For example, the angle-resolved LCI instrument reported by Wax and colleagues was developed further by Oncoscope to be sufficiently robust for broader use. “The Duke prototype was typically operated by PhD scientists and grad students who could tune up the instrument if needed and were able to instantly assess if something was not functioning correctly,” Wax said. “In contrast, the Oncoscope device needs to be stand-alone, so that a physician can operate it without difficulty.” To achieve this, they re-engineered several of the internal components to make them less vulnerable to outside influences and added automated routines to ensure calibration. The University of Illinois researchers also are working to translate their technology for use in the clinic. At the time of writing, they were waiting to receive final word about an Academic-Industry Partnership proposal they had submitted to the NIH National Cancer Institute to develop NIVI for intraoperative use, detecting molecular tumor margins during breast cancer surgery (the proposal had been scored very highly). Boppart noted several challenges to be addressed in developing the technology for clinical application. These include: (1) developing compact, portable, turnkey fiber-based sources for nonlinear optical imaging to replace the current mode-locked lasers, multi-laser systems or optical parametric oscillators that keep these techniques in the lab; (2) developing the imaging, processing and analysis algorithms to make NIVI diagnostically useful; and (3) developing the portable system cart and handheld probe for use in clinical settings. “These are all the goals for our Academic-Industry Partnership, but are also what is needed to move this field forward,” he said. UK-based Michelson Diagnostics developed and offers a clinical OCT scanner for dermatologists. Courtesy of Michelson Diagnostics Ltd. For this project, the researchers are developing handheld microelectromechanical systems-based scanners for use with NIVI in the operating room. “These, by themselves, are rather novel,” Boppart said, “because few optical probes currently exist for intraoperative use in the sterile surgical field. His startup company, Diagnostic Photonics Inc., has developed such a surgical probe for interferometric synthetic aperture microscopy, a computed imaging approach to OCT, and began clinical trials in February. The NCI proposal included both academic and industry partners, of course, but also a clinical partner: Carle Foundation Hospital of Urbana, Ill. Building strong relations in the clinical arena is especially important to translation, Boppart said. “When you have a good clinical partner, you can step into this very different environment and culture – the clinical setting – and still be accepted.” The technology still must be well designed and as unobtrusive as possible, though, if it has any chance of finding support from the medical community. “You’ll find that, where a lot of these technologies succeed, there’s minimal disruption to the standard of care,” he said. “As engineers, we tend to want to have complicated solutions, but if it causes clinical practice to change too dramatically, it’s just not going to happen.” Academic-Clinical Partnerships Jon Holmes, CEO of Kent, UK-based Michelson Diagnostics Ltd., has a few thoughts about developing technology for clinical application. He developed and offers the VivoSight OCT scanner for dermatologists. The device uses multibeam OCT to obtain higher-resolution, clearer images than can be achieved with conventional single-beam Fourier-domain OCT systems. “Many academic groups working on OCT have developed from physics or engineering departments, and so they are focused primarily on developing the underlying technology,” he said. “Put simply, their research will be published if it studies an advance in technology, whereas papers on developments in the clinical applicability (such as the probe ergonomic design) are less likely to be published.” The upshot, he continued, is that any technology developed by academic groups for biomedical applications might not be properly evaluated, and in many cases it may never reach commercial exploitation. “My advice is that physics and engineering groups should closely partner with clinical teams and work with them on a specific clinical need over a long period of time (decades) in a focused manner with a clear long-term goal of developing an exploitable device evaluated with clinical trials. Funders should also actively support this type of collaborative work.”